Abstract

Facile template–free controllable growth of freestanding polyhedron–like CoS onto microporous Ni foam with three-dimensional architecture via a mild hydrothermal technique is reported. The as-obtained CoS catalyst phase was first tailored to N-Co9S8 (nitrogen doped Co9S8), and its inherent reaction kinetics and conductivity were then enhanced through sulfur incorporation via a hydrothermal process. The electrochemical performance of the pristine CoS and a sulfur-enriched N-Co9S8 (S, N-Co9S8) electrode in alkaline 1.0 M KOH was examined. The optimized polyhedral S, N-Co9S8 structured catalyst exhibits significantly enhanced electrocatalytic activity for both oxygen evolution reaction (OER) and hydrogen evolution reaction (HER). As a result, low overpotentials of 244 and −92 mV is required to achieve the current density of 10 mA cm−2 for the OER and HER, respectively. Furthermore, when the polyhedral S, N-Co9S8 catalyst was employed as a bifunctional catalyst in a two-electrode electrolyzer cell exhibiting a cell voltage of 1.549 V at 10 mA cm−2 and demonstrates excellent long-term (50 hrs.) chronopotentiometric electrolysis at various current rate, reveals excellent bifunctional OER and HER activities at different applied current densities. The superior OER and HER activities of the S, N-Co9S8 catalyst is result of the improved electronic conductivity and enhanced intrinsic reaction kinetics, which led to the enhanced electrocatalytically active sites after the incorporation of heteroatoms in the catalyst structure.

1. Introduction

Hydrogen (H2) is considered an alternative to traditional fossil fuels (e.g., coal, petrol, and natural gas) for various energy applications, and its use could limit greenhouse gas emissions caused by the use of fossil fuels [13]. Clean, renewable, and highly efficient H2 fuel can be easily produced through the sustainable ecofriendly electrochemical water splitting technique [4, 5]. An electrolysis of Earth-abundant water is a promising approach to generate pure oxygen (O2) and H2 gases without any carbon emission. These gases are generated by two main half-cell reactions, namely (1) oxygen evolution reaction (OER) and (2) hydrogen evolution reaction (HER), respectively. These reactions occur at the anode and cathode, both of which should be made of highly efficient and durable catalyst materials that have a small overpotential () in alkaline electrolyte media [68]. Although RuO2/IrO2- (OER catalyst) and Pt-based (HER catalyst) electrode materials have been considered as highly efficient, stable, active, and reliable. However, their low abundance, high cost, and material scarcity hinder their practicability in mass production [9, 10].

The overall performance of an electrolyzer cell primarily depends on the employed active OER and HER electrocatalyst [11, 12]. To date, various transition-metal-based electrocatalysts have been investigated as individual electrocatalysts for the OER and HER in alkaline and acidic media [1317]. However, their bifunctionality lags behind than that of the precious-metal-based catalysts due to inefficient kinetics in a single electrolyte medium [3]. For example, due to its high conductivity, intrinsic kinematics, robustness, mixed valence state, and efficient and durable electrochemical activity, Co9S8 is considered a promising electrocatalyst material [1315], but it suffers from poor electrochemical activity arising from its sluggish kinetics during catalytic OER and HER processes [1820]. The catalytic activity of an electrocatalyst is mainly determined by its electronic structure, which can be altered intrinsically via the addition of heteroatoms (e.g., N and S) to the host structure [21, 22]. The unfilled outermost 2p and 3p respective orbitals of N and S exhibit promising electron-withdrawing ability and high electronegativity due to the capture of electrons from adjacent Co atoms via acceptor/donor interactions [17, 21, 23]. This can lead to a change in the intrinsic electronic structure of Co9S8 and result in the optimal adsorption and dissociation of H2O molecules on the surface due to the reduced Gibbs’s free energy [3, 7]. Moreover, the addition of sulfur to the host structure minimizes the free energy of hydrogen adsorption and weakens the interaction between the anion and the metal on the catalyst surface [12, 23, 24]. Thus, the inclusion of metal heteroatoms with numerous free electrons in the structure of an active catalyst material would be a promising strategy for the improvement of the surface state density, which results in efficient and stable catalytic OER and HER activity [3].

Inspired by the above key consideration, herein, we report the design and synthesis of sulfur-rich nitrogen-doped Co9S8 (S, N-Co9S8), which can be used as a bifunctional electrocatalyst practically at wide industrial scale, prepared via a hydrothermal experimental procedure followed by nitrogen annealing with subsequent anion-exchange process. At first, a freestanding polyhedral-like CoS catalyst electrode film was prepared through a mild hydrothermal procedure. The structural phase of the pristine catalyst changed from CoS to S, N-Co9S8 upon treatment with N2 and subsequent anion-exchange procedure. The electrochemical performance of all prepared catalysts were then investigated in an alkaline electrolyte medium (1.0 M KOH). Figure 1 shows the experimental schematic and the steps involved for the fabrication of the proposed catalysts. The optimized S, N-Co9S8 bifunctional catalyst shows an excellent electrochemical OER and HER activities, achieves the low overpotentials of 244 and −92 mV, respectively, at a current density of 10 mA cm−2. Moreover, this electrocatalyst exhibits an elevated intrinsic activity and higher reaction kinetics compared to as-prepared CoS catalyst. The S, N-Co9S8 catalyst exhibits the pretentious electrocatalytic OER and HER performances comparable to other metal-sulfide-based catalysts reported previously in KOH electrolyte medium (summarized in Figure 2(a) and 2(b)) [2532]. Moreover, the S, N-Co9S8 bifunctional catalyst shows a static voltage response at different current rates and superior ultralong stability at a high current rate of 100 mA cm−2.

2. Experimental Section

2.1. Chemical and Materials

The analytical grade chemicals were used in the present work; those were purchased from Sigma Aldrich. Cobalt chloride hexahydrate (CoCl2·6H2O), sodium citrate tribasic dihydrate (C6H5Na3O7·2H2O), thioacetamide (C2H5NS), thiourea (CH4N2S), ruthenium oxide (RuO2), commercial 20 wt.% platinum on graphitized carbon (Pt/C), and potassium hydroxide (KOH) pellets were used without any further purification. Three-dimensional (3D) Ni foam (NF, sheet size:  cm, thickness: 1.6 mm, and cell size: 450 μm) was bought from Alantum Corporation (Korea) and used after cleaning with DI water, acetone, ethanol, and HCl several times.

2.2. Fabrication of CoS, N-Co9S8, and S, N-Co9S8 Catalyst Films

First, a freestanding polyhedral network of CoS was grown on a 3D NF substrate (size: ; exposed area: ) via a hydrothermal technique [33]. The active material phase was then transformed into N-doped Co9S8 (N-Co9S8) via annealing at N2 ambient. The N-Co9S8 was further subjected to anion-exchange process to obtain the S-rich N-Co9S8 (S, N-Co9S8). In a typical synthesis, an equimolar amount of CoCl2·6H2O and C6H5Na3O7·2H2O were dissolved at ambient temperature in a 50 mL aqueous solution of C2H5NS (18 mmol) under vigorous stirring for about 30 minutes (min). The mixture solution and precleaned NF substrate were transferred into a 0.1 L Teflon-lined stainless-steel autoclave. The autoclave assembly was placed in a muffle furnace. A hydrothermal experiment was performed for 6 hours (h) at 150°C with a ramping rate of 1°C/min. The CoS electrode film was collected from the autoclave, rinsed numerous times with deionized water and ethanol, and dried in a vacuum ambient. The N-Co9S8 and S, N-Co9S8 were then prepared from the as-obtained CoS film. The CoS electrode was annealed in a N2 environment for about 3 h at 700 K to obtain N-Co9S8 and subsequently an anion-exchange process was performed for about 10 h at 120°C in a 50 mL solution of 0.1 M CH4N2S to get the S, N-Co9S8 catalyst electrode [34].

2.3. Fabrication of RuO2 and Pt/C Catalyst Films

The RuO2 and Pt/C films were fabricated using a simple drop casting approach. Commercially obtained RuO2 or Pt/C powder was separately dispersed in 1.00 mL of ethanol containing 0.15 wt.% Nafion. The as-obtained solutions were sonicated for about 1 h, drop-casted on NF (), and vacuum-dried overnight.

2.4. Characterization Methods

The crystal phase and material structure of the catalyst electrode were investigated using X-ray diffraction (XRD; model: Rigaku Smartlab instrument). The catalyst electrode’s morphology and material/chemical composition were determined using field-emission scanning electron microscopy (FE-SEM) and energy-dispersive X-ray spectroscopy (EDS; Model: JSM–6701F) techniques, respectively. The elemental binding states and material chemical fingerprint were analyzed using X-ray photoelectron spectroscopy (XPS; model: ULVAC PHI 5000 VersaProbe) and the Raman spectra analysis techniques (model: LabRam Armis instrument, Horiba Jobin Yvon), respectively. High angle annular dark field (HAADF) scanning transmission electron microscopy (STEM) and elemental line scan mapping were carried out using a JEOL-2010 microscope at 200 kV with an X-ray EDS accessory.

2.5. Electrochemical Measurements
2.5.1. Catalyst Electrode Configurations for OER and HER

The electrochemical OER and HER performance of the prepared catalyst electrodes were examined using a conventional three-electrode electrochemical analyzer (VersaSTAT system, Princeton Applied Research) in 1.0 M alkaline KOH electrolyte. The prepared (CoS, N-Co9S8, and S, N-Co9S8) catalyst electrodes were used as the working electrodes. A platinum wire and a saturated calomel electrode (SCE) were served as the counter and reference electrodes, respectively. Cyclic voltammetry (CV) curves were recorded as a function of the scan rate () within a fixed non-Faradaic voltage window range to estimate the electrocatalytically active surface sites/area () as follows: [35] where (0.04 mF for KOH) is the specific capacitance of electrolyte and , , and are the double-layer capacitance, current density, and scan rate, respectively. Linear sweep voltammetry (LSV, polarization) and a chronopotentiometry test were conducted to analyze the potential response of the catalyst for different current densities. Electrochemical impedance spectroscopy (EIS) was used to understand the charge transfer kinetics of the catalyst electrode material at the electrode/electrolyte interface.

2.5.2. Bifunctional Electrolyzer Configurations

The bifunctional electrocatalytic performance of the optimized S, N-Co9S8 catalyst electrode was investigated using an electrochemical analyzer (VersaSTAT instrument) in a two-electrode configuration. The S, N-Co9S8 catalyst was served as both the anode and cathode in the electrolyzer cell and all the measurements were conducted in 1.0 KOH electrolyte.

3. Results and Discussion

The material phase and crystal structure of the CoS, N-Co9S8, and S, N-Co9S8 catalyst electrodes were examined using the XRD technique [36]. Figure 2(c) shows the XRD patterns of the prepared CoS, N-Co9S8, and S, N-Co9S8 species. The XRD spectrum of each catalyst shows distinct diffraction peaks, indicating that the good crystallinity of the cobalt sulfide phases formed. The diffraction peaks at the bottom (black spectrum) present at 30.89°, 35.76°, 47.43°, 55.00°, and 75.03° can be assigned to the (100), (101), (102), (110), and (202) planes of hexagonal CoS (JCPDS card No. 75-0605), respectively. The diffraction peaks at the top (green spectrum) located at 25.13°, 29.48°, 31.08°, 39.72°, 44.49°, 51.92°, 54.67°, 55.12°, 73.03°, and 76.36° can be indexed to the (220), (311), (222), (331), (422), (440), (531), (442), (731), and (800) planes of cubic Co9S8 (JCPDS card No. 86-2273), respectively. Both the N-Co9S8 (green spectrum) and S, N-Co9S8 (red spectrum) catalysts reveal similar reflection planes. However, the S, N-Co9S8 catalyst shows an apparent positive shift (29.48° to 29.56°) in the diffraction peak corresponding to the reflection of the (311) plane, indicating the successful incorporation of S atoms into the N-Co9S8 lattice, which results in a lower interplanar lattice spacing because of the smaller atomic radius of the dopant atom compared to the host material [37, 38]. The absence of other noticeable peaks indicates the phase pure formation of CoS and its transformation into S, N-Co9S8. Further, the Raman spectra of the CoS, N-Co9S8, and S, N-Co9S8 catalyst were recorded to elucidate the material fingerprint (Figure 2(d)). The pristine CoS catalyst (black spectrum) shows the characteristic Raman vibrational mode of Co-S (298 and 344 cm−1), Eg (474 cm−1), F2g (519 cm−1), and A1g (675 cm−1), whereas the N-Co9S8 (green spectrum) and S, N-Co9S8 catalyst (red spectrum) shows two Co-S vibrational modes similar to the pristine CoS catalyst, among the four Raman peaks [31]. However, these two initial peaks were slightly right-shifted (298 to 304 cm−1 and 344 to 351 cm−1, respectively) compared with the vibrational mode of CoS. The remaining two peaks originated from the S-S symmetric stretching (479 cm−1) and the lattice vibrations (614 cm−1) of Co-cations on the surface of S, N-Co9S8 [39]. The blue shift in the Co-S vibrational peak could be a result of induced compressive strain associated with compressive stress, which originates from the presence of heteroatoms in the unstressed crystal structure. This analysis is consistent with the XRD result.

FE-SEM images were recorded to investigate the effect of the material phase on the morphological structure of the prepared cobalt species. Figure 3 shows the FE-SEM images of the CoS and S,N-Co9S8 catalyst electrodes. For comparison, the morphology of the intermediate-state N-Co9S8 electrode film was also recorded (Figure S1, Supporting Information). Figure 3(a) shows the FE-SEM image of the CoS catalyst electrode, revealing an accumulated polyhedral structure grown vertically with several void spaces on the 3D NF substrate that might be due to the uneven growth of the active species in the hydrothermal process. These voids assist in the relaxed reversing channel for the electrolyte and lead to the increment in the number of active sites, which is helpful in improving the catalytic efficacy [40]. A well-defined flake-like characteristic embossed texture is observed on the surface of these polyhedral structures (Figure 3(b)). The flake-like texture is less prominent for the N-Co9S8 (supporting Figure S1) electrode film. Although, the S,N-Co9S8 catalyst electrode maintains similar morphological structure as that of the CoS. However, the surface texture and topography changed significantly. A rough embossed texture is observed on the polyhedral surface. In addition to material morphology, the constituent element’s chemical composition and their distribution for the CoS, N-Co9S8, and S,N-Co9S8 catalyst electrodes were analyzed by means of an EDS technique. Supporting Figure S2 shows the obtained EDS spectra and the mapped atomic chemical compositions of the constituent elements, respectively. The obtained chemical compositions are summed up in the Tables (inset of supporting Figures S2A, B), revealing the stoichiometric atomic percentage of all constituent elements in the prepared catalyst electrodes.

The local crystal structure, lattice orientation, and the presence of constituent elements at the atomic level were further confirmed by TEM, HR-TEM, and HAADF-STEM line scanning image analyses. Figure 4(a) shows the TEM image of the CoS which consists of numerous interconnected small particles (dark region) and voids (grey region), confirming its porous polyhedral architecture. The HR-TEM image (Figure 4(b)) reveals a random orientation of local planes, which are indexed in the magnified view. The interplanar lattice spacing (d-spacing) of and  nm are associated with the reflection of the (100) and (002) planes of hexagonal CoS (JCPDS card no. 75-0605), respectively. Moreover, the HAADF-STEM line scanning (Figure 4(c)) reveals the stoichiometric existence of the constituent Co and S elements. The size of the interconnected particles increases for the N-Co9S8 (Figure 4(d)) possibly due to the fusion of smaller particles during high temperature annealing. The N-Co9S8 reveals the d-spacing of and  nm (Figure 4(e)), which are associated with the reflection of the (440) and (331) planes of the cubic Co9S8 (JCPDS card no. 86-2273), respectively. The HAADF-STEM line scanning of the N-Co9S8 (Figure 4(f)) confirms the presence of N in the Co9S8 structure. The TEM image for the S,N-Co9S8 (Figure 4(g)) reveals that accumulated particles size in the polyhedral structure decreases significantly compared to the N-Co9S8. This could be possibly due to the recrystallization upon anion-exchange and is consistent with the deformed polyhedral structure observed in the FE-SEM image of the S,N-Co9S8 (Figure 3(c) and 3(d)). The N-Co9S8 also shows a clear lattice fringe pattern (Figure 4(h)) with the d-spacing of and  nm, which are related to the reflection of the (222) and (440) planes, respectively. Moreover, the HAADF-STEM line scanning of the N-Co9S8 (Figure 4(i)) further confirms the existence of Co, S, and N elements similar to the N-Co9S8 (Figure 4(f)) and reveals increased S content.

To further support the above mentioned XRD, Raman, FE-SEM, EDS, TEM, and HR-TEM analyses, XPS measurement for the CoS, N-Co9S8, and S,N-Co9S8 catalyst electrodes was performed. Figure 5(a) shows the full-range survey spectra of the prepared cobalt species. The additional C 1 s and O 1 s peak in survey are associated with the carbon contaminant inside the vacuum chamber and the adsorbed oxygen on the film from the environment [4144]. The Gaussian curve fitting model was used to fit the narrow range XPS emission spectra [45]. The Co 2p spectrum of the CoS catalyst electrode (Figure 5(b)) shows a total of four emission peaks at 781.31, 788.30 (satellite; Sat.), 797.30, and 804.29 (Sat.) eV. The low energy peaks (initial two) are related to Co 2p3/2, and the high energy peaks (later two) are associated with the Co 2p1/2 emissions of cobalt. The spin-energy separation of 15.99 eV (≈16.00 eV; Co 2p1/2–Co 2p3/2) reveal the presence of the Co2+ state [46]. The S 2p XPS spectrum (Figure 5(c)) of the CoS catalyst electrode shows two characteristic peaks at 162.80 and 163.99 eV after the deconvolution of a broad peak, which are associated with the characteristic S 2p3/2 and S 2p1/2 emissions. The spin-energy separation of 1.19 eV (S 2p1/2–S 2p3/2) assured the divalent state of sulfur (S2) bounded to Co2+ ions of cobalt in the CoS structure. The N-Co9S8 and S,N-Co9S8 catalyst electrodes also exhibits identical spin-energy separation for Co 2p and S 2p. However, the Co 2p and S 2p spectra were slightly right-shifted compared with the pristine CoS catalyst electrode (Figure 5(b) and 5(c)). This rigid binding energy shift for N-Co9S8 compared with CoS is associated with N-doping and the Co 2p and S 2p peak positions remains almost intact after the incorporation of S in the N-Co9S8 (i.e., the S,N-Co9S8 structure). The blue shift in the binding energy is a result of the lower concentration of electrons around the cobalt nuclei because of the higher electronegativity and stronger electron trapping capability of nitrogen compared with that of sulfur in the N-Co9S8 [37]. Moreover, the Co 2p spectra for the N-Co9S8 and S,N-Co9S8 catalyst electrodes reveal additional deconvoluted peaks originating from the Co-N and Co-O binding states compared with that of the CoS catalyst electrode [47]. Furthermore, the S 2p peak intensity of the N-Co9S8 catalyst electrode decreased slightly compared with that of the pristine CoS after high temperature annealing in ambient N2. This spectrum (Figure 5(c)) reveals an additional peak related to SOxn− that could have originated from the strong interaction of the adsorbed oxygen with sulfur in air to form sulfate [31, 48]. The SOxn− peak became less prominent after the anion-exchange process, whereas the intensity of the S 2p peaks significantly increased compared with the N-Co9S8 catalyst electrode, indicating the higher S content in the S,N-Co9S8 structure. In addition, the N-Co9S8 and S,N-Co9S8 catalyst electrodes shows an additional N 1 s peak in the survey spectrum (Figure 5(a)). The high-resolution N 1 s spectra (Figure 5(d)) exhibits a broad peak related to Co-N bonding within the Co9S8 structure originating from the replacement of S by the N dopant in the lattices [37, 49, 50]. XPS analysis further confirms the formation of the CoS structure and its complete transformation into N-Co9S8 and S,N-Co9S8.

The electrochemical OER and HER performance of the prepared cobalt species were examined thoroughly via LSV, with the chronopotentiometric voltage-step profile being a function of the current density, chronopotentiometric stability, and Nyquist impedance curve analyses. Figure 6(a) shows the obtained OER LSV curves that were recorded in a potential window range between 0.00 and 1.00 V (vs. SCE) for the CoS, N-Co9S8, and S,N-Co9S8 catalysts measured at 5.0 mV s−1 in alkaline 1.0 M KOH electrolyte. All of the obtained polarization curves were initially converted into the reversible hydrogen electrode (RHE) reference scale and then with the help of solution resistance () and by using the following equations: [31, 51, 52] where is the obtained potential in the reference SCE electrode scale and is the standard potential of the SCE at room temperature. The S,N-Co9S8 catalyst demonstrates excellent electrocatalytic OER activity by attaining a low overpotential of 244 mV (vs. RHE) at a driving current density of 10 mA cm−2, whereas the pristine CoS and N-Co9S8 catalysts attains an overpotential value of 292 and 259 mV at the same current density value, respectively. For comparison, polarization curves for the RuO2 and bare NF substrate (supporting Figure S3A) were also recorded at the same experimental conditions. An almost inactive OER activity was observed for the bare NF substrate, revealing a large overpotential of 442 mV at an applied current density of 10 mA cm−2. The enhanced electrocatalytic OER activity of the S,N-Co9S8 catalyst is aroused from improved conductivity and elevated electrocatalytically active sites, which results in efficient electron/ion transport and better intrinsic reaction kinetics (Figure S4–S7 and Table S1, Supporting Information). The electrocatalytic OER performance of the S,N-Co9S8 catalyst is relatively comparable to those of a cobalt sulfide-based catalyst reported in the literature (Figure 2(a) and Table S2).

The electrochemical reaction kinetics of the catalyst material was further assessed from the associated Tafel curve slope. Figure 6(b) shows the associated Tafel plots for the CoS and S,N-Co9S8 catalyst, which were obtained from the linear portion of the polarization curves. The Tafel curves were fitted using the Tafel equation as follows: [33, 5355] where , , and are the Tafel slope, current density, and fitting parameter, respectively. The S,N-Co9S8 catalyst reveals a smaller Tafel slope of 47 mV dec−1, which is considerably lower than the Tafel slope of the pristine CoS catalyst (70 mV dec−1), N-Co9S8 (66 mV dec−1), and RuO2 (65 mV dec−1) catalysts. The smallest Tafel slope value of the S,N-Co9S8 catalyst is indicative of the catalyst’s superior intrinsic reaction kinetics compared with the pristine CoS and the N-Co9S8 catalysts and is supported by the turnover frequency (TOF) analysis (Supporting Figure S6A). Generally, the OER process requires a four-electron transfer for the reaction mechanism, while the transfer of one electron occurred at each step, leading to sluggish reaction kinetics for the OER process. A Tafel slope of 40 mV dec−1 indicates that a two-electron transfer is the rate-determining step, a Tafel slope of 60 mV dec−1 is linked to the rate-limiting step that follows the first electron transfer, and a Tafel slope of 120 mV dec−1 is associated with the turnover-limiting single electron transfer process that occurs in the catalyst’s resting state [56, 57]. The variation in the Tafel slope with increasing potential could be a result of change in the potential response following the adsorption of reaction intermediates or a change in the rate-determining step [58, 59]. The S,N-Co9S8 catalyst reveals the smallest Tafel slope, which is crucial for efficient electrocatalytic OER activity. The low Tafel slope is most likely due to the second reaction step being rate-determining. Moreover, the low Tafel slope also indicates that strong surface intermediate adsorption during the initial step hindering the subsequent step, which turn out to be rate-limiting [59]. In contrast, the CoS and N-Co9S8 catalysts trails the rate-limiting step. The obtained electrocatalytic OER activity for the CoS, N-Co9S8, and S,N-Co9S8 catalysts were highly reproducible in independent and sequential catalyst test (Figure S8A, Supporting Information).

In addition to the low overpotential of the S,N-Co9S8 catalyst obtained at 10 mA cm2, a quantitative potential response analysis using the chronopotentiometric test was performed at low and high current rates. Supporting Figure S9A shows the potential-step profile being a function of the current density. The driven current density was kept constant for an interval of 20 min at each step, and a static voltage response was obtained at each applied current rate. A linear relationship was observed between “” and “” (Figure S9B, Supporting Information). The S,N-Co9S8 catalyst maintains a smaller potential response compared with the pristine CoS catalyst at each current rate. This is the indicative of efficient electron/ion transport throughout the porous polyhedral structure, robust adhesion, and excellent conductivity of the catalysts. In addition to the electrocatalytic OER activity, catalyst stability in alkaline conditions is also a crucial factor to be examined for the real time application of the active catalyst. Supporting Figure S9C shows the long-term chronopotentiometric stability for the S,N-Co9S8 catalyst measured at various current rates of 10, 50, and 100 mA cm2. The S,N-Co9S8 catalyst maintains a steady potential response during the continuous 50 h chronopotentiometric electrolysis after a minor potential loss of 0.37% in the initial stage. The increase in the potential may be due to the partial transformation of the S,N-Co9S8 surface into CoOOH during the electrooxidation process (reversible redox reaction; Eq. (4)–(6)), as evidenced by the presence of mixed Co2+/Co3+ oxidation states. This results in the increase in the O 1 s peak intensity, as shown in the post-stability XPS emission spectra (supporting Figure S10A-C) [4, 7, 10, 26]. The post-OER stability XRD spectrum (supporting Figure S10D) also exhibits two small additional diffraction peaks related to the CoOOH phase and this phase transformation is further supported by the HR-TEM data measured after the chronopotentiometric OER stability test (supporting Figure S11), revealing the reflection of the (012) plane associated with the CoOOH phase [60]. This transformation leads to a slightly deformed morphological structure, as shown in the FE-SEM images (supporting Figure S12) recorded after long-term chronopotentiometric electrolysis. Thus, based on the post-stability XPS, XRD, HR-TEM, and FE-SEM analyses, the S,N-Co9S8 core with a thin CoOOH shell is confirmed to be the electrocatalytically active phase during the OER process [4, 6, 26]. Nonetheless, an almost identical Nyquist impedance and polarization curves before and after the long-term chronopotentiometric electrolysis (Figures S7B, S9D, Supporting Information) confirms its excellent endurance throughout the stability test for the electrocatalytic OER in alkaline KOH medium.

Since the electrocatalytic HER performance is another important factor for evaluating the overall water splitting, the HER activity of each catalyst is assessed on the basis of cathodic polarization curve. Figure 6(c) shows the polarization curves for the CoS, N-Co9S8, and S,N-Co9S8 catalysts as well as for the bare NF substrate (supporting Figure S3C) and Pt/C catalyst measured at a scan rate of 5.0 mV s−1 in 1.0 M KOH electrolyte. The S,N-Co9S8 catalyst further reveals an excellent electrocatalytic HER activity, achieving a small overpotential of −92 mV (vs. RHE; ) at 10 mA cm−2 [61], which is quite comparable to the overpotential reported in previous literatures (Figure 2(b) and Table S3) and reproduceable within an adequate tolerable range (supporting Figure S8B). Notably, the overpotentials of 123, 114, and 342 mV are required to achieve the same current density for the CoS, N-Co9S8, and NF substrate, respectively. It is well known that the Pt/C is a state-of-the-art catalyst because of its electrocatalytic HER activity. For comparison, a commercially available Pt/C was used, which shows an overpotential of 13 mV at 10 mA cm2. However, at high a current value, the S,N-Co9S8 catalyst activity is competing well with the Pt/C catalyst, suggesting that the S,N-Co9S8 catalyst activity is more favourable for HER process. This could be linked to the efficient inherent kinetics, which is supported by the Tafel, ECSA-based LSV, and TOF curve analyses (Figure 6(d), Figure S6 and Table S4, Supporting Information). As shown in Figure 6(d), the S,N-Co9S8 catalyst reveals a Tafel slope of 107 mV dec1 (Volmer mechanism), which is less than those of the CoS (120 mV dec1) and the bare NF substrate (134 mV dec1), indicating the faster H2 evolution on the S,N-Co9S8 catalyst surface [33]. The superior HER performance of the S,N-Co9S8 catalyst is associated with the incorporation of S and N into the Co9S8 catalyst structure, which reduces the activation barrier and results in enhanced electrocatalytic HER activity [31, 33].

A multistep chronopotentiometric electrolysis was performed to elucidate the quantitative voltage response of the prepared catalysts. Supporting Figure S13A shows the measured multistep voltage profile measured as a function of the current rate. As expected, the S,N-Co9S8 catalyst shows a steady potential voltage response, revealing the linear potential response with an applied current density (“”; Supporting Figure S13B) and maintains the lowest voltage response at each driven current rate. Furthermore, the chronopotentiometry test was conducted to evaluate the endurance of the catalyst during the long-term stabilities. Supporting Figure S13C shows the chronopotentiometric electrolysis curve for the S,N-Co9S8 catalyst measured up to 50 h at various current densities. The S,N-Co9S8 catalyst shows static potential response at each applied current rate without underlying the material morphology and surface topography (Figures S12C, D, Supporting Information), revealing its excellent endurance throughout the chronopotentiometric electrolysis test, and demonstrating the excellent electrocatalytic HER activity in the 1.0 M KOH electrolyte medium [3, 4, 31]. This analysis is further supported by the Nyquist impedance and polarization curve recorded before and after the long-term chronopotentiometric stability (Figures S7B, S13D, Supporting Information).

Due to its excellent electrocatalytic OER and HER activity, the bifunctional activity of the S,N-Co9S8 catalyst was investigated in a 1.0 M KOH electrolyte using a two-electrode electrolyzer cell. The S,N-Co9S8 catalyst served as both the anode (OER catalyst) and cathode (HER catalyst) for the electrolyzer. Figure 7(a) presents the polarization curves for the bifunctional S,N-Co9S8||S,N-Co9S8 measured in a voltage window range of 1.1–2.2 V at a scan rate of 5.0 mV s1, producing an electrolyzer cell voltage of 1.549, 1.693, and 1.772 V at 10, 50, and 100 mA cm2, respectively. Figure 7(b) displays the comparative full-cell performance of our S,N-Co9S8||S,N-Co9S8 bifunctional catalyst and various metal sulfide-based catalysts reported at 10 mA cm2 in an alkaline KOH electrolyte (1.0 M) medium [25, 27, 28, 33, 62, 63]. The obtained bifunctional cell voltage was competitive with that of the cobalt sulfide-based catalysts. The reliability of the S,N-Co9S8||S,N-Co9S8 bifunctional catalyst was confirmed by evaluating the full-cell electrochemical activity under the same experimental conditions. The fabricated electrolyzer cell demonstrated electrolysis activity within a suitable range (Supporting Figure S8C). A linear relationship between the potential and current density was also observed in the voltage-step profile (Figure 7(c)), for the S,N-Co9S8||S,N-Co9S8 bifunctional electrolyzer cell. This electrolyzer cell also exhibited a steady potential response (~100%; Figure 7(d)), highlighting its long-term chronopotentiometric stability, as illustrated by the post-stability polarization curve (Figure 7(a)). The excellent endurance throughout the long-term chronopotentiometric test is a result of synergy and mutual interaction between the heteroatoms, varied electronic structure, robust adhesion, high mechanical strength, enhanced conductivity, and large catalytically active sites, which results in efficient electron/ion transport and improved electrochemical reaction kinetics compared to the as deposited catalyst material.

4. Conclusion

In summary, a cost-effective synthesis of unique freestanding polyhedral-like CoS topography with microporous architecture has been accomplished via a facile hydrothermal process. An active S,N-Co9S8 catalyst was obtained from the pristine CoS catalyst through nitrogen annealing followed by anion-exchange process, and it demonstrates superior electrocatalytic OER and HER activities in alkaline KOH (1.0 M) medium. The optimized S,N-Co9S8 catalyst exhibits excellent bifunctional electrocatalytic OER (244 mV) and HER (−92 mV) activities at 10 mA cm−2 with comparatively enhanced intrinsic reaction kinetics than that of the CoS catalyst. The catalytic activity of the S,N-Co9S8 catalyst is even compared with the RuO2 and Pt/C catalysts and high current densities. Moreover, S,N-Co9S8 bifunctional catalyst demonstrates a steady potential response and excellent stability endurance throughout 50 h for both OER and HER process at various current densities. The excellent electrocatalytic performance of the S,N-Co9S8 catalyst is aroused from the enhancement in the electron transfer and ion transport at the electrolyte interface along with improved material conductivity and enhanced electrocatalytically active surface sites, resulting in the elevated intrinsic reaction kinetics (Tafel, ECSA corrected LSV, and TOF analyses) after the heteroatom incorporation that could have changed the active catalyst surface chemistry. This universal approach could be helpful for the development of highly efficient nonnoble metal-based catalysts with excellent electrocatalytic activities.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare no competing financial interests.

Acknowledgments

This study was supported by the National Research Foundation of Korea (Grant nos. 2021R1A2B5B01001796, 2021R1A2C1006113, and 2021R1A4A5031805).

Supplementary Materials

Supporting Data Information (Figures S1-S13 and Table S1-S4) associated with this article can be found in the supplementary materials. (Supplementary Materials)