Oxygen-evolution reaction by nickel/nickel oxide interface in the presence of ferrate(VI)

In this study, we investigate the effect of K2FeO4, as a new and soluble Fe salt at alkaline conditions, on oxygen-evolution reaction (OER) of Ni oxide. Both oxidation and reduction peaks for Ni in the presence and absence of Fe are linearly changed by (scan rate)1/2. Immediately after the interaction of [FeO4]2- with the surface of the electrode, a significant increase in OER is observed. This could be indicative of the fact that either the [FeO4]2- on the surface of Ni oxide is directly involved in OER, or, it is important to activate Ni oxide toward OER. Due to the change in the Ni(II)/(III) peak, it is hypothesized that Fe impurity in KOH or electrochemical cell has different effects at the potential range. At low potential, [FeO4]2− is reduced on the surface of the electrode, and thus, is significantly adsorbed on the electrode. Finally, oxygen-evolution measurements of K2FeO4 and Ni2O3 are investigated under chemical conditions. K2FeO4 is not stable in the presence of Ni(II) oxide, and OER is observed in a KOH solution (pH ≈ 13).

www.nature.com/scientificreports www.nature.com/scientificreports/ activity 20 . In 2018, Chorkendorff 's group electrochemically oxidized the nanoparticles of metallic Ni-Fe alloy in an electrolyte containing H 2 16 O to form NiFe 16 O x H y ; in the next step, they added H 2 18 O to the electrolyte. No participation of lattice oxygen in NiFe 16 O x H y was observed 21,22 .
Using 18 O-labeling experiments in combination with in situ Raman spectroscopy, it was found that lattice oxygen is involved in OER by Ni and Ni/Co LDHs, but not when Ni-Fe and Ni/Co/Fe LDHs are involved 23 .
The roles of dissolved Ni species in OER was reported by He's group 24 . Other elements do not considerably affect OER of Ni oxide. However, some groups reported that Fe ions in electrolyte react with the surface of Co, Ni, Cu, Ag, Au, but not with Ti electrodes, and form the active sites for OER 12,14 .
It was reported that nickel(II) and cobalt(II) have a significant effect on the decomposition of [FeO 4 ] 2− 25 .
Herein, OER by Ni oxide in the presence of [FeO 4 ] 2− is reported. In addition to the electrochemical experiments, OER by [FeO 4 ] 2− and Ni 2 O 3 was also investigated, and the results are reported. These findings showed that in addition to the hypothesis of the effect of Fe on the oxygen-evolving activity of Ni oxide, the effect of Ni ions on the oxygen-evolving activity of high-valent Fe active sites should also be investigated.

Results
In this study, [FeO 4 ] 2− was used as an Fe salt, and its effect on Ni oxide was investigated during OER. Using [FeO 4 ] 2− as an Fe salt has some advantages: (i) It is a new Fe salt for investigating the effect of Fe on OER by nickel oxide.
(ii) In contrast to many Fe salts, [FeO 4 ] 2− is soluble under alkaline conditions. Thus, it could be easily added to nickel oxide under the potential to allow direct observation of the effect of Fe on OER of nickel oxide. (iii) In contrast to Fe(III), [FeO 4 ] 2− has a negative charge, and its migration toward the anode is easier than the migration of Fe(OH 2 ) 6 3+ with the positive charge. (iv) In the absence of the potential or a reductant, it has relatively good stability at high pH (>12).
(v) It shows a sharp peak in the visible spectra (violet color), and thus, its concentration could be detectable.  27 .
In this study, a nickel foam was used as the source of nickel oxide. First, a Ni foam was placed under oxidation conditions at 10.0 V for 20 minutes in a two-electrode setup in a KOH solution (pH ≈ 13). Then it was used for the electrochemical investigation. Because soluble nickel compounds may affect the electrochemistry of the electrode 24 , no nickel oxide was used to purify 12 the electrolyte in our experiments.
LSV of the Ni foam in KOH (pH ≈ 13) showed that before adding [FeO 4 ] 2− , the onset of OER is observed at 1.52 V (in the paper, all potentials are reported vs. RHE) (Fig. 1a). Ni(II)/(III) oxidation is also revealed at 1.43 V (Eq. 1): The consecutive CVs are interesting in that they show the effect of the impurity of Fe and the addition of [FeO 4 ] 2− . As shown in Fig. 1a, in the absence of the [FeO 4 ] 2and the presence of the impurity of Fe impurity in KOH or electrochemical cell, an increase in OER (orange arrow) is observed during the consecutive CVs. The Ni oxidation peak shifts from 1.43 to 1.46 V during the consecutive CVs (Fig. 1a). After adding the Fe salt (the final concentration of Fe salt was 0.25 mM. If another concentration of Fe salt was used, it was written in the related captions), a significant change is observed in OER (pink arrow), and the Ni oxidation peak is observed at 1.50 V (Fig. 1a).
The same CV is observed for a pure metallic Ni foil (99.99% purity; Fig. 1b). For a pure metallic Ni foil, the first no peak is observed for Ni(II)/(III) oxidation, since no pre-treatment at high potential was performed for the foil. However, after a few consecutive CVs, the peak at 1.46 V was observed, which shows a shift to the higher potential at 100 th CV (1.48 V) (Fig. 1b). To find the effect of the amounts of Ni oxide on OER, a Ni foam without any anodization at 10.0 V was also investigated, which showed similar characteristics to the anodized Ni foil (Fig. S1).
After 100 consecutive CVs at a scan rate of 25 mV/s in the absence of adding Fe ions, the onset of OER and the peak are observed at 1.48 and 1.47 V by LSV, respectively (Fig. 1c). The changes correspond to the incorporation of Fe impurity in KOH or electrochemical cell to Ni oxide. After adding Fe salt, the onset of OER was observed at 1.39 V. The related peak to Ni(II)/(III) oxidation is observed at 1.50 V. The peak shape is significantly changed in the presence of adding Fe ions.
SWV indicated the peaks for Ni(II)/(III) oxidation before and after adding the Fe salt. Before adding the Fe salt in KOH (pH ≈ 13), a peak at 1.37 V was observed (Fig. 1d), which corresponded to Ni(II)/Ni(III) oxidation; after 50 consecutive CVs, this peak shifted to 1.39 V, which was related to the effect of impurity in KOH or electrochemical cell. After adding Fe, the peak was observed at 1.43 V (Fig. 1d). The multi-step amperometry at 1.36-2.16 V before adding the Fe salt (blue) and after adding the Fe salt (red) is shown in Fig. 1e. Before adding the Fe salt at 1.96 V, 5 mA increase in OER was observed, which was related to impurity in KOH or electrochemical cell (pink line). (2020) 10:8757 | https://doi.org/10.1038/s41598-020-65674-x www.nature.com/scientificreports www.nature.com/scientificreports/ In the next step, the effect of different scan rates on OER, Ni oxidation and reduction peaks were investigated (Fig. 2). In all recorded scan rates, the Ni oxidation/reduction peaks in the absence and presence of Fe were observed. Both oxidation and reduction peaks for Ni in the presence and absence of Fe shifted linearly by (scan www.nature.com/scientificreports www.nature.com/scientificreports/ rate) 1/2 ( Fig. 2i), which showed freely diffusing species were important to form these peaks. The oxidation peaks shifted to higher potentials, and reduction peaks shifted to lower potential in the presence of Fe ions.
In the next step, the Fe salt was added to the electrolyte in the form of solid powder, and CVs were recorded ( Fig. 3a,b). The salt slowly diffused toward the electrode, and an improvement in OER was observed. Interestingly, immediately after the interaction of the Fe salt with the electrode, a significant increase in OER was observed (see the video, SI). This could show that Fe ions on the surface of Ni oxide were directly involved in OER or were important to the activation of Ni oxide toward OERs. After 80 consecutive CVs, a constant current was observed for OER (Fig. 3a,b), which showed that a few Fe sites were enough to optimize OER by Ni oxide.
The effect of the step by step addition of Fe ions to Ni oxide was studied. To inhibit the effect of the Fe salt on the cathode, the experiment was performed in a two-cell setup. As shown in Fig. 3c, at 1.46 V in KOH (pH ≈ 13), adding the Fe salt in the first step had a significant effect. However, adding more Fe salt had less effect on OER. If Ni sites are the active sites for OER, a decrease in OER may be observed after significantly covering the Ni sites by Fe ions. Surprisingly, however after adding more Fe salt, OER was increased. Thus, it is hypothesized that Fe sites could be active sites for OER.
The effect of large amounts of the Fe salt (12.6 mM) on OER of the Ni foam was investigated (Fig. 3d,e). Adding large amounts of Fe salt had no significant effect on OER compared to a small amount of Fe salt. After 100 consecutive CVs, a constant current was observed for the electrode in the presence of the Fe salt (Fig. 3d,e).
An interesting idea is to analyze the effect of the Fe salt on the CV of Ni oxide. The subtraction (Fig. 4a, orange plot) of the CV in the absence of the Fe salt and the CV in the presence of the Fe salt is shown in Fig. 4b. It seems that the redox property of Ni(III) was used by Fe to increase in OER (B area in Fig. 4b). In other words, it is hypothesized that Ni sites, which remain at high oxidation states before the addition of Fe, are pushed to be discharged by Fe ions toward OER in the presence of Fe (B area in Fig. 4b). www.nature.com/scientificreports www.nature.com/scientificreports/ Graphite electrode has no synergistic interaction with the Fe salt and the subtraction of the CV for graphite (green plot) in the absence and presence of Fe shows the effect of Fe on the electrode toward OER (Fig. 5c). It is interesting that the green and orange plots are comparable (Fig. 4b,c), if we withdraw the Ni(II)/(III) oxidation/ reduction peaks. The orange plot is just bigger, and its onset of OER is observed at a low overpotential. The similarity between orange and green plots could show that in Fe-Ni oxide, similar to the case where Fe is on graphite, the Fe ions are active sites for OER.
If we subtract the pure effect of Fe (green plot) from the orange plot in Fig. 4d, the synergistic effect of Fe on nickel could be calculated (black plot). As shown in Fig. 4d, it is hypothesized that Ni(II)/(III) oxidation occurs before OER in pure Ni oxide and the absence of Fe. However, after adding the Fe salt, Ni(II)/(III) oxidation occurs in the same potential as the onset of OER. The onset of OER is observed at 100 mV lower by Fe-Ni oxide than Fe ions on graphite (Fig. 5d).
In the next study, the effect of the potential range on the effect of Fe on OER by Ni oxide was investigated (Fig. 5). Before adding the Fe salt at (1.56-1.86 V) range (Fig. 5a), the consecutive CVs showed no clear increase in OER. After adding the Fe salt at this range (Fig. 5a), an increase in OER (1.2 times) was observed, which was constant during the time of the reaction.  (Fig. 5b). After adding the Fe salt at this range, a few increases (1.1 times) in OER was observed, which was constant during the time of reaction (Fig. 5b).
Before adding the Fe salt at (0.86-1.86 V) range (Fig. 5c), the consecutive CVs showed an increase in OER (1.5 times) at the 150 th CV. After adding the Fe salt at this range, a few increases (1.3 times) in OER was observed, which was constant during the time of reaction (Fig. 5c).
Before adding the Fe salt at (−0.14-1.86 V) range (Fig. 5d), the consecutive CVs showed an increase in OER (1.75 times) at the 150 th CV. After adding the Fe salt at this range, a few increases (1.15 times) in OER was observed, which was constant during the time of reaction (Fig. 5d).
Before adding the Fe salt, it seems that the activation factor needed a potential of less than 1.56 V. Given the change in the Ni(II)/(III) peak, it is hypothesized that impurity in KOH or electrochemical cell is the active factor. At a low potential, Fe ions are reduced on the surface of the electrode, and thus, are significantly adsorbed on the electrode. Thus, after the 150 th CV, a high activity for Ni oxide is observed. After adding Fe impurity on the surface of the electrode at a low potential, an Fe saturation occurs for Fe ion on the surface of the electrode, and the current remains constant. The related peak to Fe(II)/Fe(III) oxidation/reduction in Ni-Fe oxide is displayed at 0.36 V/ −0.06 V, respectively (Eq. 2): SEM images show that a fresh Ni foam has a cellular structure with high porosity and a large volume fraction of pores (Figs. 6a,b, S2). After operating the Ni foam at 10.0 V, a few corrosive points (ca. 10 nm) were observed on the surface of the cellular structure of the foam (Figs. 6c,d, Fig. S3 and S4). Small particles (20-40 nm) were also observed on the surface of the Ni foam (Fig. 6c,d). Under the operation at 10.0 V, no detectable iron was observed by EDX-SEM on the surface of the foam.
X-ray photoelectron spectroscopy (XPS) was used to characterize the operated foam at 10.0 V. Ni 2p peak showed significant split spin-orbit components (17.8 eV). Ni 2p3/2 spectrum of the operated foam at 10.0 V indicated a peak at 855.3 eV, which is related to Ni hydroxide (Fig. S5a) 28 . Ni 2p1/2 was also observed at 873 eV. Satellite features for Ni 2p3/2 and Ni 2p1/2 were displayed at 861 and 879 eV, respectively.
In addition, O 1 s component after the operation showed a peak at 531.1 attributed to oxygen in Ni hydroxide, OH, and OH 2 groups on the surface of the electrode (Fig. S5b) 29 .  (Fig. 7). The HRTEM images also indicated a crystal lattice spacing of 0.21-0.23 nm, corresponding to the (101) plane of Ni(II) hydroxide (ref.: 00,001). (111) plane was also observed in the XRD pattern (Fig. S6).  XRD showed no patterns related to Fe compound on the surface of the Ni foam after amperometry at 1.46 V in KOH (pH ≈ 13) in a two-compartment electrochemical cell for 2 hours (Fig. S9). However, the patterns for the separated iron compounds indicated Fe(OH) 3 (Fig. S10)  . Only a trace amount of Fe could significantly activate Ni oxide for OER (Fig. 1). Thus, it is concluded that the separated phases such as Fe(OH) 3 and FeOOH are not necessary for oxidizing water.
Oxygen-evolving catalysts are usually powders placed onto conductive substrates by binders. These binders decrease the contact between the electrolytes, electrode, and the catalyst and results in a decrease in the electric conductivity. The stability of these electrodes also is low because of the catalyst separation from the electrodes, especially under high current densities and harsh oxygen evolution. Using adding Fe salt to Ni oxide/Ni interface needs no binder and related limitations for the binder. The answers to these questions are critical for understanding the mechanism of OER by Fe-Ni oxide. After the deaeration of a KOH solution (1.0 mM), Ni 2 O 3 was added, but no clear oxygen evolution was observed (Fig. 8).
In addition to this reaction, as shown by the electrochemical method, Ni(III) oxide in the absence of Fe ions or potential could not oxidize water. It was demonstrated that Ni(III) in the β-NiOOH-like phase has no activity. Though, it was reported that after the oxidation of Ni(III) to Ni(IV), the activity toward OER would be observed 30 .
In the next step, after the deaeration of Ni 2 O 3 in KOH (1.0 mM), the Fe(III) ions was added. Again no detectable OER was observed. This shows that Fe(III) ions cannot activate Ni 2 O 3 for OER. Regarding the Pourbaix diagram, it is also unlikely that Ni 2 O 3 could oxidize Fe(III) ions (Fig. 8).
In contrast to Ni 2 O 3 , K 2 FeO 4 is not stable, and OER is observed in the presence of this compound in a KOH solution (1.0 mM) (Fig. 8). Adding Ni(II) ions or nickel hydroxide significantly increased OER of a K 2 FeO 4 solution in a KOH solution (1.0 mM). It displays that OER could occur by the mechanism of activation of K 2 FeO 4 using Ni ions (Fig. 8c) 31 .
Another mechanism could be the formation of an Fe-Ni oxide by incorporation of Fe ions in Ni oxide. This Fe-Ni oxide could be oxidized by K 2 FeO 4 in the next state and can oxidize water in the next step. To recognize the activation of K 2 FeO 4 toward OER on the surface of Ni oxide, a saturated Ni(II) solution was prepared, and the decomposition of K 2 FeO 4 was investigated in the absence of Ni oxide, but in the presence of trace but soluble Ni(II). In the presence of a K 2 FeO 4 solution, adding a trace amount of Ni(II) ions causes no increase in OER.  In the absence of the Fe salt and the presence of the impurity in KOH or electrochemical cell, an increase in OER was displayed during the consecutive CVs. The Ni oxidation peak shifted from 1.43 V to 1.46 V. After adding the Fe salt, a significant change is observed in OER, and the Ni oxidation peak is observed at 1.50 V.
SWV indicated the clear peaks for Ni(II)/(III) oxidation before and after adding the Fe salt. In the absence of the Fe salt in KOH (pH ≈ 13), a peak at 1.37 V was observed, which corresponded to Ni(II)/Ni(III) oxidation. After 50 consecutive CVs, the peak shifted to 1.39 V, which was attributable to the effect of Fe impurity. After adding the Fe salt, the peak was observed at 1.42 V. The effect of different scan rates on OER, Ni oxidation and reduction peaks were also studied. Both oxidation and reduction peaks for Ni in the presence and absence of Fe changed linearly by (scan rate) 1/2 . The oxidation peaks shifted to higher potentials, and reduction peaks shifted to the lower potential in the presence of Fe ions. Adding large amounts of the Fe salt had no significantly stronger effect on OER compared to adding a small amount of the Fe salt. SEM images of the foam as an anode electrode (amperometry at 1.46 V in KOH (pH ≈ 13)), after adding the Fe salt, showed nanoparticles on the surface of the electrode (ca. 70-100 nm). In the next step, OER of [FeO 4 ] 2− and Ni 2 O 3 under chemical conditions was investigated. The experiments showed that K 2 FeO 4 , compared to Ni 2 O 3 , is more active toward OER. Significant OER is observed in the presence of K 2 FeO 4 and Ni(II) hydroxide.