Introduction

The prosperity of human society relies heavily on fossil fuels, which are converted into carbon dioxide and released into the atmosphere in the process of utilization, thus leading to the greenhouse effect [1, 2]. Therefore, weakening or even completely replacing the current energy structure has become an inevitable problem. Water splitting, as a form of energy conversion without harmless byproduct, has attracted the attention of many research teams. However, the popularity of this new form of energy is hampered by complex four-electron process of oxygen evolution reaction (OER), a half reaction of water splitting [3, 4]. As commercial catalysts, Ru- and Ir-based materials have played a great role in promoting the sluggish OER process [5]. Yet, it is precisely because of their high cost the industrial scale applications have been limited.

Up to now, transition metal-based phosphides [6,7,8,9,10], borides [11, 12], nitrides [13,14,15,16], selenides [17,18,19,20,21], and sulfides [22,23,24,25,26], etc. showed excellent performance in accelerating the OER process. Among of them, nickel sulfide has been widely concerned because of its excellent electrical conductivity and high intrinsic activity. For instance, Zhang et al. [27]. prepared Ni-S/graphene heterostructure electrocatalyst, that requires an overpotential of 300 mV to afford OER at 10 mA cm−2. Similarly, Chen et al. reported the superior OER activity (297 mV @ 11 mA cm−2) of Ni-S nanosheet fabricated by hydrothermal method [28]. However, these electrocatalytic performances of nickel sulfide used in OER process are far from meeting the requirements of mass production and need to be improved urgently. In fact, the modification of the above sulfides and even most catalysts is mainly based on two points: morphology optimization and the electronic structure modulation [29, 30]. So far, quite a few research results emerge from zero-dimensional quantum dots to three-dimensional hybrid structures. For instance, Mohanty et al. successfully synthesized MoS2 quantum dots, which possess outstanding OER performance (370 mV @ 10 mA cm−2 and 39 mV dec−1) [31]. Dai and co-workers fabricated self-assembled Ni0.7Co0.3P nanoflower, which presented an enhanced performance for OER (230 mV @ 10 mA cm−2 and 78 mV dec−1) [32]. Among the reported nanostructures with different morphologies, the vertically aligned nanoarray structure with larger surface area and more exposed electrochemically active sites are of great significance to improve the electrochemical performance for OER [33, 34]. In addition, adjusting the electronic structure of the catalyst is another effective strategy for improving the electrocatalytic activity. For nickel-based sulfides, the addition of Ni/Fe-based hydroxides can significantly promote the catalytic activity of OER [35,36,37]. Liu synthesized a NiCo2S4/NiFe LDH interface by hydrothermal possess, wherein the electronic interactions between NiCo2S4 and NiFe LDH have been regulated to suppress the free energy of intermediate, thus enhancing the catalytic activity [37].

Based on the above analysis, it is significant to combine morphology optimization and regulate the electronic structure at the same time to obtain excellent catalytic performance of nickel sulfide, thus, a controllable two-step electrochemical deposition method for constructing a neoteric Fe(OH)3/Ni9S8 nanoarrays on etched stainless steel was first developed as efficient and stable OER electrocatalyst. The advantages of particular structure are stated below: (1) The open vertical array of Fe(OH)3/Ni9S8 nanosheets not only exposed a wealth of catalytic active sites, but also increased the opportunity for electrolyte solution to contact with catalytic active sites and oxygen release during OER process. (2) The interaction between Fe(OH)3 and Ni9S8 creates a more favorable interface for the OER. (3) The outer layer of Fe(OH)3 prevents the Ni9S8 layer from direct contact with the electrolyte and thus avoiding the additional Ni oxidation potential. (4) The self-supporting growth catalyst deposited on conductive substrate can avoid the covering of active sites by binder and thus further improve the OER performance. (5) The structure prepared by electrodeposition will make the active materials firmly attached on ESS, thus ensuring its durable performance. Therefore, under the condition of inheriting the above advantages, the Fe(OH)3/Ni9S8/ESS has displayed ultra-low overpotential (206 mV @ 10 mA cm−2), which outperforms the RuO2/ESS (310 mV @ 10 mA cm−2). This systematic study provides an inspiration for the design of low-cost and high efficiency catalysts for boosted OER performance.

Experimental section

Synthesis of Ni9S8/ESS

First of all, the stainless steel (SS) was cut into 3*2 cm2 pieces, then washed with deionized water and alcohol for several times, and finally put it into 4 M HCl and etched for 25 min by ultrasonication. Subsequently, the etched stainless steel (ESS) was placed in an electric oven at 60 °C for 8 h prior to electrodeposition.

The typical synthesis of nickel sulfide on ESS substrates as follows. The electrolyte for the first electrodeposition includes 0.26 M thioacetamide (TAA) and 0.05 M nickel nitrate hexahydrate (Ni(NO3)2.6H2O). The deposition of precursor was executed under the DC stabilized power supply. During the electrodeposition process, carbon rod was used as anode and the ESS was used as cathode, and the potentiation deposition (5 V) was selected for 40 s. After the deposition, the surface of the electrode was washed with deionized water and dried at 60 °C in an electric oven.

Synthesis of Fe(OH)3/Ni9S8/ESS

The electrode prepared above was used as the Cathode for the secondary deposition, and the electrolyte was replaced with 0.02 M ammonium ferrous sulfate [(NH4)2Fe(SO4)2], and the anode was still a carbon rod. Deposition in constant current mode (2 mA cm−2) was used for 60 s, and then the surface of the electrode was washed with deionized water and dried in the 60 °C oven.

Synthesis of RuO2/ESS

As a reference group, 2 mg cm−2 RuO2 loaded ESS was prepared as follows; first, 5 mg of RuO2 was dispersed in 750 μL of deionized water (DW), 100 μL of Nafion and 150 μL of alcohol, then ultrasonication was performed until the mixture was evenly distributed. Finally, the mixture was dropped onto ESS. Next, the sample prepared above is then left overnight in an electric oven at 60 °C (denoted as RuO2/ESS).

Material characterizations

X-ray diffractometer (XRD) diffractometer was used to measure the composition and crystal phase of samples treated with different processes, which recorded on X’pert PRO PEP (Cu Ka) with 2θ from 10 to 80 degrees. Materials loaded on the substrate are identified by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha) mainly the valence and composition. The micromorphology of the prepared samples was obtained from scanning electron microscopy (SEM, ZEISS Gemini 300, OXFORD Xplore) and transmission electron microscopy (TEM, FEI Tecnai G2 F20, OXFORD X-max 80 T).

Measurements of electrocatalytic properties

The related OER performance tests were performed on a typical three-electrode electrochemical workstation. The sample to be tested served as the working electrode, saturated Hg/HgO electrode was served as the reference electrode, and a carbon rod was used as the counter electrode. Cyclic voltammetry (CV) was employed before the linear voltammetry scanning (LSV) between 0 and  0.8 V (vs. Hg/HgO) at a scan rate of 0.1 V s−1. The LSV used to determine the OER performance curve, which is scanned in reverse direction (0.8–0 V vs. Hg/HgO) at a low scanning rate (2 mV s−1). All potentials measured by electrochemical workstations have been corrected using the Nernst equation: ERHE = EHg/HgO + 0.923 V. Finally, the overpotential is the potential at the reversible hydrogen electrode corresponding to 10 mA cm−2 in the LSV curve minus 1.23 V. The Tafel slope is calculated by Tafel formula based on LSV data. The impedance spectra of different samples were tested by setting the frequency in the range of 0.1 Hz–100 kHz at the voltage of 0.513 V vs. Hg/HgO. The cyclic voltammetry was obtained by scanning from 40 to 200 mV s−1 with 50 mV s−1 interval in the non-Faraday region (0.953–1.053 V vs. RHE). The long-term stability was determined by 12 h chronoamperometry test at a constant voltage of 1.44 V vs. RHE. The stability of the catalyst was tested after 500 cycles of voltammetry in 1 M KOH.

Results and discussion

Material characterization

The fabrication process of as-synthesized vertical nanoarrays is depicted in Fig. 1. In this method, the smooth surface of stainless steel (Fig.S1a, e, g) is modified for better deposition of the active material. After 30 min of acid etching, the surface of the stainless steel becomes rough and owns cracks and pits on its surface (Fig.S1b, f, h), which provides suitable surface conditions for the uniform deposition of the precursor. At the same time, it is worth mentioning that, compared with the untreated stainless steel, the overpotential and Tafel slope of the stainless steel after acid etching are greatly reduced (Fig.S4). Subsequently, Ni9S8 nanoarrays was formed on the ESS (Fig.S1c, i) after the first electrodeposition. Finally, Fe(II) ions in solution reacted with reduced oxygen and water to form Fe(OH)3 at the surface of Ni9S8 nanoarrays to construct Fe(OH)3/Ni9S8 composite catalyst on ESS (Fig.S1d, j) in the second electrochemical deposition process.

Figure 1
figure 1

Schematic illustration of synthesis procedure for Fe(OH)3/Ni9S8/ESS

Full size image

In addition, in the second deposition process, we took deposition time as the independent variable and 10 s interval as the time gradient. Through the test, we found that with the increase of deposition time, the catalytic performance also increased significantly and reached the maximum value at 40 s, so the later experimental tests were all based on the second 40 s.

The morphologies of as-papered samples were obtained by SEM. According to Fig. 2a and Fig. 2b, the vertically arranged nanoarrays grow uniformly on the surface of the ESS, and the morphology of the one and two-step deposition is basically the same with slightly differences, which indicates that the newly deposited species on the surface have no effect on the morphology. The remaining vertical open structure allows more open space for the catalyst to be in contact with the electrolyte, and thus facilitating charge transfer and mass transport. In order to clarify the elements contained in the secondary deposition products, the overall morphology EDS mapping (Fig. 2c) was carried out and the results showed that Fe, Ni, S and O were uniformly distributed on the microstructure surface. To indicate the composition of the growing material on the ESS matrix, the XRD tests were performed and the results are shown in Fig. S2a., All the XRD patterns show that the samples during different processes have the same three strong peaks of the austenite, which originated from the ESS substrate. And, it can be seen that after each electrodeposition, the sample displays a weaker peak than previous sample, which respectively indicates the generation of nickel sulfide and iron hydroxide. However, due to the strong inherent peak of stainless steel and the thin active material deposited on the SS surface (Fig. S2d), the phase peak of as-prepared catalyst is not prominent. Thus, we then use XPS to figure out the main form of elements in the surface of the samples. As shown in Fig.S2b, Ni, Fe, S and O elements all exist in the prepared Fe(OH)3/Ni9S8/ESS catalyst, which is consistent with the EDS result. The 2p region of nickel divides (Fig. 3a) into two regions, including Ni 2p1/2 and 2p3/2. The three peaks at 854.98 (Ni2+), 855.2 (Ni3+), and 859.8 eV (satellite peak) belong to the Ni 2p3/2 region, and the binding energies located at 871.3 (Ni2+), 873.2 (Ni3+), and 877.8 eV (satellite peak) are ascribed to Ni 2p1/2 region. These results about the peak of nickel accord with the characteristic peaks of Ni9S8 species [38,39,40]. It is noteworthy that the Ni 2p peak of the Fe(OH)3/Ni9S8/ESS shifts in a negative direction compared with the Ni9S8/ESS (Fig.S2c) without the addition of the second step of iron hydroxide (Ni 2p3/2 peaks appeared at 853.4 (Ni2+) and 854.8 eV (Ni3+); Ni 2p1/2 peaks located at 870.8 (Ni2+), and 872.6 eV (Ni3+)), which is caused by the interaction of hydroxide of iron with nickel sulfide [41, 42]. Moreover, in S 2p spectrum (Fig. 3b), the binding peaks displayed at 162.7 and 164.7 eV are corresponding to Ni-S bonding structure [24, 38,39,40, 43]. Meanwhile, the peak at about 168.5 eV belongs to the sulfate species on account of surface oxidation [24, 44].

Figure 2
figure 2

SEM images of a Ni9S8/ESS b and Fe(OH)3/Ni9S8/ESS. c Overall morphology and EDS elemental mappings of Fe(OH)3/Ni9S8/ESS

Full size image
Figure 3
figure 3

a Ni 2p, b Fe 2p, c S 2p and d O 1 s core-level XPS spectra of Fe(OH)3/Ni9S8/ESS

Full size image

According to the high-resolution XPS spectra of Fe 2p (Fig. 3c), the binding energies of the fitting peaks at 711.1 and 724.2 eV are corresponding to Fe 2p3/2 and Fe 2p1/2 regions, respectively, which confirmed that Fe3+ plays a dominant role in the catalyst [45,46,47]. As for the O 1 s spectrum, three peaks (Fig. 3d) occur mainly in three areas are related to hydroxyl groups (531.7 eV), lattice oxygen (530.3 eV) and surface-adsorbed water molecules (553.5 eV) [44, 48, 49]. Among them, OH group is the main component, and a small amounts of lattice oxygen may form as a result of surface oxidation of iron substances when exposed to air.

The microstructure and composition of as-synthesized samples were further characterized by TEM. As shown in Fig. 4a and 4c, both the precursor and the final sample are composed of ultra-thin nanosheets, which is equivalent to that of the second deposition process. Moreover, the HRTEM image (Fig. 4b) of the precursor nanoarrays matches the (043) crystal plane of Ni9S8 with a lattice spacing of 0.213 nm. As for the nanosheets after the second deposition, it can be seen that not only the (043) plane of the precursor still exists from Fig. 4d, but also the (400) crystal plane (0.208 nm) of Fe(OH)3 is generated, which further proves that the existence of hydroxide of iron and the nickel sulfide in the composite. Therefore, after analyzing the characterization results of SEM, XRD, XPS and TEM, we conclude that the composition of the nanoarray is a composite of Ni9S8 and Fe(OH)3.

Fig. 4
figure 4

a TEM image of Ni9S8/ESS. b HRTEM image of Ni9S8/ESS. c TEM image of Fe(OH)3/Ni9S8/ESS. d HRTEM image of Fe(OH)3/Ni9S8/ESS

Full size image

Electrochemical performances of Fe(OH)3/Ni9S8/ESS

Electrochemical performance tests were performed on an electrochemical workstation in 1.0 M alkaline solution (KOH). Figure 5a shows the LSV curve of stainless steel in various processes and RuO2/ESS as the reference group. It can be seen from Fig. 5a and Fig. 5c that the first deposited vertical nanoarrays have shown excellent electrochemical properties compared to the original substrate (η10 = 416 mV), etched substrate (η10 = 350 mV) and even superior to commercial RuO2 catalyst ( η10 = 310 mV) at a current density of 10 mA cm−2. It is noticeable that the Ni9S8/ESS displays a weak oxidation peak around 1.39 V (vs RHE) due to the transformation Ni(II) to Ni(III) [50, 51]. After the second electrochemical deposition, the overpotential (η) of Fe(OH)3/Ni9S8/ESS at 10 mA cm−2 decreased by 88 mV compared to the Ni9S8/ESS. One of the most obvious changes is the disappearance of oxidation peaks, which is caused by the deposition of iron hydroxide on the surface of nickel sulfide. As a result, the Ni9S8 layer can prevent the nickel species from direct exposure to the alkaline environment and therefore does not require a higher potential to overcome the oxidation peak [36]. In addition, compared with the single Ni9S8 precursor and RuO2, after iron deposition the current density of Fe(OH)3/Ni9S8/ESS significantly increases to more than 100 mA cm−2, making the catalyst more practical.

Figure 5
figure 5

a OER polarization curves of Fe(OH)3/Ni9S8/ESS, Ni9S8/ESS, RuO2/ESS, ESS, and bare SS. b Tafel plots for the OER, and c Tafel slopes and overpotentials of Fe(OH)3/Ni9S8/ESS, Ni9S8/ESS, RuO2/ESS, ESS, and bare SS. d Stability test of Fe(OH)3/Ni9S8/ESS. The inset is the LSV curve of Fe(OH)3/Ni9S8/ESS before and after 500 CV cycles

Full size image

The OER kinetics of the catalyst was assessed using the Tafel slope. As shown in Fig. 5b, c, the as-synthesized catalyst possesses the lowest Tafel slope (81.79 mV dec−1), which is lower than those of Ni9S8/ESS (94.69 mV dec−1), ESS (106.04 mV dec−1), RuO2/ESS (98.63 mV dec−1), and SS (133.14 mV dec−1), indicating that the Fe(OH)3/Ni9S8/ESS displayed the fastest OER kinetics among the as-prepared samples. For a practical catalyst, stability is a non-negligible property. Herein, we did 500 cycles CV scan and then we conducted a LSV test, and the obtained curve was almost the same as the curve that obtained after the first scan (the inset of Fig. 5d), which confirm its durability. Then, we performed a stability test with chronopotentiometry (CP) near 10 mA cm−2, and Fig. 5d shows that the current did not begin to decay even after 12 h, that signifies the long-term OER performance of the catalyst. Moreover, the morphology of Fe(OH)3/Ni9S8/ESS after long-term OER process was further carried out by SEM measurement. As shown in Fig.S6, the Fe(OH)3/Ni9S8/ESS still maintain the previous ultra-thin nanosheets structure, which indicates its good structure stability in the OER process. To further study the charge transfer process at the interface of Fe(OH)3/Ni9S8/ESS nanoarray electrode, we used EIS to conduct impedance test. According to Fig. 6a (Table S1), the fitting semicircle diameter decreases from small to large in the following order; Fe(OH)3/Ni9S8/ESS (1.769 Ω) < Ni9S8/ESS (1.795 Ω) < ESS (1.81 Ω) < SS (1.88 Ω), which demonstrates the small interface resistance and fast kinetics of the Fe(OH)3/Ni9S8/ESS. Meanwhile, the magnitude of charge transfer resistance (Rct) also accords with this order, Fe(OH)3/Ni9S8/ESS (2.169 Ω) < Ni9S8/ESS (2.543 Ω) < ESS (6.86 Ω) < SS (50.51 Ω), indicating that the catalyst of final deposition has better charge transfer properties. From the LSV curve and the Tafel slope, the prepared catalyst has excellent performance and fast kinetics toward OER. As an index to estimate the available active sites of the electrocatalyst, a double-layer capacitance based on the electrochemical active surface area was calculated [52, 53] through CV plots (Fig. S3) in the non-Faradaic potential area (0.953–1.053 V vs. RHE) at different scan rates. As expected, (Fig. 6b), Fe(OH)3/Ni9S8/ESS has the highest Cdl value, indicating that it possesses the most active sites compared to the other reference samples.

Figure 6
figure 6

a Electrochemical impedance spectroscopy and b double-layer capacitance (Cdl) data of electrocatalysts fabricated in this study

Full size image

Conclusion

In summary, Fe(OH)3/Ni9S8/ESS nanoarrays were synthesized by electrodeposition of Fe(OH)3 on a Ni9S8 nanoarrays precursor using a template-mediated strategy. Benefited from the nanoarrays morphology superiority and the coupling effect of Fe(OH)3 and Ni9S8, as-constructed catalyst exhibits excellent performance than those of Ni9S8/ESS, RuO2/ESS and most non-noble metal catalysts-fabricated on current collectors (Table S2) toward OER in 1.0 M alkaline electrolyte. In particular, the superior performance was ascribed to the enhanced electron transfer and abundant active sites exposure of nanoarrays architecture constructed on ESS. Moreover, the layer-by-layer electrodeposited strategy was expected to prevent the oxidation of active Ni9S8 during the water oxidation, and hence resulted in an exceptional long-term OER ability in a 12 h continuous i-t test without any current density decay. This work provides a facile synthesis strategy with novel structural design for the development of efficient and low-cost electrocatalysts.