Highly Efficient Oxygen Evolution Reaction Enabled by Phosphorus Doping of the Fe Electronic Structure in Iron–Nickel Selenide Nanosheets

Abstract The electronic structure of active sites is critically important for electrochemical reactions. Here, the authors report a facile approach to independently regulate the electronic structure of Fe in Ni0.75Fe0.25Se2 by P doping. The resulting electrode exhibits superior catalytic performance for the oxygen evolution reaction (OER) showing a low overpotential (238 mV at 100 mA cm−2, 185 mV at 10 mA cm−2) and an impressive durability in an alkaline medium. Additionally, the mass activity of 328.19 A g−1 and turnover frequency (TOF) of 0.18 s−1 at an overpotential of 500 mV are obtained for P─Ni0.75Fe0.25Se2 which is much higher than that of Ni0.75Fe0.25Se2 and RuO2. This work presents a new strategy for the rational design of efficient electrocatalysts for OER.


Introduction
Electrochemical water splitting using intermittent renewable energy is a highly attractive approach for producing hydrogen without CO 2 emission. [1][2][3] The anodic oxygen evolution reaction (OER) is kinetically sluggish due to the four protoncoupled electron transfer kinetics and the oxygen─oxygen bond formation. [4][5][6][7] Currently, noble metal-based catalysts such as iridium and ruthenium oxides (IrO 2 or RuO 2 ) are recognized as the most active OER catalysts although as precious metals, their cost and low earth abundance makes the technology competitively unviable against fossil fuels. [8] Recently, transition metalbased (ranging from metal oxides, phosphides, chalcogenides, to DOI: 10.1002/advs.202101775 emerging single-atom) catalysts have been considered as promising candidates to replace noble metal-based catalysts for OER due to their low cost, excellent activity, and high stability. [9][10][11][12] In particular, considerable efforts have been devoted to developing nickel-iron-based OER electrocatalysts. [7,13,14] Among them, NiFeselenides have been widely investigated as ideal OER candidates due to the high electronic conductivity, diversity of stable crystal phases, and adjustable electronic structure. [15,16] Additionally, engineering the porosity, selenium vacancy, and the polarized electronic spin of Fe/Ni further enables the optimization of absorption/desorption of reaction intermediates and gas release. [17][18][19][20] Crucially, although the nickel-iron (Ni-Fe)-based bimetal electrocatalysts exhibit remarkable OER performance, there is still no scientific consensus on whether nickel or iron is the active center. [21] According to the Sabatier principle, the studies on metal hydr(oxy)oxides suggest that Ni might be the active site due to the optimal interaction strength with OH ad . [22] The presence of Fe is thought to affect the charge contribution leading to high valence Ni cations thereby enhancing their OER performance. [23] Indeed, the absorption of Fe impurities was reported to exert a partial-charge transfer activation effect on Ni improving the conductivity of Ni based electrocatalysts. [24] The study by Hu et al. revealed that Fe sites dominate the catalysis and the activity of Fe sites are 20-200 times higher than those of Ni sites in NiFe layered double hydroxides (LDHs). [25] In related work, Chen et al. proposed that Fe 4+ species are not directly responsible for the OER activity but the theoretical results suggested that high spin Fe 4+ leads to efficient formation of an active O radical intermediate. They further suggest that Ni 4+ catalyzes the subsequent O─O coupling, and it is the synergy between Fe and Ni that is responsible for the optimal performance for OER. [26,27] The results todate show that the electronic structures of Ni and Fe are integral to the mechanism which drives enhanced OER activity and understanding and tuning these electronic structures is the key to unlocking the pathways involved. More recently, elemental doping (Co, Fe, etc.) has also emerged as an efficient strategy to regulate the electronic structure of target materials. [18,20,[28][29][30][31][32][33] In this study, we have developed a facile approach to independently regulate the electronic structure of Fe in Ni 0.75 Fe 0.25 Se 2 nanosheets by P doping.

Results and Discussion
A schematic of Ni 0.75 Fe 0.25 Se 2 and its P doped analogue P─Ni 0.75 Fe 0.25 Se 2 is outlined in Figure 1a with the synthetic route described in Scheme S1, Supporting Information. Briefly, hexagonal NiFe layered double hydroxide (LDH) [ 34] The broad peak at around 25°is ascribed to the carbon cloth substrate ( Figure S3, Supporting Information). [35,36] Compared with NiSe 2 , the peaks were shifted to higher angles due to the incorporation of Fe, further confirming the formation of Ni 0.75 Fe 0.25 Se 2 (Figure 1b-inset). 12 It is worth noting that the crystal structure remains unchanged after P doping ( Figure 1b). Notably, the interlinked nanosheet struc-ture was preserved after doping with the thickness of nanosheets increasing slightly to ≈50 nm (Figure 1c Figure S8, Supporting Information) corroborated that the existence and homogeneous distribution of Ni, Fe, Se, and P elements within the sample. The composition was further determined by the inductively coupled plasma analysis spectrometry (ICP, Table S1, Supporting Information) to be Fe/Ni/Se/P∼1/3/8/0.03.
Compared with the X-ray photoelectron spectroscopy (XPS) of Ni 0.75 Fe 0.25 Se 2 , the presence of peaks located at 128.6 and 133.4 eV (Figure 2a) were assigned to P─M (Fe or Ni) and P─O bands due to surface oxidation, respectively, further confirming that P was successfully doped into Ni 0.75 Fe 0.25 Se 2 . [37,38] The P content determined by XPS (Table S2, Supporting Information) is much higher than the value obtained by the ICP examination, suggesting that the P element is mainly distributed close to the surface. In the region of Fe 2p (Figure 2b), the peaks can be fitted with two prominent peaks at 711.7 and 725.4 eV with two satellite peaks, indicative of Fe 3+ . [17,37,39] After P doping, the two main peaks of Fe 3+ were shifted slightly to a lower binding energy, while the spectrum of Ni 2p showed negligible change ( Figure S9, Supporting Information), suggesting that the electronic structure of Fe was modulated by introducing P. To confirm the modulation effect, Mössbauer analysis was applied and the results in Figure 2c reveal that a doublet with an isomer shift ( ) of 0.34 mm s −1 and quadrupole splitting (Δ) of 0.44 mm s −1 for Ni 0.75 Fe 0.25 Se 2 were observed, confirming the highspin, Jahn-Teller-distorted Fe 3+ species, similar to those reported previously. [26] As for P─Ni 0.75 Fe 0.25 Se 2 , the doublet peaks can be fitted with the signals of Fe 2+ ( = 0.32 mm s −1 ) and high spin Fe 3+ ( = 0.73 mm s −1 and Δ = 1.00 mm s −1 ). [40,41] The Mössbauer data with the XPS results proves that P doping in Ni 0.75 Fe 0.25 Se 2 can enrich the electron cloud around Fe 3+ . The electronic structure of Fe and Ni was further investigated by X-ray absorption spectra (XAS). The Ni-K edge X-ray absorption near-edge spectra (XANES) of the samples before and after P doping completely overlaps, indicating that the electronic structure of Ni remained unchanged after P doping (Figure 2d). Furthermore, the corresponding Fourier-transformed k 3 -weighted (k) function (Figure S10, Supporting Information) also signifies that the bonding environment of Ni atom is basically unchanged with obvious peaks at 2.44 and 1.62 Å corresponding to Ni─Se/Fe/Ni bonds and Ni─O bond, respectively. [20,42] For the Fe K-edges XANES (Figure 2e), the curve of P─Ni 0.75 Fe 0.25 Se 2 shifts to lower energy than Ni 0.75 Fe 0.25 Se 2 , indicating P doping can significantly reduce the valency of Fe 3+ . Another observation is that the peak intensity of Fe-Se/Fe/Ni for P─Ni 0.75 Fe 0.25 Se 2 decreases significantly compared with Ni 0.75 Fe 0.25 Se 2 ( Figure S11, Supporting Information). These results imply that P bonds more readily with Fe than Ni and that P-doping causes severe surface structural disorder. [20,42,43] The linear sweep voltammetry (LSV) polarization curves are displayed in Figure 3a. P─Ni 0.75 Fe 0.25 Se 2 exhibits significantly enhanced OER activity compared with Ni 0.75 Fe 0.25 Se 2 and RuO 2 . It needs only a low overpotential ( ) of 192 mV to drive a current density of 10 mA cm −2 , which is 20% lower than Ni 0.75 Fe 0.25 Se 2 and 60% lower than RuO 2 (the inset in Figure 3a). For Tafel analyses of the catalysts to evaluate the electrocatalytic kinetics (Figure 3b), P─Ni 0.75 Fe 0.25 Se 2 exhibits a substantially smaller Tafel slope of 31.5 mV dec −1 than Ni 0.75 Fe 0.25 Se 2 (43.7 mV dec −1 ) and RuO 2 (57.6 mV dec −1 ), indicating faster kinetics. In order to rule out the effect of the contact resistance and compare the performance of P─Ni 0.75 Fe 0.25 Se 2 with the reported electrocatalysts, iR-correction was applied for P─Ni 0.75 Fe 0.25 Se 2 (Figure 3a,b). P─Ni 0.75 Fe 0.25 Se 2 -iR delivers a current density of 10 mA cm −2 at an ultra-low overpotential of 185 mV with a small Tafel slope of 27.2 mV dec −1 , suggesting an impressively higher OER catalytic activity than related electrocatalysts (Table  S3, Supporting Information). The mass activity (at = 0.50 V) of P─Ni 0.75 Fe 0.25 Se 2 is 328.19 A g −1 , which is 3.19 times higher than that of Ni 0.75 Fe 0.25 Se 2 (102.90 A g −1 ) and 1.28 times RuO 2 (256.68 A g −1 ) suggesting P-doping greatly improves the OER activity of Ni 0.75 Fe 0.25 Se 2 (Figure 3c and Table S4, Supporting Information). Similarly, the TOF in Figure 3d shows that the value of P─Ni 0.75 Fe 0.25 Se 2 is the largest within the studied potential range, indicating its outstanding intrinsic electrocatalytic activity. These experimental results distinctly demonstrate that regulating the electronic structure of Fe in Ni 0.75 Fe 0.25 Se 2 by P doping is a viable route to improve its OER catalytic activity. The work further confirms the active role of Fe in Ni 0.75 Fe 0.25 Se 2 for OER.
The durability of P─Ni 0.75 Fe 0.25 Se 2 was evaluated by chronopotentiometric measurement at 100 mA cm −2 (Figure 3e). The current remained steady for 120 h without any appreciable increase in potential and the LSV curve of P─Ni 0.75 Fe 0.25 Se 2 remains unchanged after 5000 scans (The inset in Figure 3e). Additionally, the XRD patterns shows no change compared with the initial P─Ni 0.75 Fe 0.25 Se 2 ( Figure S12, Supporting Information), sug-gesting that no new crystal was formed under the OER stability test. The XPS results indicate that the binding energy of Fe 2p and Ni 2p was shifted slightly to higher energy due to the formation of metal oxide/hydroxide species on the surface of P─Ni 0.75 Fe 0.25 Se 2 during the OER process ( Figure S13, Supporting Information). [12,17,20,38] Similarly, the intensity of the P─M peak decreased significantly to form P─O species. These in-situ generated species contribute to the enhanced stability. [44] Importantly, the P─Ni 0.75 Fe 0.25 Se 2 electrode delivered a Faradaic efficiency of ≈100% for OER (Figure 3f), indicating that the observed current exclusively originated from OER rather than other processes.
The electrochemical double-layer capacitance (C dl ) of P─Ni 0.75 Fe 0.25 Se 2 (6.25 mF cm −2 ) is higher than RuO 2 (5.92 mF cm −2 ) and Ni 0.75 Fe 0.25 Se 2 (4.13 mF cm −2 ), indicating P doping indeed increases the number of active sites, which is beneficial for the OER process ( Figure S14, Supporting Information). When the LSV curves are normalized by the electrochemical surface area (ECSA) to exclude the contribution of larger ECSA for OER performance, the results shown in Figure 4a indicate that P─N 0.75 Fe 0.25 Se 2 shows much better OER activity than RuO 2 and Ni 0.75 Fe 0.25 Se 2 samples, reflecting that the enhanced OER activity is not only attributed to the increased ECSA but also the improved intrinsic activity of the catalyst due to the optimized electronic structure. [45] In addition, the P doping showed minimum influence on the morphology of Ni 0.75 Fe 0.25 Se 2 , confirmed by the comparable surface area before and after P doping ( Figure  S15a,b, Supporting Information). [46,47] The LSV curves normalized by the BET surface area further confirmed the improved intrinsic catalytic activity by P doping ( Figure S15c, Supporting Information). The activation energy (E a ) of different catalysts was calculated by measuring the currents at different temperatures ( Figure S16, Supporting Information). [48,49] Figure 4b shows that the E a derived from the slopes of the Arrhenius plot is 16.  (Figure 4e). On the other hand, the capability of adsorption and desorption for the oxygen-containing intermediates (*OH, *O, *OOH) is comparably or even more critical for the performance of OER electrocatalysts. [21,22,50,51] The adsorption strength of OH − ions on Ni 0.75 Fe 0.25 Se 2 and P─Ni 0.75 Fe 0.25 Se 2 during OER was further verified based on the Laviron analysis. [52] As shown in Figure 4f and Figures S17 and S18, Supporting Information, P─Ni 0.75 Fe 0.25 Se 2 exhibits a larger K s (0.14 s −1 ) than Ni 0.75 Fe 0.25 Se 2 (0.09 s −1 ), suggesting that the enhanced adsorption capability of P─Ni 0.75 Fe 0.25 Se 2 sites for OH * intermediate facilitates the OER process, [52][53][54] in good agreement with the results of O 2 -TPD.

Conclusion
We have proposed a unique way to independently regulate the electronic structure of Fe in Ni 0.75 Fe 0.25 Se 2 by P doping. The optimized electronic structure of the resulting catalyst has been studied and confirmed using XPS, Mössbauer spectra, and XANES spectra. The corresponding electrode exhibits outstanding OER activity and durability to achieve a benchmark current density of 10 mA cm −2 at an ultralow overpotential of 185 mV. The mechanistic investigation reveals that P doping endows the Ni 0.75 Fe 0.25 Se 2 electrocatalyst with enhanced conductivity, www.advancedsciencenews.com www.advancedscience.com optimized adsorption of oxygen-containing intermediates, and a reduced kinetic barrier. This work provides an in-depth insight into understanding the effect of P doping in Ni 0.75 Fe 0.25 Se 2 . The use of doping to regulate the electronic structure of a single metal site in multinary transition metal electrocatalysts opens new pathways for enhancement of OER in a wide range of systems.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.