Hierarchical Nickel-Cobalt Dichalcogenide Nanostructure as an Efficient Electrocatalyst for Oxygen Evolution Reaction and a Zn-Air Battery.

A unique three-dimensional (3D) structure consisting of a hierarchical nickel-cobalt dichalcogenide spinel nanostructure is investigated for its electrocatalytic properties at benign neutral and alkaline pH and applied as an air cathode for practical zinc-air batteries. The results show a high oxygen evolution reaction catalytic activity of nickel-cobalt sulfide nanosheet arrays grown on carbon cloth (NiCo2S4 NS/CC) over the commercial benchmarking catalyst under both pH conditions. In particular, the NiCo2S4 NS/CC air cathode shows high discharge capacity, a narrow potential gap between discharge and charge, and superior cycle durability with reversibility, which exceeds that of commercial precious metal-based electrodes. The excellent performance of NiCo2S4 NS/CC in water electrolyzers and zinc-air batteries is mainly due to highly exposed electroactive sites with a rough surface, morphology-based advantages of nanosheet arrays, good adhesion between NiCo2S4 and the conducting carbon cloth, and the active layer formed of nickel-cobalt (oxy)hydroxides during water splitting. These results suggest that NiCo2S4 NS/CC could be a promising candidate as an efficient electrode for high-performance water electrolyzers and rechargeable zinc-air batteries.

Introduction

Splitting water into pure <span class="Chemical">hydrogen and oxygen to generate sustainable green hydrogen energy has been intensively studied in recent years, which can replace fossil fuel use.[1,2] However, the efficiency of water splitting has so far been limited by the lack of sustainable catalysts toward the oxygen evolution reaction (OER) that can accelerate the kinetics.[3−5] So far, IrO and RuO2 are the best-known OER catalysts, although their high cost and scarcity limit their widespread use.[4] Meanwhile, some promising attempts have been devoted to developing an efficient nonprecious metal OER catalyst under alkaline conditions. However, an almost harsh alkaline medium presents severe corrosion and related environmental issues.[6,7] In this regard, someday, the splitting of water at neutral pH from ocean or river would be the target goal to satisfy renewable future hydrogen energy.[4] It is thus highly required to develop efficient OER electrocatalysts that can operate in both alkaline and neutral media for overall water splitting even though it is relatively tough searching for those catalysts.

Nowadays, various nonprecious transition metal-based catalysts are being explored, for example, transition <span class="Chemical">metals,[8] transition-metal oxides,[9,10] chalcogenides,[11−14] phosphides,[15−17] hydroxides/oxyhydroxides,[18,19] carbides,[20] borides,[21] and so on. Although lots of established catalysts have been reported concerning their excellent OER activity under alkaline conditions, only a few of them could still maintain their catalytic activity in neutral media. Cai et al. reported that the amorphous cobalt sulfide porous nanocubes showed a low OER onset potential of 1.5 V, comparable to that of RuO2 (1.49 V).[22] However, a still substantial overpotential of 570 mV is needed to generate 4.59 mA cm–2 in phosphate-buffered solutions (PBSs; pH 7.0), whereas it could generate 10 mA cm–2 current density at 290 mV in 1 M KOH (pH 14.0). Similarly, sulfur-incorporated NiFe2O4 nanosheets (NSs) on nickel foam (S–NiFe2O4/NF) developed by Liu et al. exhibited a remarkably enhanced water-splitting performance for both OER and hydrogen evolution reaction (HER) as a bifunctional electrode under both alkaline and neutral conditions.[23] The S–NiFe2O4/NF still requires 1.921 V to deliver 10 mA cm–2 in 1 M PBS (pH 7.4) for overall water splitting in three electrode systems, mostly occurring during OER with an overpotential of 494 mV. As air cathode catalysts for Zn–air batteries, Prabu et al., demonstrated a highly active one-dimensional structure of a spinel NiCo2O4 catalyst in rechargeable Zn–air batteries and Li–O2 batteries.[24,100] Recently, Meng et al. constructed Co0.85Se nanocrystals in situ coupled with N-doped carbon with a metal–nitrogen–carbon (M–N–C) structure and short diffusion pathways for the transport of electron/ion to improve the Zn–air battery performance.[25,26] For instance, Wu et al. reported zinc cobalt sulfide, the nanoneedle (NN) arrays grown on the carbon fiber paper electrode catalyst, which enables the Zn–air battery operation with an overpotential of 0.85 V and a long cycle life time of up to 200 cycles at 10 mA cm–2 as well as comparable water-splitting performance.[27] Wang et al. proposed Co3FeS1.5(OH)6 hydroxysulfides serving as a superb air electrode catalyst with a low overpotential of 0.84 V and prolonged cyclability over 36 h test for 108 cycles at 2 mA cm–2.[28]

In pursuit of high electrochemical performance in water electrolyzers and for a Zn–air battery application, the spinel bi<span class="Chemical">metallic sulfide NiCo2S4 with abundant redox chemistry has been considered to be the most promising electrochemically active material, which exhibits 2 orders of magnitude larger than that of NiCo2O4 and ∼104 times better electric conductivity than conventional single-metal compounds.[12,29] Moreover, the stable spinel structures of bimetallic sulfide with a formula of AB2S4 possess plentiful exposed edge sites, leading to a higher electrochemical activity. Therefore, it has been widely applied for supercapacitors, Li-ion batteries, and a counter electrode for dye-sensitized solar cells as well as in water electrolyzers.[30] For example, in our group, NiCo2S4 nanowire arrays were directly grown on 3D Ni foam (NiCo2S4 NW/NF) as a water-splitting catalyst and applied in an alkaline water electrolyzer. Because of its intrinsic properties, large surface area, and well-separated NW structures, NiCo2S4 NW/NF afforded continuous water-splitting reaction of generating hydrogen and oxygen gas at a cell voltage of only 1.63 V to generate 10 mA cm–2 current density.[31] Ma et al. developed 3D networked porous NiCo2S4 nanoflakes on NF, which can offer more exposed active sites and easy transport of electrons and ions, thereby leading to significantly improved HER activity and stability.[32] The outstanding OER performance of the spinel bimetallic sulfide NiCo2S4 has also become a promising application for rechargeable Zn–air batteries involving reversible OER and ORR.

Also, hybridizing NiCo2S4 with the 3D structure of NS or NN with a <span class="Chemical">carbon cloth (CC) (NiCo2S4 NS/CC or NN/CC) substrate by an in situ growth hydrothermal approach can be an efficient way to enhance the robustness of the electrode. It could also help to minimize the agglomeration of NiCo2S4 nanostructures and the detachment during long-term operation, and make faster ion/electron kinetics. Also, the advantages of CC, such as flexibility, high conductivity, and corrosion/dissolution resistivity might lead to enhanced catalytic activity and stability in a wide pH range.[33] Meanwhile, tuning the nanostructure and the morphology, as well as porosity, can be another promising strategy to produce numerous exposed catalytic active sites on the catalyst surface.[17]

On the basis of our knowledge, we describe the physical and electrochemical properties of NiCo2S4 NS/CC as a highly active OER catalyst in both neutral and alkaline media, which have not been thoroughly investigated so far. Mainly, this is the first time a bimetallic sulfide, the NiCo2S4-based material, is reported to catalyze OER under a neutral condition. The catalytic performance of NiCo2S4 NS/CC is remarkably enhanced compared to the recent reports, particularly under a neutral condition. The NiCo2S4 NS/CC electrocatalyst exhibits the lowest OER overpotentials of 260 and 402 mV to generate 10 mA cm–2 in alkaline and neutral media, respectively. Specifically, it exhibits a low Tafel slope of 123 mV dec–1 and a high turnover frequency (TOF) of 8.17 × 10–3 s–1 at 1.63 V applied potential to drive 10 mA cm–2 current density under neutral conditions, confirming superior intrinsic activity with a substantial electrochemical active surface area (ECSA) of NiCo2S4 NS/CC compared with commercial RuO2/CC and other previously reported OER electrocatalysts. In addition, the constructed NiCo2S4 NS/CC air cathode for primary and rechargeable Zn–air batteries exhibits high discharge capacity, a narrow overall overpotential, and a long cycling life time exceeding the benchmark for precious metal-based electrodes.

Experimental Methods

Material Synthesis

Cobalt nitrate hexahydrate [Alfa Aesar, Co(NO3)2·6H2O], nickel nitrate hexahydrate [Sigma-Aldrich, Ni(NO3)2·6H2O], urea (Sigma-Aldrich, CH4N2O), and sodium sulfide hydrate (Sigma-Aldrich, Na2S·xH2O) were used to synthesize the electrodes. A piece of CC (NARA CELL-TECH, 0.7 cm × 0.7 cm) was utilized with further treatment with ethanol. The NiCo2S4 NSs grown on CC (NiCo2S4 NS/CC) were prepared through a two-step hydrothermal process. A nickel nitrate of 0.004 M and cobalt nitrate of 0.008 M was dissolved in 120 mL of deionized water; further, 0.012 M urea was added. The obtained solution was transferred into a Teflon-lined stainless steel autoclave of 200 mL capacity, and a piece of CC was immersed in the solution. The autoclave was heated at 120 °C for 8 h in an electric oven. After the first step, the electrode was washed with deionized water several times to eliminate unreacted residues. Consequently, sodium sulfide flakes were dissolved in deionized water to prepare a 0.2 M sulfide solution for a sulfurization process. This sulfur-containing solution was again transferred to the autoclave and heated at 160 °C for 8 h in an electric oven. After cooling down to room temperature naturally, we washed the synthesized electrode several times with ethanol and deionized water, followed by the drying step in the vacuum oven at 60 °C overnight. For comparison, NiCo2S4 NN arrays on CC (NiCo2S4 NN/CC) were prepared by changing the temperature of the first hydrothermal step from 120 to 130 °C and maintaining the heating time of 8 h. To synthesize NiCo2O4 with NSs morphology, which is grown on CC (NiCo2O4 NS/CC), the electrode after the first step of the hydrothermal growth process was annealed at 450 °C for 2 h in an air atmosphere.

Microstructural Characterizations

The morphology and element compositions were studied on a field-emission scanning electron microscopy (FE-SEM, Hitachi-S4800, 3 kV) system equipped with a Horiba Scientific energy dispersive spectrometer and using transmission electron microscopy (TEM, Hitachi HF-3300, 300 kV). The crystal structures of all catalysts were examined by powder X-ray diffraction (XRD, Rigaku MiniFlex600). The composition of the catalyst was studied using X-ray photoelectron spectroscopy (XPS, Thermo-Scientific ESCALAB 250Xi).

Electrochemical Measurements

For electrochemical measurements, the OER catalytic performance was evaluated by linear sweep voltammetry (LSV) with a low scan rate of 1 mV s–1 in an electrolyte of 1 M <span class="Chemical">KOH and a PBS without purging oxygen. The OER performance was evaluated in a three-electrode configuration directly using synthesized electrodes such as NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC as a working electrode (1.0 mg cm–2), a saturated calomel electrode (SCE) as a reference electrode, and a Pt wire as the counter electrode. Similarly, commercial RuO2 cast onto CC was used as a working electrode (1.4 mg cm–2), Pt wire as a counter electrode, and SCE as a reference electrode. The potentials reported were converted to the reversible hydrogen electrode (RHE). All electrochemical results were iR-corrected, considering the ohmic resistance from the electrolyte. The current densities presented in this paper are normalized concerning the geometric surface area of the electrode. The cyclic voltammetry (CV) was performed in N2-saturated 1 M KOH at room temperature with a scan rate of 10 mV s–2. Electrochemical impedance spectroscopy (EIS) was performed within a frequency range of 0.01 Hz to 0.1 MHz.

The ECSA is calculated by following an established methodology reported in the literature.[31] In detail, through the cyclic voltammogram obtained in a non-faradaic region at various scan rates (1, 2.5, 5, 10, 20, and 50 mV s–1), double-layer capacitance (Cdl) can be estimated. By plotting the anodic and cathodic current densities against the scan rate, the obtained linear slope value is Cdl. Finally, the ECSA can be obtained from the following equation:

Cs denotes the specific capacitance of a flat, smooth surface of the electrode material, which is assumed to be 26 μF cm–2 for Ni- and Co-containing materials.

Zinc–Air Battery Fabrication and Testing

For full-cell zinc–air battery evaluation, the as-synthesized NiCo2S4 NS/CC or commercial catalyst of <span class="Chemical">RuO2 + Pt/C/CC was used as an air-breathing cathode (1.54 cm2). Typically, the NiCo2S4 NS/CC air cathode was prepared with a loading amount of 1.0 mg cm–2. The commercial catalyst-based cathode was fabricated by coating with 2 mg of RuO2 and 40 wt % Pt/C on CC to achieve a loading of 1.30 mg cm–2. A polished zinc plate with a thickness of 0.1 mm, 6 M KOH solution with 0.2 M zinc acetate, and a Whatman glass microfiber filter membrane were prepared as an anode, an electrolyte, and a separator, respectively, to assemble the zinc–air battery with a coin cell (MTI Korea) configuration. The specific capacity and the charge–discharge curves were reported with a battery analyzer (BST8-3), which consumed ambient air. The discharge capacity of the primary zinc–air cell was normalized to the consumed mass of Zn metal, whereas the current density was normalized to the area of the electrode.

Results and Discussion

The hierarchical bimetallic <span class="Chemical">sulfide NS arrays/CC hybrids were developed using an in situ two-step hydrothermal method, as graphically represented in Scheme . In the first step of the hydrothermal process, nickel nitrate hexahydrate and cobalt nitrate hexahydrate in a stoichiometric ratio were dissolved in deionized water and then urea was added. The solution was transferred into an autoclave and heated at 120 °C for 8 h, forming cobalt–nickel carbonate hydroxide hydrate NS arrays on a CC substrate. Subsequently, after oxidation reaction, a solution of sodium sulfide flakes dissolved in deionized water was prepared for the next sulfurization process. The anion exchange reaction from Co32–/OH– anions to S2– anions occurred at 160 °C for 6 h, thereby leading the complete phase transformation from cobalt–nickel carbonate hydroxide hydrate to nickel–cobalt sulfide on the CC. The morphology and composition of NiCo2S4 NS/CC were studied by FE-SEM and TEM. The bare CC consists of interconnected fibers with a smooth surface, as shown in Figure a. After the first hydrothermal reaction, numerous NS arrays are stacked onto the surface of CC with a rough surface shown in Figures b and S1. Consequently, the second hydrothermal process for sulfurization treatment is conducted, and its NS-like morphology perpendicular to a substrate with a rough surface still preserves its architecture (Figure e). From the top view of NiCo2S4 NS/CC, as shown in Figure c,d, the NiCo2S4 NSs that are interconnected with each other and form a porous architecture with submicron-size pores, providing ample space for the fast diffusion of redox ions during the reaction, uniformly cover the surface of CC. For comparison, the needle-like shape of NiCo2S4 vertically grown on the surface of CC with the average length of 3 μm is also synthesized to understand the effect of morphology on catalytic activities in both alkaline and neutral medium (Figure S2a–c). Also, the NiCo2O4 NS/CC was fabricated to have the same morphological characteristics as NiCo2S4 NS/CC (Figure S2d–f).

Scheme 1

Schematic Preparation Process of Self-Supported NiCo2S4 NS/CC Nanostructures

Figure 1

FE-SEM images of (a) bare CC, (b) NiCo-precursor/CC after the first step in the hydrothermal process, (c–f) NiCo2S4 NS arrays grown on the surface of CC at different magnifications.

The XRD patterns of NiCo2S4 NS/CC, <span class="Chemical">NiCo2S4 NN/CC, and NiCo2O4 NS/CC are presented in Figure a. After oxidation reaction in the first step of the hydrothermal process, nickel–cobalt carbonate hydroxide hydrate NS arrays are uniformly formed on the CC, as shown in Figure S1 (ICDD 00-040-0216). During the sulfurization in the second hydrothermal step, the nickel–cobalt carbonate hydroxide hydrate phase is completely transformed into spinel nickel–cobalt sulfide without destroying the original nanostructures. The diffraction peaks of NiCo2S4 NS/CC and NiCo2S4 NN/CC at 16.2°, 26.7°, 31.4°, 38.1°, 50.3°, and 55.1° are assigned to the (111), (220), (311), (400), (511), and (440) planes of cubic-phase NiCo2S4, respectively (ICDD 00-043-1477). The high-resolution TEM image in Figure b reveals the interplanar distance of 0.28, 0.23, 0.16, 0.33, and 0.18 nm, corresponding to the (311), (400), (440), (220), and (511) planes of NiCo2S4, respectively, confirming the successful formation of NiCo2S4. The formation of NiCo2S4 NS/CC is confirmed based on the energy-dispersive X-ray (EDX) using TEM, which shows that the atomic percentages of nickel, cobalt, and sulfur are 15.2, 30.7, and 54.1 at. %, respectively, as shown in Figure S3. Note that the two characteristic peaks at 26° and 43° for all prepared electrodes are attributed to the CC (ICDD 01-074-2329). The diffraction peaks of the NiCo2O4 NS/CC catalyst obtained after the first step in the hydrothermal process and heat treatment are consistent with the standard pattern of NiCo2O4 (ICDD 01-073-1702). Meanwhile, the NN-like morphology of NiCo2S4 NN/CC can be confirmed from the TEM elemental mapping images shown in Figure S4.

Figure 2

(a) XRD patterns of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC catalysts. (b) High-resolution TEM images of NiCo2S4 NS/CC. High-resolution XPS deconvolution spectra of the NiCo2S4 NS/CC catalyst for (c) Ni 2p, (d) Co 2p, (e) S 2p, and (f) C 1s.

(a) XRD patterns of NiCo2S4 NS/CC, <span class="Chemical">NiCo2S4 NN/CC, and NiCo2O4 NS/CC catalysts. (b) High-resolution TEM images of NiCo2S4 NS/CC. High-resolution XPS deconvolution spectra of the NiCo2S4 NS/CC catalyst for (c) Ni 2p, (d) Co 2p, (e) S 2p, and (f) C 1s.

To further characterize the chemical composition of electrodes, the XPS analysis is carried out, and results are given in Figure c–f. As shown in Figure c, the Ni 2p spectrum consists of two spin–orbit doublets of Ni2+ and Ni3+, including Ni2+ at 853.8 eV for Ni 2p3/2 and 873.8 eV for Ni 2p1/2, and Ni3+ at 857.7 eV for Ni 2p3/2 and 875.6 eV for Ni 2p1/2.[24,30,34,35] The XPS spectrum of Co 2p (Figure d) contains well-resolved peaks of Co2+ 2p3/2, Co3+ 2p3/2, Co2+ 2p1/2, and Co3+ 2p1/2 at 798.8, 793.9, 783.0, and 778.9 eV, respectively, implying the co-presence of Co2+ and Co3+ species in NiCo2S4 NS/CC.[23,34] The S 2p XPS spectrum (Figure e) shows two peaks at 163.0 and 161.8 eV, which are assigned to metal–sulfur bonds and the low coordination state sulfur ion that exists at the surface of NiCo2S4 NS/CC, respectively, with the satellite peak appearing at 170.1 eV in Figure e.[34,35] The C 1s spectrum of NiCo2S4 NS/CC is deconvoluted into four peaks located at 284.8, 285.5, 287.0, and 291.3 eV, which correspond to the C 1s orbital of C–C (sp2), C–C (sp), C–O, and π–π interactions, respectively (Figure f). The additional π–π interaction indicates the strong interactions between NiCo2S4 NSs arrays and the CC, which can minimize contact resistance to generate a direct electron pathway.[29] The binding energy values of Ni 2p, Co 2p, and S 2p are matched well with the previous reports on NiCo2S4-based materials.

The OER catalytic activity of NiCo2S4 NS/CC was first evaluated with a three-electrode setup using a low scan rate of 1 mV s–1 to eliminate the capacitive current effects in alkaline solution (1 M KOH, pH = 14). For comparison, NiCo2S4 NN/CC, NiCo2O4 NS/CC, bare CC, and RuO2/CC benchmarking OER catalysts were also tested under the same condition. All the synthesized electrodes in this work are directly used as free-standing oxygen-evolving electrodes, including conventional RuO2/CC, to avoid possible influencing factors. The LSV polarization curves and Tafel plots for all samples are revealed, as shown in Figure a,b. The NiCo2S4 NS/CC exhibits a superior catalytic activity toward OER with a low onset potential of only 180 mV. Moreover, the overpotential of 260 mV is required to generate 10 mA cm–2, which is smaller than that of NiCo2S4 NN/CC (316 mV), NiCo2O4 NS/CC (368 mV), RuO2/CC (322 mV), and bare CC (484 mV). Also, it requires only a 280 mV overpotential to afford 10 mA cm–2 when the scan rate is 5 mV s–1 (Figure S5). It is lower than those of many other reported nonprecious metal-based OER electrocatalysts tested under 1 M KOH conditions, such as NiCo2S4 NWs/graphdiyne foam (300 mV), NiCo2S4/NF (306 mV)—, NiCo2S4 NAs/CC (310 mV), and so on.[22−24,29,31,36−39] A detailed comparison is further summarized in Table S1. The cyclic voltammograms for NiCo2S4 NS/CC and NiCo2O4 NS/CC in the potential region from 1.0 to 1.8 V show a broad peak at 1.36 and 1.22 V, indicating the redox behavior of Ni2+ and Ni3+, respectively (Figure S7). The catalytic kinetics of OER is evaluated by Tafel plots in alkaline medium. The Tafel slope of NiCo2S4 NS/CC is 72 mV dec–1, which is lower than that of all other electrodes, such as RuO2/CC (87 mV dec–1), NiCo2S4 NN/CC (84 mV dec–1), NiCo2O4 NS/CC (114 mV dec–1), and bare CC (178 mV dec–1), indicating a more favorable rate of OER at the NiCo2S4 NS/CC electrode. The favorable kinetics of OER on NiCo2S4 NS/CC is also supported by EIS analysis to measure the charge transfer resistance during OER (Figure S6). The charge transfer resistance of NiCo2S4 NS/CC, which forms NSs morphology is 8.94 Ω at 1.5 V versus RHE, smaller than that of NiCo2S4 NN/CC with NN-like architectures (16.3 Ω). In contrast, NiCo2O4 NS/CC shows at least five times higher charge transfer resistance than that of NiCo2S4 NS/CC and NiCo2S4 NN/CC under the same applied potential because of the lower electrical conductivity of NiCo2O4 NS/CC.

Figure 3

OER in alkaline media (1 M KOH, pH = 14): (a) LSV polarization curves for OER and (b) Tafel plots of NiCo2S4 NS/CC, along with NiCo2S4 NN/CC, NiCo2O4 NS/CC, RuO2/CC, and CC electrodes for comparison. (c) Cyclic voltammogram measured in a non-faradaic region at various scan rates for the NiCo2S4 NS/CC electrode. The inset shows the plot of anodic and cathodic charging current density vs different scan rates. (d) Current density based on intrinsic catalytic activity vs voltage curves. (e) TOF and the specific (intrinsic) activities of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC electrodes at η = 350 mV. (f) Time dependence of the current density for NiCo2S4 NS/CC at a fixed potential of 1.55 V for 162 h.

OER in alkaline media (1 M KOH, pH = 14): (a) <span class="Chemical">LSV polarization curves for OER and (b) Tafel plots of NiCo2S4 NS/CC, along with NiCo2S4 NN/CC, NiCo2O4 NS/CC, RuO2/CC, and CC electrodes for comparison. (c) Cyclic voltammogram measured in a non-faradaic region at various scan rates for the NiCo2S4 NS/CC electrode. The inset shows the plot of anodic and cathodic charging current density vs different scan rates. (d) Current density based on intrinsic catalytic activity vs voltage curves. (e) TOF and the specific (intrinsic) activities of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC electrodes at η = 350 mV. (f) Time dependence of the current density for NiCo2S4 NS/CC at a fixed potential of 1.55 V for 162 h.

To better understand the different OER catalytic activities of NiCo2S4 NS/CC, including NiCo2S4 NN/CC and NiCo2O4 NS/CC catalysts, the ECSA and roughness factor (RF) of all electrodes are determined to estimate the real catalytic activities in the same pH condition. It can be easily calculated based on the double-layer capacitance (Cdl) through CV in a non-faradaic region at different scan rates of 1, 2.5, 5, 10, 20, and 50 mV s–1 (Figure c). The NiCo2S4 NS/CC electrode shows over twofold higher ECSA value of 5.5 mF cm–2 than that of NiCo2S4 NN/CC (2.0 mF cm–2) and NiCo2O4 NS/CC (2.3 mF cm–2), respectively (Figure S8). This indicates that plenty of catalytically active sites for OER might form on NiCo2S4 NS/CC.

The surface roughness for all electrodes was also calculated by dividing the estimated ECSA to the geometric area of the electrode, and the values of 27.5, 10, and 11.5 were achieved for each electrode, NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC, respectively. On the basis of these results, 2D NS architecture arrays can offer larger space and have a rougher surface; hence, they lead to more electrochemical active sites on the catalyst surface. As a result, the excellent electrocatalytic performances of the NiCo2S4 NS/CC electrode can be partially ascribed to the high ECSA and consequently highly exposed active sites.

We further calculate the TOF, which could provide the intrinsic OER catalytic activities of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC electrodes. The TOF of various electrocatalysts was derived using the following equation:J is the geometric current density at a specific overpotential. A denotes the geometric area of the electrode. The number of electrons consumed for generating 1 mol of O2 from water is 4. F is the Faraday constant value of 96 485 C mol–1. m denotes the mole numbers of active materials.[40] On the basis of this calculation, at an overpotential of 350 mV, the TOF of NiCo2S4 NS/CC and NiCo2S4 NN/CC is calculated as 7.46 × 10–2 and 4.12 × 10–2 mol O2 s–1, respectively (Figure e). In sharp contrast, NiCo2O4 NS/CC has the lowest TOF value of 1.19 × 10–2 s–1. Moreover, it is almost fourfold higher than that of the IrO catalyst (0.89 × 10–2 s–1), indicating that the NiCo2S4 NS/CC is highly efficient toward OER.[41]

The specific activity of catalysts with different surface areas or loading is calculated with current normalization by the catalyst RF. NiCo2S4 NS/CC can deliver a high specific current density of 2.67 mA cm–2, whereas NiCo2O4 NS/CC can produce only 0.96 mA cm–2 at the same overpotential of 350 mV. Even though their morphology appears to be similar, that is, the density and the distribution of NiCo2S4 NSs on the CC are virtually the same as those of NiCo2O4 NSs on CC, NiCo2S4 NS/CC outperforms NiCo2O4 NS/CC. This result shows that NiCo2S4 NS arrays on CC have a higher intrinsic OER catalytic activity than that of NiCo2O4 NS arrays on CC (Figure d), which can be explained by the difference in the crystal structure of NiCo2S4 and NiCo2O4. NiCo2S4 that formed closely packed arrays of large S2– anions with nickel and cobalt metal cations in different oxidation states occupying the tetrahedral and octahedral sites, respectively, possesses more octahedral active sites of Co(III) concerning NiCo2O4, which has smaller anions of O2– in the spinel structure.[31,42] On the basis of the previous literature, σ* orbital (eg) occupation–related metal cations at octahedral sites are mostly coordinated with electrocatalytic activities.[43] In view of this point, NiCo2S4 NS/CC might afford better OER intrinsic activity compared to NiCo2O4 NS/CC.

The electrocatalytic activity toward the OER of NiCo2S4 NS/CC is also evaluated in PBS (pH = 7) as well as control samples as shown in Figure a. Similar to the OER activity trend in alkaline media, the NiCo2S4 NS/CC catalyst exhibits the highest OER performance compared with other electrodes. Surprisingly, NiCo2S4 NS/CC requires only 321 and 402 mV to afford 5 and 10 mA cm2, respectively. However, RuO2/CC as a state-of-the-art OER catalyst needs an extremely large overpotential of 700 mV to deliver 5 mA cm–2 current density. At the same time, NiCo2S4 NN/CC and NiCo2O4 NS/CC require at least 368 and 460 mV to produce 5 mA cm–2, respectively. Bare CC shows negligible OER performance. The catalytic activity of NiCo2S4 NS/CC for OER in neutral media is exceptional compared to that of many electrodes reported recently, such as CoS4.6O0.6 (η = 570 mV for 5 mA cm–2),[19] ultrathin Co3S4 NS (η = 650 mV for 3.27 mA cm–2),[40] Co3O4 nanorod (η = 385 mV for 1 mA cm–2),[44] Co-Pi NA/Ti foam (η = 450 mV for 10 mA cm–2),[15] Co–Bi NSs/graphene (η = 570 mV for 14.4 mA cm–2),[18] and Fe–Ni–P (η = 429 mV for 10 mA cm–2).[14] The detailed comparison is summarized in Table S2. Figure b shows the Tafel plots of all electrodes for a better understanding of the obtained catalytic behavior. The Tafel slope of 123 mV dec–1 in a neutral electrolyte for NiCo2S4 NS/CC is achieved. It is the smallest value among NiCo2S4 NN/CC (125 mV dec–1) and NiCo2O4 NS/CC (203 mV dec–1) and comparable to that of RuO2/CC (115 mV dec–1), which in turn favors the kinetics of OER. Notably, in comparison with NiCo2O4 NS/CC, the NiCo2S4 NS/CC catalyst presents a lower Tafel slope value, originating from the increase in electrical conductivity as well as more plentiful electrocatalytic active sites correlated with its intrinsic activities.

Figure 4

OER in neutral media (phosphate buffer, pH = 7): (a) LSV polarization curves for OER. (b) Corresponding Tafel plots of NiCo2S4 NS/CC, along with NiCo2S4 NN/CC, NiCo2O4 NS/CC, RuO2/CC, and CC electrodes for comparison. (c) Cyclic voltammogram measured in a non-faradaic region at various scan rates for the NiCo2S4 NS/CC electrode. (d) Anodic and cathodic current density vs scan rate plot of NiCo2S4 NS/CC. (e) Mass activities and TOF of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC electrodes in PBS. (f) Time dependence of the current density for NiCo2S4 NS/CC at a fixed potential of 1.6 V for 11 h.

OER in neutral media (phosphate buffer, pH = 7): (a) <span class="Chemical">LSV polarization curves for OER. (b) Corresponding Tafel plots of NiCo2S4 NS/CC, along with NiCo2S4 NN/CC, NiCo2O4 NS/CC, RuO2/CC, and CC electrodes for comparison. (c) Cyclic voltammogram measured in a non-faradaic region at various scan rates for the NiCo2S4 NS/CC electrode. (d) Anodic and cathodic current density vs scan rate plot of NiCo2S4 NS/CC. (e) Mass activities and TOF of NiCo2S4 NS/CC, NiCo2S4 NN/CC, and NiCo2O4 NS/CC electrodes in PBS. (f) Time dependence of the current density for NiCo2S4 NS/CC at a fixed potential of 1.6 V for 11 h.

We further measure the double-layer charging of electrodes via scan-rate-dependent CVs to estimate the effective surface areas for catalytic activity. The potential range in which the non-faradaic region was chosen with the potential window of 0.04 V centered at an open-circuit voltage (OCV) of each system.[34] The electrochemical double-layer capacitance for NiCo2S4 NS/CC is 0.044 mF cm–2, whereas those for NiCo2S4 NN/CC and NiCo2O4 NS/CC are 0.025 and 0.023 mF cm–2, respectively, indicating the rougher surface of the NiCo2S4 NS/CC electrode. It is noticeable that NiCo2S4 NS/CC still possesses almost twofold higher electrochemical double-layer capacitance than NiCo2O4 NS/CC in neutral media, which is probably because of the well-aligned hierarchical NSs architecture and the formation of numerous electrochemically active sites. The TOF at the overpotential of 400 mV in neutral media is evaluated to compare the intrinsic activities of NiCo2S4 NS/CC with those of other comparison electrodes. The calculated TOF for NiCo2S4 NS/CC is 9.89 × 10–3 s–1, which is much larger than those for previously reported cobalt-based catalysts, including Co3S4 (1.32 × 10–3 s–1 at η = 500 mV),[40] Co–Pi (∼2 × 10–3 s–1 at η = 410 mV),[45] Co–Bi (1.5 × 10–3 s–1 at η = 400 mV),[46] and Co3O4 (≥0.8 × 10–3 s–1 at η = 414 mV),[40] further suggesting the remarkable OER catalytic activity of NiCo2S4 NS/CC under neutral conditions. The NiCo2S4 NN/CC for which NiCo2S4 NN arrays are grown on CC indicates a TOF of 4.83 × 10–3 s–1, implying its lower intrinsic activities compared with that of the NiCo2S4 NS/CC electrode, whereas the TOF of NiCo2O4 NN/CC is calculated as 2.47 × 10–3 s–1, which shows the lowest value among the three electrodes. Therefore, the remarkable electrocatalytic activity of NiCo2S4 NS/CC can partially originate from the higher electrochemical surface area and the direct contact between NiCo2S4 NS arrays and the CC, which facilitate fast electron transfer as well as enhanced mass transportation. In addition, (i) the intrinsic electrocatalytic activity of the NiCo2S4 with larger anions compared with NiCo2O4 so as to expose more cation active sites; (ii) the enough void space among interconnected NiCo2S4 NSs, which allows facile redox ion diffusion; (iii) the 2D morphology of NiCo2S4 NSs that yields a large contact area between the catalyst and the electrolyte; and (iv) the formation of the nickel–cobalt (oxy)hydroxide active layer on its surface, which will be discussed later, all contributed to the superb performance of NiCo2S4 NS/CC in the OER.

The long-term operation of the OER catalyst is a critical issue for practical application. Figure f exhibits the chronoamperometric (CA) curve of the NiCo2S4 NS/CC measured at 1.55 V potential in alkaline media (1 M KOH). After the 29 h CA test, the electrode entirely stabilizes and retains a current density of 10 mA cm–2 (without iR corrected) and then is reduced to 85% of its original activity over 160 h long-term operation. The inset of Figure f shows the linear polarization curves of NiCo2S4 NS/CC before and after 1000 CV cycles with 50 mV s–1 scan rate and a potential range from 1.2 to 1.7 V. Accordingly, there is no difference in the LSV curve recorded after 1000 CV cycles, indicating its high stability. Meanwhile, we detected losses of larger than 24, 29, and 47% in their current densities within 50 h for NiCo2S4 NN/CC, NiCo2O4 NS/CC, and RuO2/CC, respectively, at the same potential of 1.55 V in 1 M KOH solution. We also performed the durability test for same electrodes under the neutral condition. NiCo2S4 NS/CC presents excellent durability for 11 h, achieving 6 mA cm–2 at 1.6 V versus RHE with only 10% loss. During CA, O2 gas bubbles were visibly observed from the NiCo2S4 NS/CC electrode and dissipated quickly into the electrolyte. The NiCo2S4 NN/CC and NiCo2O4 NS/CC electrodes show a dramatic catalytic activity loss of 45 and 54%, respectively. Moreover, RuO2/CC almost lost its catalytic activity after 4 h durability test. The morphological robustness of NiCo2S4 NS/CC was examined by post-OER FE-SEM analysis under alkaline and neutral conditions (Figure S12). Maintaining morphology with negligible damage is another convincing evidence of the structural robustness of NiCo2S4 NS/CC observed in the FE-SEM micrographs.

Recently, Li et al.[46] reported that the nonoxide transition <span class="Chemical">metal-based chalcogenides, especially cobalt selenide catalysts, usually oxidize during the OER under the basic condition and progressively transform to the corresponding TM (oxy)hydroxides, which is proposed to be the true active species to catalyze the OER.[47] In the case of the Co3Se4/CF electrode, the XPS peak intensity of Se virtually disappears after a 3 h chronopotentiometric electrolysis duration, and after 12 h, Co3Se4 is converted to CoOOH. Similarly, in our study, we investigated the composition of the electrode after 160 h CA operation by XPS (Figure S11) to confirm the real surface species of NiCo2S4 NS/CC. The XPS Ni 2p shows that Ni2+ at 853.8 eV for Ni 2p3/2 and 873.8 eV for Ni 2p1/2 visibly disappears and the peaks located at 855.7 and 873.2 eV are assigned to Ni3+ species of the nickel (oxy)hydroxide.[31,48] Moreover, the binding energy shift of Ni 2p for 1.4 eV reveals the occurrence of electron transfer during extended CA electrolysis. Similarly, the XPS Co 2p3/2 peak is deconvoluted into two peaks of 780.7 and 782.3 eV, which represent the formation of Co(OH)2 and CoOOH, implying the formation of a higher valence state of cobalt (Co3+).[49,50] Meanwhile, the peak intensity of S 2p was weakened, whereas the two strong peaks of O 1s spectra were observed at 531.3 and 532.7 eV, indicating the O–H bond in NiCoOOH and the adsorption of H2O on the surface of NiCo2S4 NS/CC, respectively.[48−50] The XPS results demonstrate that in situ electrochemical tuning of nickel–cobalt sulfide to nickel–cobalt mixed (oxy)hydroxide phase occurred, which is highly active for the OER catalytic activity attributed to the enhanced surface area and electrochemically active sites. This transformation might change the electronic states and the interactions with intermediate products during OER. Hence, it leads to the catalyst becoming more catalytic active for OER, which is also shown in other chalcogenide materials.[46,50]

The post-OER durability measurement for over 11 h in the neutral medium was also carried out using XPS analysis to confirm the chemical composition (Figure S12). Similarly, in an alkaline environment, the Ni2+ peak disappeared from the surface of catalyzed <span class="Chemical">NiCo2S4 NS/CC and transformed to Ni3+ of TM (oxy)hydroxides with binding energy values at 855.7 and 873.2 eV as well as satellite peaks at 865.1 and 879.9 eV. Also, the new peaks formed at 779.9 and 781.1 eV are also assigned to the (oxy)hydroxides phase.[47−50] It is noticeable that similar phenomena of in situ electrochemical tuning for NiCo2S4 NS/CC have occurred under a neutral condition, achieving an increase in the surface area as well as electrochemically active sites for primarily improved catalytic activity for OER.

To validate the practical application of the NiCo2S4 NS/CC catalyst, a primary zinc–air battery was demonstrated and fully discharged to 0.6 V at a current density of 5 mA cm–2 (Figure a). The NiCo2S4 NS/CC cathode shows an OCV of 1.18 V with a specific capacity of 722 mA h g–1, which is almost 88.1% utilization of theoretical capacity (∼820 mA h g–1), whereas the commercial catalyst-based cathode shows an OCV of 1.31 V with a discharge capacity of 590 mA h g–1. Moreover, the galvanostatic discharge–charge cycling performance was evaluated at a current density of 5 mA cm–2 with a 5 min discharge followed by 5 min charge for each cycle (Figure b). For the initial cycle of NiCo2S4 NS/CC, the rechargeable battery discharged at 1.11 V versus Zn, with the corresponding charging potential of 1.90 V giving an overall overpotential of 0.79 V, which increased only 0.04 V (1.95 V for charge and 1.12 V for discharge potential) after 30 h battery operation (173 cycles). However, in the case of RuO2 + Pt/C/CC, the potential gap between charge and discharge increased continuously from 0.61 to 1.00 V even after 1350 min (135 cycles) cycling. The superior cycling durability over 173 cycles with a high discharge capacity of 722 mA h g–1 indicates the excellent electrocatalytic activity and stability of NiCo2S4 NS/CC for zinc–air batteries.

Figure 5

Zn–air battery performance: (a) specific discharge capacities of primary zinc–air batteries with NiCo2S4 NS/CC and commercial RuO2 + Pt/C/CC air cathodes. (b) Comparative galvanostatic charge–discharge profiles of rechargeable zinc–air batteries based on NiCo2S4 NS/CC and RuO2 + Pt/C/CC air cathodes at 5 mA cm–2 in 10 min interval cycles.

A two-electrode alkaline water electrolyzer was developed for full <span class="Chemical">water splitting with NiCo2S4 NS/CC and Pt/C/CC as the anode and cathode, respectively, in 1 M KOH solution (Figure ). To achieve a current density of 10 mA cm–2, a voltage of 1.53 V was needed for water splitting with gas evolution on both electrode surfaces, showing more advantages to split water than precious metal-based electrodes, which requires higher cell voltages of 1.64 V.

Figure 6

Overall water electrolysis: the polarization curves based on NiCo2S4/CC//PtC/CC and commercial RuO2/CC//Pt/C/CC electrodes with a scan rate of 5 mV s–1 in 1 M KOH solution. The inset is the photograph of the two-electrode configuration.

Overall water electrolysis: the polarization curves based on <span class="Chemical">NiCo2S4/CC//PtC/CC and commercial RuO2/CC//Pt/C/CC electrodes with a scan rate of 5 mV s–1 in 1 M KOH solution. The inset is the photograph of the two-electrode configuration.

Conclusions

In summary, the hierarchical spinel bimetallic <span class="Chemical">sulfide nanostructures in situ grown on the CC were investigated for their electrochemical properties in different pH media and evaluated for their capability in practical primary and rechargeable zinc–air batteries. The most active NiCo2S4 NS/CC electrode can catalyze the OER at an overpotential of 260 mV at 10 mA cm–2 with good durability of over 160 h operations under an alkaline condition. Moreover, the NiCo2S4 NS/CC electrode still maintained its superior OER catalytic activity under the neutral condition. The enhanced intrinsic catalytic properties, morphology-based advantages of nanostructures, and the generation of the Ni–Co oxyhydroxide active layer were considered responsible for the excellent OER performance in water splitting. Especially, the in situ fabricated NiCo2S4 NS/CC-integrated air cathode exhibits excellent durability and electrocatalytic activity in zinc–air batteries compared with the precious metal-based catalyst. This work supports a snapshot of the rational design and construction of nonprecious electrode materials with excellent catalytic activity and durability for the future practical system toward OER in water electrolyzers and zinc–air batteries.