In recent decades, increasing demand for energy and severe environmental implications have spurred the switch from fossil fuels to cleaner energy (such as solar, wind, and tidal energies)^[1]. However, owing to daily, seasonal, and regional variables, these renewable energies are unpredictable, restricting their significant success in the global energy mix. Water splitting devices, rechargeable metal-air batteries, polymer electrolyte membrane fuel cells (PEMFCs) and CO₂ reduction are several kinds of novel energy conversion and storage technologies^[2,3,4,5,6]. Even though the architecture of these devices differs, the fundamental processes are almost the same. In a rechargeable metal-air battery, metal dissolution and deposition on the negative electrode (metal), while the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) on the positive electrode (air)^[2]. For the electrochemical CO₂ reduction reaction (CO₂RR), the OER process can serve as the anode reaction^[7]. An unmet challenge for this reaction is the lack of efficient oxygen evolution electrocatalysts (OECs), using a non-noble metal electrocatalyst to pair with the CO₂RR process for high overall energy conversion efficiency^[7]. On the other hand, electrochemical water splitting, which is known as the most viable way of producing pure hydrogen quickly and an indirect way of storing electrical energy obtained from intermittent sources in the form of chemical fuels like H₂, comprises two half-cell reactions: hydrogen evolution reaction (HER) at the cathode and OER at the anode^[8,9]. The overall reaction is predicted to occur at 1.23 V vs. RHE (2H₂O → 2H_2(g) + O_2(g) where E = 1.23 V)^[10]. However, large overpotentials associated with both OER and HER prevent this. The general alkaline OER process in water splitting is summarized in Figure 1^[11]. OER is the kinetically sluggish one because it involves four protons and four electrons coupled transfer with the formation of an oxygen-oxygen (O-O) bond in which each elementary step has its own activation energy. And the one with the highest activation energy is the slowest elementary step, making it the rate-determining step in OER^[12]. Therefore, OER dominates the overall efficiency of the two technologies because it pertains to the creation of O-O bonds and includes four proton-coupled electron transfer processes^[13,14]. As a result, highly OECs are required to accelerate the OER or decrease the overpotential to mitigate the energy loss inherent in energy conversion technologies.

The innovative combination of materials science and water splitting electrochemistry opens new pathways for creating different OECs. IrO₂ and RuO₂ are considered as benchmark catalysts due to their low onset potentials for triggering the OER. However, small reserves and high prices restrict their utilization at a larger scale^[15,16]. To address these problems associated with rare earth elements, scientists have focused on non-precious metals-based OECs, mostly composed of 3d transition metals^[17,18]. In the start, oxides and hydroxides of these 3d transition metals-based OECs have been investigated under alkaline conditions^[15]. Beyond oxides and hydroxides, significant efforts have been dedicated to develop highly efficient electrocatalysts for OER from cost-effective and naturally abundant materials, such as layer structure materials^[18], metal chalcogenides^[19], metal nitrides^[20], metal phosphide^[21], metal borides^[22], metal carbides^[23], and organometallics^[24]. According to recent studies, all non-precious metal catalysts containing anions other than oxides and hydroxides undergo surface chemical reconstruction, resulting in the production of oxides/hydroxides on the surface in contact with the alkaline electrolyte^[25,26]. As a result, regardless of the materials utilized for OER under alkaline circumstances, their (oxy)hydroxides are always the real-time catalyst during OER. Among these transition metal (oxy)hydroxides materials, NiFe-(oxy) hydroxides have been proved to be the most potential materials in alkaline electrolytes^[27]. According to many studies, pure Ni and Fe oxides/hydroxides are not efficient OER electrocatalysts, while their complexes have shown higher OER activity^[28].

The most active electrocatalyst for OER was thought to be Ni(OH)₂/NiOOH at the start of the evolution of 3d transition metals-based OECs. Subbaraman and co-workers suggested that the OER activity trend of 3d transition metal divalent cations goes in the order of Mn²⁺ < Fe²⁺ < Co²⁺ < Ni^2+[29]. And it was believed to be true and accepted widely until it was found that a trace amount of Fe (~ 1 ppm) largely enhanced the OER activity of Ni²⁺ based electrocatalysts^[30,31]. In the late 1980s, the extraordinary OER performance of Ni(OH)₂ anodes-based alkaline batteries was observed in Fe-containing (up to 1 ppm) KOH electrolyte with smaller cell voltages than pure KOH^[32]. This work found that Fe doping from the electrolyte into the Ni(OH)₂ electrode reduced both the cell voltage and the OER onset overpotential in an alkaline medium. And this was the first time that researchers discovered the Fe effect on increasing the OER activity of Ni(OH)₂. It was observed that the OER activity of a Fe-doped β-Ni(OH)₂ electrocatalyst increases linearly as the quantity of Fe is high enough to enter the β-Ni(OH)₂ lattice. And as a result, it was assumed that Fe might be the active site for OER^[33]. Therefore, Ni-based OECs attracted great attention, and a huge amount of research was conducted to determine the processes by which Fe improves the OER activity of Ni-based OECs.

Without a doubt, Fe incorporation has been the driving force behind the significant OER improvement seen with Ni-based OECs in unpurified, reagent grade KOH electrolyte so far. OER activities were being massively reported for these materials without realizing that Fe in concentration as low as 1 ppm may significantly change the OER activities of Ni-based OECs. Based on these findings, much research has been published in literatures detailing the various functions of Fe in the matrix of Ni-based OECs. Important roles stated in literatures with experimental pieces of evidence are: 1) Incorporation of Fe in place of Ni in their respective hydroxides/(oxy)hydroxides matrices makes them harder to be oxidized further and thus alters the redox electrochemistry of Ni, 2) When Fe is incorporated into Ni-based OECs, it causes structural changes, 3) It is well-known fact that Fe³⁺ serves as the strongest transition metals-based Lewis acid with a stronger electrophilic nature and notably affects the electronic properties of the other cations in which being incorporated; hence it changes the Ni oxidation states and aids in the formation of active Ni⁴⁺ sites, and 4) Fe³⁺ ions in NiFe-based OECs with a 1:0.33 ratio depicts a superexchange magnetic contact that improves OER electrocatalysis by allowing electrons to hop.

Almost all researches involving Ni-based OECs and Fe incorporation have revealed that the most appropriate and active composition of Ni:Fe for OER is 1:0.2 ~ 0.3. NiOOH is an excellent electronic conductor, with no change in conductivity before and after Fe inclusion. Boettcher and co-workers have recently discovered that the effective conductivity of NiOOH films diminishes as the cycle number increases and becomes independent of the quantity of Fe incorporated into the film^[34]. The decrease in conductivity with increasing cycle number did not affect the OER activity and substantially increased. Figure 2 illustrates the cyclic voltammetric (CV) cycles of a NiOOH film electrode in a KOH electrolyte with and without Fe-spiking, demonstrating an increase in OER activity despite a reduction in effective conductivity.

As a result, it is apparent that the observed increase could not be attributed only to the change in conductivity that occurs when Fe is incorporated into Ni-based OECs. The fact that Fe in FeOOH is insulating, which is a big contradiction. In this way, how could it improve the conductivity of NiOOH when combined? Bell and co-workers proposed a plausible mechanism in which Fe³⁺ ions were incorporated into β-NiOOH lattices in smaller amounts and exhibited an extraordinarily shrunken Fe-O bond length, resulting in Fe³⁺ sites in NiOOH lattice getting the required conductivity not from their incorporation but from the conductivity of NiOOH lattice. They also discovered that increasing Fe concentration results in the development of insulating FeOOH surfaces, which decreases the OER activity. The incorporation of Fe into Ni-based OECs could increase the OER activities, which was ascribed to the change in effective conductivity (potential dependent), but it cannot be the sole one reason.

Fe incorporation or doping has a significant impact on electrochemical oxidation and reduction characteristics of Ni(OH)₂. The incorporation of Fe effect was first observed from the shifting of the Ni(OH)₂ oxidation peak to NiOOH during electrochemical water oxidation^[35]. Boettcher and co-workers observed comparable anodic shifts of Ni(OH)₂ → NiOOH oxidation peak when aged the electrodeposited Ni(OH)₂ film in TraceSelect KOH (~ 36 ppb Fe) and reagent grade KOH (~ 1 ppm Fe) for 1 hour (Figure 3(A-E))^[36]. When the Fe concentration reached 5%, they found a similar change in the Ni:Fe co-deposited film in pure KOH(Fe free). However, there was no such anodic shift in Ni(OH)₂ → NiOOH oxidation peak in the Ni:Fe co-deposited film with 25% Fe content. The most intriguing result of this research was the reduction in the strength of the Ni(OH)₂ → NiOOH oxidation peak when aged in Fe-free KOH and the development of a new peak at a high overpotential, which was ascribed to the production of Ni⁴⁺ ions. They also discovered that OER activity was reduced with aging time and attributed it to the reduction in structural flaws in Ni(OH)₂. However, Ni(OH)₂ films aged in TraceSelect KOH (~ 36 ppb Fe) and reagent grade KOH (~ 1 ppm Fe) showed a substantial increase in OER activity. This suggests that incorporation of Fe boosts activity, although the anodic shift of Ni(OH)₂ → NiOOH oxidation peak in these cases is opposite to what was anticipated. Furthermore, aging had no effect on the co-deposited NiFe-(oxy)hydroxide films. They concluded that the OER activity is intrinsic to NiFe-(oxy)hydroxide sheets and not linked to structural flaws in Ni_0.75Fe_0.25OOH. However, the observed anodic shift of Ni(OH)₂ → NiOOH oxidation peak in their research could not be fully explained.

Later on, Strasser and co-workers published an intriguing result based on their in-situ/operando differential electrochemical mass spectrometric (DEMS) and X-ray absorption spectroscopic (XAS) studies on NiFe mixed oxides under OER conditions^[37]. In this research, they discovered that nearly all Fe³⁺ ions in NiFe mixed oxides remained in the oxidation state of 3+, and that 75% of Ni remained in the 2+ and 3+ states, while the remaining Ni was oxidized to the 4+ state (shown in Figure 4(A-B)). These findings were remarkably similar to those previously reported^[36]. Furthermore, they hypothesized that the kinetic competition between metal oxidation and metal reduction during O₂ evolution may be responsible for insignificant detectable high valence metal deposition^[38]. When there was no Fe or very little Fe (less than 10%), the kinetic competition preferred metal oxidation, but when there was a lot of Fe, the kinetic rivalry favored the metal reduction phase during O₂ evolution. This means that for the substantial OER improvemnet in NiFe mixed oxide catalysts, the required concentration of Fe should be more than 10%. Moreover, in their explanation for why the oxidation peak strength of the Ni(OH)₂ → NiOOH decreased with increasing Fe content was also the kinetic competition and this resulted in most of the Ni remained unoxidized upon polarization in the 2+ state. This is the main reason that anodic shift for (oxy)hydroxide peak occurred in Ni-based OECs. The faradic efficiency of OER decreased with increasing the oxidation (Ni(OH)₂ → NiOOH) peak which rises as the pH of the reaction rises. The findings of this research highlight the importance of Fe concentration, electrolyte strength, and support material during catalyst fabrication. Although there is an explanation for why anodic shift and intensity drop occur in Ni-based OECs with increasing Fe content, but enhancing mechanism by redox inactive Fe³⁺ ions is still a subject of ambiguity.

Gray and colleagues recently published comparable findings in a study of NiFe-LDH (LDH: layered double hydroxide) inducing OER in water-depleted conditions, i.e., in acetonitrile^[39]. They discovered the production of a cis-dioxo-Fe(IV) intermediate immediately before OER. They found that when this intermediate was separated and put into neutral water, it produced hydrogen peroxide, and generated oxygen in case of alkaline water. This was one of the strongest arguments indicating and supporting Fe⁴⁺ ions formation in the presence of Ni during catalyzing water oxidation, whereas previously it was claimed that some parts of Ni were changed to 4+ oxidation state upon Fe incorporation. It was also discovered that the highly active Ni_xFe_1-xOOH catalyst has two active sites with remarkable difference in OER rate constants, and the percentage of fast sites matched well with the fraction of Fe in the catalyst^[40]. These findings clarified that Fe and Ni, both are active sites for OER, with Fe sites being significantly faster. This may be due to the optimal bond energies of intermediates when coordinated with Fe. On the other hand, OER intermediates have a weaker adsorption contact with Ni where they create more covalent M-O bonds, which need a larger potential to break in order to develop O₂. This is why, in the absence of Fe, Ni(OH)₂ needs a greater overpotential to start the oxygen evolution.

As a result of these massive efforts to understand the intrinsic phenomena, phase, and structure of NiFe-based OECs, researchers have reported several conflicting findings. Here are some common and general findings from past researches: 1) Fe³⁺ incorporation/doping into NiOOH generates a greater number of catalytically active sites than a simple rock salt structure with hydroxides, and the content of Fe³⁺ dissolves up to 25% into Ni(OH)₂/NiOOH matrices under applied voltages or during electrolyte aging, making it a superior OEC^[41]. 2) The production of Ni⁴⁺ is reduced when Fe³⁺ concentration rises^[37]. 3) The active site is Fe³⁺, although it is inactive when presented alone as FeOOH at lower overpotentials. This is because the only first monolayer of FeOOH is OER active and electrolyte-permeable at lower overpotentials^[42]. 4) The active sites are Fe³⁺ ions that replaced Ni³⁺ ions in NiOOH matrices at the edges, corners, and defect sites. Those who are deeply embedded in the majority of NiOOH do not take part in OER^[34]. 5) The active Ni²⁺ ions are simply stabilized by Fe³⁺, as indicated by the anodic shift of the Ni(OH)₂ → NiOOH oxidation peak with the decreased integrated charge^[38]. 6) Fe³⁺ ions have a 6-percent shorter Fe-O bond length in NiOOH matrices, suggesting that Fe³⁺ ions may have partly higher oxidation states, and active sites in both NiFe systems are anticipated to be Fe³⁺. The conductive medium needed for Fe³⁺ to catalyze OER is provided by NiOOH matrices^[43]. 7) Fe³⁺ in the conductive Fe₇S₈ phase is OER active in the same way as Fe³⁺ in NiOOH matrices^[44]. 8) Both Fe⁴⁺ and Ni⁴⁺ are formed upon polarization, and they create a synergism in which the high spin (HS) d₄ Fe⁴⁺ stabilizes the oxyl radical generated with Ni⁴⁺ through superexchange contact, while the low spin (LS) d₆ Ni⁴⁺ promotes the production of O-O^[45].

From previous studies, it was not fully cleared that which site is playing an active role in enhancing the activity and how the structure of catalyst evolves during the reaction. This is hampered due to insufficient information about the electronic structure of the composite during the chemical reaction. To observe the electronic and structural changes in the composite during a redox reaction, Corrigan et al. conducted an in-situ⁵⁷Fe Mössbauer spectroscopic study^[46]. They said that the electronic changes at Fe site are not due to the redox process in Fe, but attributed to the host lattice in a composite material. Moreover, an intimate combination of Fe and Ni on an atomic scale rather than the mixture of Ni(OH)₂ and FeOOH particles exhibited better performance for OER rather than either of the single component hydroxides. It was also expected that highly oxidized Fe sites in composite hydroxide may play a key role in OER. Friebel et al. investigated the NiFe oxide structure using in-situ/operando XAS and concluded that most of the Fe remains in the 3+ oxidation state during the reaction and computational studies by Nørskov et al. implicated Fe³⁺ as the active sites for OER⁴³. Moreover, NiFe-based catalyst was investigated at potentials where only Ni²⁺ could be oxidized but not enough to promote OER^[46]. And a change observed in ⁵⁷Fe isomer shift (δ) value from 0.32 to 0.22 mm·s^-1 was ascribed to transfer of electron density away from Fe³⁺ species upon oxidation of the Ni centers.

The presence of Fe⁴⁺ in the OER was first clearly observed in 2015 when Stahl et al. investigated NiFe-LDH electrocatalyst during OER^[47]. They used in-situ/operando⁵⁷Fe Mössbauer technique to track the Fe oxidation state in 3:1 NiFe-LDH and Fe oxide catalysts, while polarizing them in the OER zone. Interestingly, as shown in Figure 5, the existence of Fe⁴⁺ was confirmed upon polarization and under OER conditions, which increased with increasing potential. They found that the pure Fe oxide did not show any Fe⁴⁺ in the OER reaction, demonstrating that Fe⁴⁺ can only be stable in the NiOOH lattice. This also further illustrated that Ni_xFe_1-xOOH, in addition to Ni⁴⁺ generation in the OER reaction, also has Fe⁴⁺ and that the presence of NiOOH lattice favors the generation of Fe⁴⁺. As shown in Figure 5(B), Mössbauer spectrum under open circuit conditions exhibited a doublet with a δ value of 0.34 mm·s^-1 and quadrupole splitting (Δ) value of 0.46 mm·s^-1, and did not show any change at 1.49 V vs. RHE as shown in Figure 5(C) (potential below the onset). At 1.62 V where a significant OER activity recorded, a shoulder peak at δ = -0.27 mm·s^-1 appeared, which indicates 12% of the Fe sites shown in Figure 5(D). Further increasing the potential to 1.76 V generates a stronger peak for Fe oxidation corresponding to almost 21% of the total Fe (Figure 5(E)). After bringing back the potential to 1.49 V, the current density dropped to baseline, but Fe oxidation peak appeared in Mössbauer spectrum accounting for almost ~ 20% of total Fe as shown in Figure 5(F). After 48 hours without any applied potential, the Fe oxidation peak disappeared (Figure 5(G)).

On the other hand, in the case of ⁵⁷Fe-enriched hydrous Fe oxide, they did not observe such a behavior in Mössbauer spectra under all the conditions as shown in Figure 5(H-J). And the spectra showed only a doublet (δ = 0.36 ~ 0.37 mm·s^-1, Δ = 0.64 ~ 0.67 mm·s^-1). The significant feature of NiFe catalyst was the appearance of Fe oxidation peak under an applied potential (> 1.5 V). This new peak was fitted either as a singlet (δ = -0.27 mm·s^-1) or as a doublet (δ = 0.0 and Δ = 0.58 mm·s^-1), and is consistent with the oxidized Fe species as Fe⁴⁺. This was the first direct evidence of the Fe⁴⁺ formation under reaction conditions. Furthermore, the presence of these Fe⁴⁺ species even after lowering the potential where the activity dropped to baseline indicated that these species were not directly responsible for the observed activity but playing a key role in enhancing the OER activity. This discovery showed that Fe³⁺ in NiFe-LDH lattice is redox-active under OER conditions and produces highly OER-active Fe⁴⁺ species. This now explains why Fe exhibits such a significant OER increase when doped to Ni host oxides/hydroxides. Another intriguing result of this research was that no such oxidation occurs when Fe³⁺ is isolated in Fe hydroxide. As a result, a suitable host like NiOOH is required for Fe³⁺ to be redox-active and OER boosting. Therefore, they believed that the stable host lattice of NiOOH is responsible for the production of Fe⁴⁺. Because Fe⁴⁺ species are not anticipated to be identified under the measurement circumstances, they hypothesize that these observed Fe⁴⁺ species are situated at an interior location in the layer and are surrounded by Ni. And suggested these undiscovered Fe⁴⁺ species as potential active locations if they exist, but could not give any clear understanding about the role of Fe during OER. XAS results showed that the bond length of Fe-O in NiFe-based catalysts was significantly shorter than that of Fe³⁺-O, which indicated that Fe should be in higher valence states during the OER process, and supported the results of in-situ/operando Mössbauer studies^[43]. This work gave a strong and clear indication for the existence of high valance iron which surely presents in the system and may have a crucial role in OER, but not much clear. The results of oxygen intermediates study also supported the presence of high-valent iron^[48]. However, there is still debate on the mechanism of NiFe catalyst during OER reaction, whether Ni is active sites or Fe is active sites.

Later on in 2018, NiFe Prussian blue analogues (NiFe-PBAs) were studied for OER through in-situ/operando XAS technique^[49]. It was observed that NiFe-PBA was first transformed into Ni(OH)₂ and then into NiOOH_2-x under the applied potential due to deprotonation, which showed an excellent activity at very low overpotential. They called the deprotonation process a reversible process which partially generated Ni⁴⁺ as a result of charge compensation due to deprotonation. In other words, in-situ generated catalyst NiOOH_2-x during OER was considered as responsible for higher activity which can be transformed reversely to Ni(OH)₂. This indicated that the structure of the catalyst cannot be stabilized under reaction condition and could be changed by applying potential. In fact, this study gave one clear example about dissociation of catalysts under OER condition for in-situ/operando spectroscopic characterization which is a major draw-back for the in-situ/operando study.

It is apparent from the above significant findings that there is no clear-cut mechanism, and it is difficult to reach a consensus on how Fe improves the OER activities of Ni-based OECs and what the structure or real phase of catalyst is playing active role in OER, and how it could be stabilized. There are some studies of NiFe-based OECs using in-situ/operando techniques, but the problems in such studies are that they could not determine the active sites for OER and failed to fully explain the underlying mechanism for OER. Moreover, we believe that the higher activity is crucial, but it should not be the only concern for any OER catalyst. The stability of the structure under OER conditions is also very important because the stability of the catalyst structure under OER conditions is crucial for commercial applications.

In the past, traditional electrochemical techniques such as Tafel slopes and electrochemical impedance spectroscopy (EIS) typically offer only indirect and partial information regarding reaction kinetics, making it difficult to deduce the mechanism of OER^[50]. Motivated by this challenge, the structural and valence states of the NiFe-based OECs have also been determined using different in-situ/operando techniques to study the OER mechanism, identify the actual active phase, and demonstrate the functional roles of Fe and Ni during OER^[51]. A variety of cutting-edge techniques including Raman spectroscopy, XAS, ambient pressure X-ray photoelectron spectroscopy (APXPS), Mössbauer spectroscopy, UV-visible (UV-vis) spectroscopy, DEMS, and surface interrogation scanning electrochemical microscopy (SI-SECM) were used to monitor the OER process in order to identify reaction intermediates and electrochemical activity on the surface of OECs during the OER^[52,53]. These in-situ/op-erando techniques can provide information about the structural and electronic states of NiFe-based electrodes. Figure 6 graphically shows several selected in-situ/operando spectroscopic characterization techniques which have been successfully applied to capture the NiFe-based OER electrocatalytic intermediate species, investigating the OER pathways and mechanism^[54].

Raman spectroscopy could be used to observe vibration, rotational, and other low frequency modes of the NiFe catalysts^[55]. The in-situ/operando Raman spe-ctroscopy could monitor the intermediates during the electrochemical systems under the applying test voltage in aqueous media, which could provide real-time reaction information. This technique helps to understand how an electrochemically driven reaction occurs. The in-situ/operando XAS could serve as a valuable technique for studying the electronic structure and local geometric structure of catalyst materials under working conditions. The XAS technique includes two regions: X-ray absorption near the edge structure (XANES) and extended X-ray absorption fine structure (EXAFS). Combing with the in-situ/operando XAS techniques, Friebel et al. found that the Ni_0.75Fe_0.25OOH catalyst has an α-Ni(OH)₂ structure at a low potential and a γ-NiOOH structure at a high potential. With the potential increasing, the shrinkages of the metal-oxygen and metal-metal distances indicate that in α-Ni(OH)₂ and γ-NiOOH, Fe has replaced Ni^[43]. Moreover, Görlin et al. found that the carbon-supported NiFe catalyst has a larger proportion of Ni⁴⁺ centers than the non-carbon-supported catalyst via the in-situ/operando XAS tool. They attributed it to the fact that the electrolyte can more easily contact more OER active centers^[38]. The XPS is served as a useful tool for obtaining surface element composition, and chemical oxidation and electronic state of materials. However, due to the need for ultra-high vacuum (UHV), many in-situ applications are subject to constrained conditions. APXPS based on synchrotron radiation can overcome the limitations of UHV conditions^[56]. Ali-Lcytty et al. conducted in-situ APXPS experiments on NiFe based catalysts under OER conditions. They found that when the electric potential is increased from 0 to 0.3 V (vs. Ag/AgCl), i.e., Ni(OH)₂ is oxidized to NiOOH, the O/OH ratio in-creases significantly. In addition, the Ni and Fe 2p XPS spectra also show the oxidations of Ni and Fe during OER process^[57]. The in-situ/operando UV-vis spectroscopy is used to track the metal oxidation process that occurs in Ni(OH)₂. Görlin et al. used in-situ/operando UV-vis spectroscopy to track the metal oxidation process in NiFe based electrocatalyst. It was found that Fe inhibited the oxidation form Ni²⁺ to Ni^3+/4+ during the OER process. Combing the in-situ/operando UV-vis spectroscopy, they concluded that OER was carried out at the center of Ni²⁺ in the NiFe-based catalyst^[37]. Moreover, Görlin et al. used in-situ/operando DEMS to track the Faraday charge distribution from the addition of the catalyst to the product formation and the redox process of the catalyst. The lower metal redox charge indicates that the average valence of Ni in the NiFe-based catalyst is lower than that of nickel oxide catalysts. They concluded that the Fe inhibits the oxidation of Ni during OER and seems to have a stabilizing effect on low valent Ni, thereby promoting OER^[37].

Among several other in-situ/operando techniques, ⁵⁷Fe Mössbauer spectroscopy is a strong method with very high energy resolution that uses the three main hyperfine interaction parameters of ⁵⁷Fe δ, Δ, and magnetic hyperfine field (B) to determine the oxidation state, electron spin configuration, symmetry, and magnetic information of Fe centers in materials^[58]. Therefore, our focus is not only to investigate the role of Fe in enhancing the OER activity of NiFe-based OECs but also stabilizing and confirming the structure of catalysts during OER through in-situ/operando⁵⁷Fe Mössbauer spectroscopy. We hope that one may better understand how Fe doping/incorporation improves the OER activity of Ni-based OECs and design a better catalyst system for OER in alkaline media, providing new solutions for the further deep study of NiFe-based OER catalysts towards practical applications.

Mössbauer spectroscopy has progressively become a standard technique for analyzing catalysis processes soon after Mössbauer effect was discovered^[59]. This is unique for revealing the “black box” of catalysis because of its in-situ/operando application capabilities^[60]. The main applications of Mössbauer spectroscopy in catalysis research are: 1) identification of active sites or active phases for catalytic processes; 2) investigation of correlations between catalyst structure and catalytic performance; and 3) characterization of catalysts during reaction activation and deactivation under ex-situ and/or in-situ/operando conditions. There are limited numbers of Mössbauer-active elements such as iron (⁵⁷Fe), tin (¹¹⁹Sn), antimony (¹²¹Sb), gold (¹⁹⁷Au), nickel (⁶³Ni), ruthenium (⁹⁹Ru), iridium (¹⁹¹Ir and ¹⁹³Ir), and neptunium (²³⁷Np) etc., which can be studied using Mössbauer spectroscopy. This limitation is due to several criteria such as suitable lifetime of nuclear excited state, transition energy, easy accessibility, and handling. Among these different elements, ⁵⁷Fe is the most studied and well-known Mössbauer nuclide.

Mössbauer spectroscopy, which involves the emission of γ-rays from radioactive nuclei and the absorption of these γ-rays by other nuclei of the same element, is a powerful technique for studying materials by measuring the hyperfine interactions that arise from the coupling of nuclear moments with electric and magnetic fields acting on the atomic nucleus. The uses of Mössbauer spectroscopy in catalysis and theoretical concepts have been described before in previous reviews where they mainly focused on physical concept of Mössbauer spectroscopy rather than its applications in catalysis^{[54, 61-62]}. Furthermore, it is yet unclear that which sector is best suited for a certain Fe-based catalyst, as well as the exact function of ⁵⁷Fe Mössbauer spectroscopy in various catalysis fields. As a result, in-situ/operando⁵⁷Fe Mössbauer spectroscopy is a valuable tool for determining the link between OER behavior and Fe structural characteristics in Fe-based catalysts during the OER process.

The Mössbauer effect is the absorption and emission of γ-rays from a recoilless nuclear resonance phenomenon, analogous to the acoustic resonance of two tuning forks with the same frequency f_s = f_r for sender (s) and receiver (r). A nucleus with energy E_e at an excited state (with Z proton and N neutron) undergoes transitions to the ground state with energy E_g by producing γ-rays of energy E_γ. The γ-rays may be absorbed by another nucleus of the same type (identical Z and N) in its ground state, resulting in a transition to the excited state of energy E_e. This phenomenon is called γ-rays recoilless resonance absorption as schematically shown in Figure 7(A). Resonance absorption can only be seen if the emission and absorption lines overlap enough. When γ-rays with energy E_γ are emitted or absorbed in freely moving atom (molecule) of mass m, it undergoes a recoil effect with energy E_R which is many orders of magnitude greater than the natural line-width (Г), hence resonance is not possible in a freely moving atom or molecule (gas, liquid). As a result, the Mössbauer eff-ect cannot be seen in the presence of freely moving atoms or molecules, such as those in a gaseous or liquid form. Recoilless emission and absorption of γ-rays are conceivable in the solid state, and basically un-shifted transition lines may (at least partly) overlap, resulting in nuclear resonance absorption.

German scientist, Rudolf Ludwig Mössbauer, while he was 27 years old in graduate school, proved this effect experimentally by freezing the source and absorber close to liquid nitrogen temperature, where atoms are firmly bonded in the lattice and the recoil effect is much reduced. Rudolf Ludwig Mössbauer was able to explain this phenomenon using quantum mechanics, demonstrating that atoms as lattice harmonic oscillators do not change their quantum mechanical state upon emission or absorption of γ-rays for a certain probability, which is relatively high for hard and low for soft materials (zero-phonon processes)^[59]. Nuclear resonance absorption of recoilless γ-rays, named as Mössbauer effect, soon after he was awarded the Nobel prize in physics 1961, may be detected only for this proportion, known as the Lamb-Mössbauer fraction^[63].

Figure 7(A) describes the overall Mössbauer effect phenomena in which the mean lifetime of the excited state of the parent nuclei is very important. For example, the first excited state of ⁵⁷Co has the mean lifetime of around 100 ns which is normally being used for ⁵⁷Fe Mössbauer experiment. As a result of the radioactive decays, the excited state of ⁵⁷Co decays to ⁵⁷Fe excited state by electron capture process. The excited state of Mössbauer nucleus decays and comes to ground state. The decay of ⁵⁷Co generates two excited states of ⁵⁷Fe where almost 99% decay occurs for the lowest energy level which is 14.4 keV and is known as the first excited state from ground state of ⁵⁷Fe. The ground state of ⁵⁷Fe does not undergo any kind of decay and the decay constant is zero for it due to zero energy level. Each nuclear level possesses specific energy, spin, parity, and decay constant k mean probability of decay per unit time.

As shown in Figure 7(B), due to the electrical mo-nopole interaction, the electrical quadrupole interaction, and the nuclear Zeeman interaction, the nuclear energy levels of ⁵⁷Fe (ground and first excited states with nuclear spins of 1/2 and 3/2, respectively) can be split by electric or magnetic fields acting at the nucleus. The hyperfine interactions are able to cause a positive or negative shift of the peak position from zero position of the relative velocity (Doppler velocity), and the shape of the ⁵⁷Fe Mössbauer spectrum can be changed from a single absorption peak to double and even sextet. In Figure 7(B) schematically shows the shift and splitting of nuclear energy levels and the corresponding ⁵⁷Fe Mössbauer spectral shape changes when the three kinds of hyperfine interactions are considered solely. In actual samples, the three kinds of hyperfine interactions are often existed simultaneously at iron nuclear positions, and the peak position of the resulting Mössbauer spectrum, the peak spacing and the shape of the Mössbauer spectrum will be more complex.

Usually there are three main parameters obtained from Mössbauer spectroscopy in data form by curve fitting measured Mössbauer spectra, which gives information about electronic, structural and magnetic properties of investigated material include: 1) Isomer shift (δ) arises from Coulombic interaction between the nucleus and the s-electrons, 2) quadrupole splitting (Δ) arises from the interaction of electric field gradient with the electric quadrupole moment of the nucleus, and 3) magnetic hyperfine field (B) arises from the interaction between the nuclear magnetic dipole moment and the effective magnetic field at the nucleus. That is, in turn, through the observation of the γ-rays recoilless nuclear resonance phenomenon, the Mössbauer effect, obtaining measured Mössbauer spectrum, analyzing it by considering the various hyperfine interactions existed in the nuclear location, we can study the configuration and distribution of orbital electrons, the physical properties of catalytic material, structural properties and chemical bonding properties, etc.

A standard Mössbauer spectrometer mainly consists of a radiation source (normally ⁵⁷Co in a Rh or Pd metal matrix for ⁵⁷Fe Mössbauer spectroscopy), a velocity transducer, a device for velocity calibration, waveform generator and synchronizer, γ-rays detection system, multichannel analyzer, computer, and an optionally cryostat (low temperature) or oven (high temperature) attached in the setup to control the temperature during measurement, which make this instrument suitable to operate at a wide range of temperature conditions. Schematic illustration of (A) configuring a Mössbauer spectrometer and (B) observing modes of a Mössbauer spectrum is shown in Figure 8. In transmission mode as shown in Figure 8(B), an absorber (sample under study) is placed between the source and detector to observe the negative peaks corresponding to absorption of 14.4 keV γ-rays. In backscattering mode, the source and absorber are placed on the same side to observe positive peaks corresponding to absorption of 14.4 keV γ-rays, and the conversion electron Mössbauer spectrum (CEMS) is produced by detecting only the emitted internal con-version electrons by using specific CEMS detector.

Mössbauer spectra are generated by the velocity sweep method in which a moving driver or velocity transducer is used to move the source or sample repeatedly (so-called Doppler moving) with a specific velocity, while γ-rays are transmitted or emitted continuously through the sample and counted on synchronization channels. Usually, the source is mounted on a vibrational axis of an electromagnetic transducer which moves according to voltage waveform applied on the driving coil of the systems, and acceleration with a triangular waveform frequency of this transducer is between 5 ~ 50 Hz. The detecting system is used to detect and store the signal counts in a memory of typically 1024 channels. An external clock subdivides the time intervals of waveform applied to the driven system into number of channels which synchronizes the channel number in the memory and instantaneous velocity of the source by advancing memory address one by one. The triangular waveform and 1024 memory channels produce two mirror-imaged Mössbauer spectra each with 512 channels. With sufficiently good velocity linearity, the two spectra can be easily combined (to increase the signal to noise ratio) to give one 512 channel Mössbauer spectrum.

After obtaining data from experiment, the next step is analysis and fitting of the spectra to determine the active phase structure and valence states of Fe in the catalyst. The fitting of data is usually approximated after comparison with previously available data in the literature or database system. The past data can help in the attribution and screening of the object phase, valence state, spin state and coordination structure based on the analytical and fitted results. Therefore, the database system was developed which is a globally connected network of different centers for Mössbauer effect reference and data. These centers collect data from the Mössbauer analysis of the published literature where it can be used for fitting and analysis of future materials. One of the biggest centers of Mössbauer effect, where the reference and data collected from the whole world for different Mössbauer samples, is Mössbauer Effect Data Center at Dalian Institute of Chemical Sciences, Chinese Academy of Sciences (https://medc.dicp.ac.cn/)^[64].

Furthermore, there are several sophisticated methods and software available for analyzing and interpretation of Mössbauer data. Among different types of software available in market, Mosswin is one of the most sophisticated software for data analysis and fitting (http://www.mosswinn.com/)^[65]. Fitting methods and model assumptions can have a significant impact on the interpretation of Mössbauer spectra under certain situations due to overlapping of data peaks usually. Therefore, members of a research group use one common software and analysis method for spectral analysis, and any discrepancy in interpretation due to the usage of different programs could be reduced. Mössbauer spectra have been interpreted using different physical models, but only a few of these models have been compared and published.

A variety of physical models are used by the analysis software to produce a model spectrum with which measured spectrum could be compared, fitted, and analyzed. The fitting of the Mössbauer spectrum should not only be carried out on the obtained data, but first also consider a theoretical model. This is because fitting could get superficially unreasonable chi-squared value ( χ²) if based on unphysical model. Commonly, there are three different line shapes employed for modeling of a Mössbauer spectrum as shown in Figure 9. The combination of Lorentzian and Gaussian line resultant is called as the Voigt line shape which is referred to as a pseudo-Voigt function since it is a linear mixture of the two shapes.

Error in analysis is obvious and it varies from laboratory to laboratory. Doublet areas are typically stated to no better than one significant place after the decimal. Isomer shift (δ) and quadrupole splitting (Δ) values are often ± 0.02 mm·s^-1, and magnetic hyperfine field (B) values are exceedingly erratic. In general, Mössbauer technique can detect down to 1% of the total normal Fe in the sample, but challenging for the samples containing less than 0.1% weight Fe, the Mössbauer signal is available to be largely enhanced if the enriched isotope ⁵⁷Fe is used for preparing the samples.

Taking an example of NiFe-(oxy)hydroxide OECs, generally two symmetrical peaks of almost equal height are expected for pre-OER powder-based catalyst, indicating the presence of only one kind of Fe³⁺ species. This kind of spectrum could be simply approximated and fitted with two Lorentzian lines, only one doublet confirms the one kind of Fe³⁺ species in the system. In our work recently published, all spectra of NiFe-(oxy)hydroxide OECs were fitted to Loren-rentzian profiles by the least-squares method, and the fit quality was controlled by the standard χ² and misfit tests (MossWinn program)^[66].

In this way, the ⁵⁷Fe Mössbauer spectral parameters could be determined, including the isomer shift (δ), the electric quadrupole splitting (Δ), the full width at half-maximum (Γ_exp), and the relative resonance area (A) of the different components of the absorption patterns. In the case of two different kinds of Fe³⁺ species in the system which could arise due to difference in coordination at the surface and inside the lattice, the fitting of the spectrum may not be simple and could require fitting with two doublets or one doublet and one singlet. The presence of high valence Fe such as Fe⁴⁺ could be confirmed if the spectra could not be fitted with only one doublet and an isomer shift is observed around zero and even negative side. Moreover, stability of catalyst structure can also be confirmed by comparing pre-OER and post-OER spectra by fitting equally. Any difference in the spectra of two states indicates that catalyst structure is not stable and phase changes might occur which could create ambiguity about the final results. Therefore, it is strongly recommended that in-situ/operando study is conducted for the catalyst which does not undergo any kind of structural transformation under applied potential or during OER.

Mössbauer spectroscopy becomes very important tools for the characterization of NiFe-based OECs since a very small amount of Fe significantly enhances the OER activity of these materials. Later in the 1980s, it was received a huge attention to the research on electrochemical water splitting processes utilizing non-precious metal oxides and hydroxides to replace the precious metal ones and costly anodic materials. Corrigan conducted one of the significant researches in which he discovered that when Fe impurities are injected into NiO anodes, the OER activity increases^[32]. Afterward, several other transition metal cations were investigated for their impact on OER performance by impregnating into NiO electrodes^[35]. But not a comparable OER-enhancing impact was observed from any other d-block element except Ce⁴⁺, which showed similar improvement in OER activity of NiO. Nonetheless, the degree of Ce⁴⁺-induced enhancement was near to that of Fe³⁺-induced enhancement. However, until 2014-2015, no significant effort was made to give a logical and acceptable explanation for this enhancement. And for more than two decades, these findings of Fe impurities-based enhanced OER performance of Ni-based materials were completely hidden.

Since Mössbauer study of NiFe-OECs provided some reasonable explanations to this unprecedented OER activity, it became crucial to develop and improve the in-situ/operando Mössbauer electrochemical setup and experimental understandings. Firstly, we discuss the key challenging issues and parameters to develop an in-situ/operando Mössbauer instrument for OER reaction to observe the evolution of the chemical state of the iron dynamically in the catalyst and structure of the catalyst during electrochemical reactions. The in-situ/operando setup consists of ⁵⁷Fe Mössbauer spectrometer connected with electrochemical station to conduct electrochemical reactions and Mössbauer analysis simultaneously. A typical in-situ/operando Mössbauer electrochemical setup, which was self-developed in our lab, is shown in Figure 10(D). This setup consists of three main parts: 1) Mössbauer spectrometer, 2) electrochemical station, and 3) locally designed in-situ/operando OER electroche-mical reaction cell. In our lab, we use single line ⁵⁷Fe Mössbauer radiation source ⁵⁷Co(Rh) with 14.4 keV level γ-rays, which could reduce attenuation in electrolyte and obtain a satisfactory signal-to-noise ratio attached with CHI660E electrochemical station. The Doppler velocity of the spectrometer is calibrated with respect to α-Fe at 298 K according to the international recommended standardization. Generally, before the in-situ/operando OER characterization experiment, the electrolyte is saturated with nitrogen or argon to remove the dissolved oxygen. Therefore, accessories such as a gas cylinder are added as shown in Figure 10(D).

We have self-developed several kinds of in-situ/operando electrochemical reaction cells for ⁵⁷Fe and ¹¹⁹Sn Mössbauer spectral measurements using one kind of commercialized Teflon material without containing any iron impurity, which is a cheap, stable, and easily machined material with very low γ-ray absorption property. An electrochemical reaction cell with high versatility for in-situ/operando Mössbauer spectral measurements is shown in Figure 11. Very thin windows of about ~ 0.6 mm thickness with the diameter of about ~ 10 mm aligned with the source and detector were carved into both walls of the working electrode area for the facile passage of γ-rays. The distance of the two windows was only a few millimeters to reduce the attenuation of γ-rays and enough to immerse 3 ~ 5 pieces of working electrodes parallel to each other with a distance of less than 1 mm. Two small peristaltic pumps, such as NKCP-S04B type model, can be connected with the reaction cell where the electrolyte enters from lower part of reference electrode and after passing through working electrode area leaves from upper side of counter electrode. This kind of flow style can help to maintain the pH during the electrochemical reaction. The flow rate of both pumps keeps same during the reaction and the amount of electrolyte in a beaker should be enough such as 100 ~ 200 mL to maintain the pH of 0.1 mol·L^-1 KOH solution.

The fabrication of electrode for in-situ/operando Mössbauer test is very sensitive task and needs high attention. For this purpose, the key step is to prepare ink of the catalyst. In a typical OER experiment in our lab, 10 mg of catalyst powder is added into solution containing 40 μL Nafion, 480 μL isopropanol, and 480 μL deionized water in a 5 mL plastic vial. The suspended solution of catalyst is kept in sonication for six hours to obtain a uniform slurry which is important for uniform and stable deposition of catalyst on substrate. The catalyst ink must be evenly coated on substrate such as carbon paper or carbon foil before the test and a conductive copper foil is connected with the substrate as shown in Figure 10(A). A front view of self-designed Mössbauer-electrochemical reaction cell connected with electrochemical station is shown in Figure 10(B). Figure 10(C) is a picture of the in-situ/operando Mössbauer-electrochemical cell placed inside the Mössbauer instrument ready for coinstantaneous OER reaction and Mössbauer measurement. A lead plate with 3 mm thickness, with a window of the same size as the in-situ/operando cell, is pressed to the cell’s front, ensuring that the Mössbauer γ-rays only can enter from the cell’s window, and the Mössbauer γ-rays in other directions are shielded by the lead plate. We call this specific lead plate a Mössbauer spectrometer collimator.

Generally, 5 ~ 10 mg Fe cm^-2 of absorber thickness is recommended in case of normal iron containing substance used in catalyst preparation for a conventional ⁵⁷Fe Mössbauer spectra measurement, in which the ⁵⁷Fe nucleus, one kind of the Fe isotopes with only ~ 2.2% natural abundance. For the in-situ/operando electrochemical ⁵⁷Fe Mössbauer testing, the catalyst is recommended to be prepared using enriched ⁵⁷Fe isotopes for enhancing the resonant absorption signals. Compared with the window material and the carbon paper substrate, the electrolyte solution is the greatest affecting the Mössbauer γ-rays, so, the distance between the two windows of the reaction cell should be as close as possible to be minimized. This is also the main reason for the recommended use of enriched ⁵⁷Fe isotopes for electrocatalyst preparation. Generally, the enriched ⁵⁷Fe isotopes can be purchased on the market, the purity is basically 95% and above, which is provided in the form of metallic iron foil or iron oxide. It can be dissolved in a medium such as hydrochloric acid, nitric acid or sulfuric acid, and prepared into the corresponding salts as stock solution for use.

It is well-known fact that only a small amount of catalyst is needed for one OER test, but for the in-situ/operando⁵⁷Fe Mössbauer studies enough Fe content in the catalyst is required to detect good quality signals. Depositing a too thick catalyst layer on one working electrode is not recommended as only surface particles will take part in OER while the bottom particles may not take part, hence signals could be mixture of unreacted and reacted Fe in OER catalyst. Therefore, to make it sure that every particle of the catalyst is taking part in OER, and enough amount of catalyst is presented to detect signals, we need several pieces of well-prepared working electrodes placed parallel to each other as shown in Figure 11(B) (3 ~ 5 pieces of electrodes deposited with the same catalyst ink placed at the distance of less than 1 mm). This will facilitate every particle of the catalyst to take part in the OER reaction and enough amount of Fe content to detect good quality signals.

On the other hand, we prefer a continuous flow of electrolyte rather than changing electrolyte after some time as using a small amount of electrolyte and highly active electrocatalyst may result in change of pH of the electrolyte which could alter the reaction dynamics. Hence, a continuous flow of electrolyte through reactor will prevent this kind of pH from changing during the reaction even at longer run. And at the same time, also enough amount of electrolyte must be presented inside the reaction cell during the flow process but only at counter and reference electrode sites, while the working electrode should have very less amount only to make sure electrolyte contact with the catalyst for reaction so that a good quality of signals could be obtained to study briefly. The formation of bubbles on the electrode surface is inevitable during OER especially at higher potential, which could affect the signal quality, so using multilayers of catalyst and exposure to limited amount of electrolyte could largely suppress this problem but not completely.

The electrochemical tests should be conducted in a systematic manner such as first pre-OER should be conducted for catalyst powder without substrate. Then after depositing the catalyst powder on the substrate to study the effect of substrate because if there is any effect we should consider during analysis and fitting process to normalize. After pre-OER tests, first in-situ/operando test should be conducted at open circuit potential (OCP) to study the effect of electrolyte. Usually, the absorption of γ-rays or strength of the signals could be decreased in the presence of electrolyte, but this problem could be lessened if experiment continues for longer time. Next, one test should be conducted at an applied potential lower than the potential required for OER, i.e., lower than the onset potential. This will show whether there is any change of electrochemical state of Fe species at lower potentials or not. After these initial experiments, now we can go for onset potentials and higher potentials. After studying at different potentials, one should test the post-OER immediately after taking out the electrolyte and removal of applied potentials, and then after 48 hours to 72 hours when the electrode is fully dried. This will give a clear understanding about the structure stability and time required for recovering the original structure capability of the catalyst. If the pre-OER and post-OER spectra of a catalyst are similar and show symmetric, then it could be considered that the catalyst has a stable structure and does not undergo any structural transformation, and the role of the chemical state change of Fe species could also be confirmed under OER catalyst working conditions. One most important point here is that the current density must be stable during whole reaction at every applied potential, which will not only confirm the catalyst stability but also help in determining the real structure of the catalyst.

As we have discussed above that presence of Fe⁴⁺ was first time observed by Stahl et al. but they could not completely explain the role of high valence state of Fe in the OER^[47]. And it was assumed that this high valence Fe either plays a direct role as an active site during OER or it enhances the activity of Ni sites. Later on, in-situ/operando XANES study of NiFe-PBAs indicated that NiFe-PBAs were transformed to NiOOH_2-x during OER, which generate Ni⁴⁺ hence the OER activity was improved. These contradictory and incomplete claims indicated two major flaws in these studies: 1) the real role of Fe⁴⁺ could not be clearly understood and explained and 2) the reversible structural transformation of NiFe-OECs during the OER, making it unclear that which phase of NiFe-OECs is responsible for higher OER activity. Therefore, we tried to solve both these challenging issues and address well in our following studies.

We have so far discussed the effects of doped/in-corporated Fe in Ni-based OECs, their activities, possible factors responsible for higher activity. But it is important to highlight the suitable method for their synthesis and in-situ/operando studies. We here recommend a suitable and facile method for the synthesis of NiFe-(oxy)hydroxide OECs by using PBAs as pre-cursors, which was newly developed in our lab^[67]. We know that PBAs have the typical chemical formula A_lM_n[Mm(CN)₆)]·xH₂O, where A indicates the alkali metal ions such as Na⁺ and K⁺ etc., while M and M represent the transition metal cations^[68,69]. In the PBAs crystal structures, cyanide groups serve as bridges between transition metal ions (M²⁺-C≡N-M³⁺). Due to the difference in electronegativities of C and N, the corresponding M and M metals show high and low spin states, respectively. These materials are useful for different applications such as batteries, gas storage, catalysis, energy storage, sensors, drug delivery, and environmental cleaning^[69]. One of the most pro-mising applications of these materials is the use of these materials as a precursor for the synthesis of inorganic nanostructures^[70]. It is now possible to synthesize volumes of PBAs with specified compositions and morphologies. PBAs have also gained attention due to diverse and tunable morphologies and sizes. But effective and beneficial use of PBAs mainly depends on processing and synthesis steps to control shape, composition, and size according to the desired application.

In first part of our case study here, the structural stability issue of NiFe-OEC was investigated and the real phase of the catalyst was confirmed during OER before going for in-situ/operando Mössbauer study to further elucidate and understand the mechanism. For this purpose, composites of NiFe-PBAs/carbon nano-tubes (CNTs) with the optimized amount of CNTs were synthesized and their OER activities were compared with that of NiFe@C obtained from high-temperature pyrolization of NiFe-PBAs. The NiFe-PBAs/CNTs OER activity with the optimum amount of 8wt.% of CNTs was found to be significantly superior to NiFe@C as shown in Figure 12(A), and the Tafel slope showed the smallest value of 83 mV·dec^-1 among other samples as shown in Figure 12(B). The X-ray diffraction (XRD) results before and after de-position and activation by CV of NiFe-PBA/CNTs (8wt.%) on carbon paper indicated that the crystal structure was changed after activation as shown in Figure 12(C). The XRD peaks intensities were significantly reduced after activation, which suggests that the crystal structure of NiFe-PBAs was broken during the activation process. The characteristic peaks representing NiFe-PBAs completely disappeared after 125 cycles, indicating that the NiFe-PBA structure had been completely transformed into other structures.

The XRD results after the 500 cycles were similar to 125 cycles, which suggests that there is no further change in crystal structure after 125 cycles. Furthermore, ex-situ⁵⁷Fe Mössbauer spectral measurement shown in Figure 12(D) before and after activation of NiFe-PBAs showed that after activation no resonance absorption signal could be observed, which proved that the iron cyanide large anions had been completely dissolved into the electrolyte during the reaction and the structure no longer belonged to NiFe-PBAs. This was further confirmed by ex-situ and in-situ/operando Raman spectra as shown in Figure 12(E-F), which exhibited the same results. Hence, the dissociation of [Fe(CN)₆]^3- after activation indicated the formation of new Ni-O stretching vibration band which belongs to Ni(OH)₂/NiOOH. Therefore, this study revealed that NiFe-PBAs/CNTs OER catalysts will com-pletely undergo dissociation of [Fe(CN)₆]^3- upon electrochemical activation in the alkaline condition generating Ni(OH)₂/NiOOH. In this stage, this Ni(OH)₂/NiOOH can be considered to the real active phase, which is more reasonable to claim that why NiFe-PBAs has high activity in the OER.

In the first part of our case study above, we confirmed that the spontaneous dissociation of the NiFe-PBAs in the alkaline solution generated Ni(OH)₂/NiOOH, serving as potential and stable structures which do not undergo any kind of further transformation under the applied potential once it is generated. In addition, we previously developed a novel and scalable strategy, containing a “copolymer-co-morphology” conception, to compositionally control synthesis of various types of PBAs which can be topotactically derived into corresponding double oxides^[68]. Therefore, a series of NiFe_m-(oxy)hydroxides (NiFe_m-O_xH_y, m refers to the atomic ratio of Fe/Ni) catalysts were synthesized by the topotactic transformation of NiFe_m-Fe PBAs in an alkaline solution. The NiFe_mFe PBAs materials with different iron contents doped in the Ni sites were prepared by our previously developed strategy^[71].

One of the novelties in our second part case study here is the topotactic transformation of NiFe_m-Fe PBAs into NiFe_m-O_xH_y. It was observed from the ex-situ⁵⁷Fe Mössbauer analysis of NiFe_m-Fe PBAs that there are two kinds of Fe³⁺ with different spin states in the system. This indicates that two different Fe³⁺ species are on two different crystallographic positions and all the spectra could be reasonably fitted with two doubles as shown in Figure 13(A-D). The relevant amount of different kinds of Fe³⁺ in every sample containing different amounts of Ni/Fe ratio is given in Table 1. The sensitivity of ⁵⁷Fe Mössbauer spectroscopy is so high that it could even differentiate between two different crystallographic positions of Fe³⁺ which is usually not possible with other characterization techniques. This makes ⁵⁷Fe Mössbauer spectroscopy more sophisticated and sensitive to study Fe containing materials. Moreover, increasing the Fe ratio increases the Fe³⁺ amount in the Ni sites, which, later on, will be transformed into NiFe_m-O_xH_y after treatment with KOH. Interestingly, only one type of high spin Fe³⁺ was observed in ex-situ⁵⁷Fe Mössbauer analysis for NiFe_m-O_xH_y, and increasing Fe doping without further influence on the one doublet spectral shape and parameters as shown in Figure 13(E-H) and also listed in Table 2.

Figure 14(A) shows that the NiFe_0.2-O_xH_y exhibited a significant decrease in onset potential after activation through 100 cycles of CV where phase changes occurred from α-phase Ni(OH)₂ structure to γ-phase NiOOH structure. Ex-situ⁵⁷Fe Mössbauer analysis shows a negligible change in the δ values, but there were significant changes in the Δ values before and after electrochemical activation as shown in Figure 14(B-C). Iron has the coordination structure of 6-oxygen coordinated octahedral in both crystal phases of α-phase Ni(OH)₂ and γ-phase NiOOH, with smaller Δ value indicating the octahedral symmetry should be higher in later one than the former one. This also further illustrates the high energy resolution of the ⁵⁷Fe Mössbauer spectroscopic technique where even subtle changes in the electron density and the electric field distribution at the iron nuclide position can be clearly identified. Similar trend was also observed in ex-situ and in-situ/operando Raman spectra where peak shifting and significant rise in peaks intensity were observed after activation as shown in Figure 14(D-E). The LSV results for different NiFe_m-O_xH_y catalysts shown in Figure 14(F-G) indicate that NiFe_0.2-O_xH_y exhibited the lowest overpotential of 263 mV at the current density of 10 mA·cm^-2 and the Tafel slope with a value of 35 mV·dec^-1 was observed (Figure 14(H)). The catalyst NiFe_0.2-O_xH_y showed a great stability for 100 hours while exhibiting the current density of 100 mA·cm^-2.

In summary, in the second part of our case study work, the high-performance OER catalyst of NiFe-(oxy)hydroxides (NiFe_m-O_xH_y) was firstly developed through the topotactic transformation of NiFe_m-Fe PBAs in an alkaline solution. Furthermore, the phase purity and stability were thoroughly confirmed by ex-situ and in-situ/operando Raman and ex-situ⁵⁷Fe Mössbauer spectroscopies in combination with other several kinds of conventional techniques. It is indicated that phase structure of NiFe_0.2-O_xH_y was irreversibly transformed from α-Ni(OH)₂ to γ-NiOOH by applying an anodic potential which is different from that of previous report mainly obtaining from in-situ/operando XAS characterizations^[49]. The amount of Fe was optimized, and the excellent OER electrochemical activity and stability were also demonstrated. That is to say, these catalysts are confirmed to be more stable with no changes into any other structure once it is generated after CV activation.

After making clear the stability issues, the NiFe_0.2-O_xH_y catalyst with the optimized iron proportion was selected to probe the reaction mechanism of NiFe_m-O_xH_y OER electrocatalyst using the self-developed in-situ/operando electrochemistry Mössbauer characterization instrument as shown in Figure 10(D)^[66]. This in-strument has also been applied in ORR^[72] and CO₂RR^[73], which provides a high level of research tool for the preparation of highly efficient catalysts of electrolytic water, carbon dioxide reduction, fuel cell development and even for detecting the structural and electronic properties of single atom catalysts^[72,73]. The id-entifications of durable and non-durable FeN_x sites in Fe-N-C materials for PEMFCs towards practical applications have also been successfully investigated, and the results will be published soon as like that reported by the well-known French team^[74].

The in-situ/operando⁵⁷Fe Mössbauer results indicated that Fe was only in 3+ oxidation state in NiFe_0.2-O_xH_y at lower applied potentials, such as 1.22 V and 1.32 V, and the spectral shapes are almost the same as the one collected at the open circuit voltage (OCV) as shown in Figure 15(A), where only one doublet is observed with the equal intensity of the two peaks and can be ascribed to high spin Fe³⁺. However, as the applied potential increased and reached to 1.37 V, the intensity of the two peaks began to become unequal, and a small shoulder peak started to develop with the δ value of around -0.25 mm·s^-1 as shown in Figure 15(B) which can be attributed to Fe^4+[47]. The observed spectral shape has been changed more significantly, indicating that the content of Fe⁴⁺ further increases as the applied anode potential increases (Figure 15(C-D)). The amount of Fe⁴⁺ further increased to ~ 40% as the applied potential reached 1.57 V as shown in Figure 15(E). And after that Fe⁴⁺ completely disappeared as the applied potential was removed, indicating the crucial role of Fe⁴⁺ species in OER (Figure 15(F)). The pre-OER and post-OER spectra were similar, which indicates the excellent structure stability of the catalyst. The stability was also confirmed through measurement of current density which remained stable throughout the in-situ/operando Mössbauer measurements (Figure 15(G)). Moreover, combining with the cyclic voltammetric curves (Figure 15(H)), the generation of Fe⁴⁺ (~ 12% in total) was also observed at around the onset potential 1.42 V vs. RHE of OER (Figure 15(C)). As listed in Table 3, it is different from the previous in-situ/operando Mössbauer tests, in which Fe⁴⁺ was observed at a higher potential rather than around onset potential^[47]. Therefore, the in-situ produced abundant Fe⁴⁺ at onset potential and different applied potentials suggested that Fe⁴⁺ has critical role in OER as shown in the last Figure 15(I). Therefore, it can be concluded that the amount of high-valent iron species in-situ produced in the NiFe_0.2-(oxy)hydroxide has a positive correlation with its water oxidation reaction performance, which further deepens the understanding in the mechanism of NiFe-based electrocatalysts.

In this tutorial review, NiFe_m-Fe PBAs are used as probe precursors for preparing highly efficient NiFe-(oxy)hydroxide (NiFe_m-O_xH_y) OER electrocatalysts by a novel topotactic transformation method and the home-made in-situ/operando electrochemical ⁵⁷Fe Mö-ssbauer spectroscopic technology independently developed is used to conduct in-depth research on the OER real active intermediates and working mechanism. Here we introduced in detail the application of in-situ/operando⁵⁷Fe Mössbauer technique in the electrochemical OER test process, including its working principle, instrumentation, the design of reaction cell, NiFe-based OECs sample detection, data analysis, simulation and study case results discussion. The in-situ/operando⁵⁷Fe Mössbauer technique has also been successfully applied in the ORR, CO₂RR, etc., the authors hope to be able to provide a high level of research tool for the preparation of highly efficient catalysts of electrolytic water, carbon dioxide reduction, fuel cell development and even for detecting the structural and electronic properties of single atom model catalysts.

The low cost, excellent OER activity, and better stability of NiFe-based catalysts make them ideal candidates. In order to develop novel NiFe-based OECs and their use in commercial applications, firstly, it is crucial to understand the OER process of NiFe-based OECs under realistic working conditions and to investigate their structure-activity relationship. Recently, several in-situ/operando characterization techniques have been developed to study the OER process at the solid/liquid interfaces under real-time applied potential to understand how the chemical and electronic structure changes take place in NiFe-based catalysts, the effect of electrochemical environment, and reaction intermediates interacting with each other. The focus of all these in-situ/operando techniques was to investigate the transformation of Ni(OH)₂ to NiOOH and the changes in oxidation states of Ni and Fe. And most of the results indicated that NiOOH is an active phase for OER, while both Ni and Fe are important for high OER activity in NiFe-based OECs. But all these in-situ/operando characterizations could not bring researchers to reach a consensus whether Fe is an active site or Ni as some in-situ/operando results indicated that Fe acts as an active site for OER, while NiOOH helps in the stabilization of Fe^4+[75]. On the other hand, some others suggested that Ni is the active OER center as α-Ni(OH)₂ nanostructure showed better OER activity compared to RuO₂^[76]. Moreover, the optimal interaction strength of Ni with OH_ad also satisfies the Sabatier criteria for the required design of OER catalysts^[29]. There could be several factors that are con-sidered as the potential factors for different active sites during in-situ/operando characterizations. For example, during in-situ/operando studies, it becomes difficult to find out real active sites in the presences of several other disturbing factors while focusing on selective intermediates. Another factor is the different configurations of in-situ/operando cells for the study of materials under in-situ/operando conditions. As per the current status of in-situ/operando characterizations analysis, not any single technique has shown comprehensively the ability to study the phase transformation, changes in valence state and variation in morphology under OER.

Mössbauer spectroscopy is a powerful technique to study the local electronic structure, local coordination, bonding, and oxidation states of the element under an investigation in nanostructure materials. Moreover, it can help in the identification of the several in-situ/operando redox processes characterizing the intermediate phases which lead to understand the possible catalytic reaction mechanisms and pathways. Recently, Mössbauer spectroscopy has gained huge attention in electrocatalysis where it has been found more useful for the detection of primary active sites in different Fe-based materials. A deep understanding of probing the local electronic structure, identifying the active sites and phases, determining the crystal structure, and tracking the oxidation state and environment change of various Fe-based catalysts is critical for further development in catalysis fields. For instance, the active sites of NiFe-based OECs could be clearly probed with the development of quasi-in-situ cells to in-situ/operando analyze the properties of catalysts. Further development of high temperature in-situ/operando or quasi in-situ cells would greatly promote the investigation in the phase transformation of Fe-based materials.

To address key problems, more comprehensive in-situ/operando techniques are highly desired to determine the active sites accurately and the variation in morphologies of NiFe-based catalysts under OER conditions, and those techniques influencing the interfacial interaction between catalysts and electrolyte/electrodes are crucial to be investigated simultaneously. For example, transmission and scanning electron microscopies may detect surface reconstruction and interface structure if they can overcome the difficulties of utilizing liquid electrolytes in a high vacuum environment. Although NiFe-based electrocatalysts have shown excellent OER activity, but lack of long-term stability prevents them from use in commercial applications. Therefore, understanding about degradation mechanism of catalysts, which includes variations in morphology, composition, crystal structure, and the number of active sites, is very important to design stable and highly active materials. This demands a precise and set of combining multiple complementary in-situ/operando characterization techniques to detect all the above-mentioned changes simultaneously. Hence, it will be beneficial to understand the activity-structure relationship in different catalysts by combining different in-situ/operando characterization techniques and theoretical calculations to determine the active sites for OER. This will help in understanding the OER mechanisms comprehensively, and will guide the design of the robust and efficient catalysts for OER to commercial applications.

This work was financially supported by the National Natural Science Foundation of China (No.: 21961142006) and the International Partnership Program of Chinese Academy of Sciences (No.: 121421KYSB20170020).