In Situ Growth of MoS2 Nanosheet Arrays and TS2 (T = Fe, Co, and Ni) Nanocubes onto Molybdate for Efficient Oxygen Evolution Reaction and Improved Hydrogen Evolution Reaction.

Rationally designing efficient and low-price bifunctional electrocatalysts for oxygen evolution reaction (OER) and hydrogen evolution reaction (HER) are vitally important to bring solar/electrical-to-hydrogen energy conversion processes into reality. Herein, we report on a synthetic method that leads to an in situ growth of ultrathin MoS2 nanosheets and transition metal disulfide nanocubes onto the surface of Fe1/3Co1/3Ni1/3MoO4 nanorods for the first time. Such hybrids are found to serve as a bifunctional electrocatalyst with high activities for OER and HER, as represented by an impressive anodic and cathodic current density of 10 mA cm-2 at 1.53 and -0.25 V, respectively. More importantly, the performance for OER is even better than that of IrO2, the conventional noble metal electrocatalyst. These striking observations were interpreted in terms of the combination of strongly synergistic effect of multimetal components, large amount of exposed active site, and superaerophobia. The present methodology has been confirmed universal for synthesizing other molybdate solid solutions, which would open up new possibilities for designing novel non-noble bifunctional electrocatalysts for OER and HER.

Introduction

Electrochemical water splitting is widely considered as a promising and sustainable route for hydrogen production because electrical energy can be supplied from renewable energy resources such as wind turbines, hydropower, photovoltaic cells, and so forth.[1−3] Enhancing electrocatalytic activities of oxygen evolution reaction (OER) and hydrogen evolution reaction (HER), the half reactions of water splitting, is vital for solving environment and energy issues, particularly for OER.[4,5] At present, IrO2 and Pt/C are popularly used as the electrocatalysts for OER and HER, which however have the demerits such as high cost, scarcity of noble metals, and monofunction.[6,7] All these drawbacks make both IrO2 and Pt/C unsuitable in wide applications. Hence, developing bifunctional electrocatalysts with an attractive price and quality becomes more and more important in a clean energy field.

Until now, many strategies have been tried for the development of efficient bifunctional electrocatalysts with improved OER and HER activities. These strategies can be primarily divided into several categories: the first one is to enhance the synergistic effect between interfaces in multimetal components.[8−10] Zou and co-workers fabricated Ag/CoP core–shell heterogeneous nanoparticles which have shown an excellent OER performance.[10] The second one is to increase the number of active sites.[11−13] Xie and co-workers prepared defect-rich MoS2 and Ni(OH)2 nanosheets and achieved significantly improved HER and OER performances when compared to the defect-poor samples. The third one is to design superaerophobic electrodes that can accelerate gas evolution behavior and improve electrocatalytic performance at high reaction rates.[14,15] Jiang and co-workers fabricated a 3D-nanostructured Cu film for hydrazine fuel cells.[14] Given that all these strategies could improve the activity of OER and HER to some extent, we wonder whether remarkable catalytic activity can be obtained when we unify all these strategies in a material system.

Heterostructures, integrating distinct components with different functionalities in one system, give a chance to answer the above question because of their strong synergistic effect. However, most reported heterostructures were only two components as of now, which is not enough to settle this problem. Motivated by this, we attempted to initiate a strategy, in which MoS2 nanosheet arrays and transition metal disulfide nanocubes (TS2, T = Fe, Co, and Ni) were grown in situ onto the surface of transition metal molybdate solid solution. This novel strategy is expected to show merits of strong synergistic effect, large amount of exposed active sites, and superaerophobia.

To this end, for the first time, we report a synthetic method that leads to a highly efficient MoS2–TS2–transition metal molybdate electrocatalyst for OER and HER. The obtained material could be considered as an integrator of ultrathin MoS2 nanosheet arrays, TS2 nanocubes, and transition metal molybdate solid solution nanorods. We show that the integrator exhibits a high electrocatalytic activity toward both OER and HER, thanks to its special composition and structure. Most notably, by modulating the transition metal composition, an anodic current density of 10 mA cm–2 at an applied potential of 1.53 V is achieved, which is better than those of IrO2 and commercial noble metal electrocatalysts. Furthermore, the catalyst also achieved a cathodic current density of 10 mA cm–2 only at −0.25 V (vs RHE), which is superior to that of the MoS2 catalyst.

Experimental Section

Preparation of Materials

All chemical reagents used were of analytical grade without further treatment.

Synthesis of 1D Porous TMoO4 (T = Ni, Co, and Fe) Nanorods

Porous TMoO4 nanorods were synthesized via pyrolysis of the TMoO4·H2O precursor in air atmosphere.[16] The synthetic procedure could be briefly described as follows: 3 mmol sodium molybdate was dissolved in 30 mL of deionized water under vigorous stirring to get solution A. Transition metal nitrate or sulfate (3 mmol) was dissolved in 30 mL of deionized water to form solution B. Then, solution B was dropped into solution A under vigorous stirring. Finally, the resulting suspension solution was kept for heating at 150 °C for 12 h and cooled down naturally. The resultant powder was collected after centrifugation and washed several times with deionized water and ethanol and subsequently dried at 60 °C. To fabricate TMoO4 nanorods, the dried powder was moved to an alumina crucible and then heated to 500 °C at a heating rate of 1 °C/min and maintained at 500 °C for 2 h. The whole procedure was conducted under a constant N2 gas flow to maintain an inert atmosphere. For convenience, the product was marked as TMO (T = Ni, Co, and Fe). Specific names for samples are listed in Table .

Table 1

Names and Abbreviations of All Samples in This Worka

transition metal or ratios	TMoO4·H2O (TMH)	TMoO4 (TMO)	MoS2/TMS2/TMoOx (TMS)
Co	CMH	CMO	CMS
Ni	NMH	NMO	CMS
Fe/Co/Ni = 1:1:1	FCNMH	FCNMO	FCNMS
Ni/Fe = 4:1	NFMH	NFMO	NFMS
Ni/Co = 1:1	NCMH	NCMO	NCMS

T, F, C, N, and M represent transition metals, Fe, Co, Ni, and Mo, respectively; H represents hydration, O represents oxide, and S represents sulfidation.

Synthesis of 1D Porous TMoO4 (T = Ni/Co, Ni/Fe, or Ni/Co/Fe) Nanorods

The procedure for preparing 1D TMoO4 nanorods was similar to that for TMoO4, except for solution B of TMoO4 that contained given ratios of transition metal salts (Ni/Co = 1:1, Ni/Fe = 8:2, and Fe/Co/Ni = 1:1:1). The obtained sample was named TMO, for example, Ni0.5Co0.5MoO4 was marked as NCMO (see Table for details).

Synthesis of 1D and 2D Hybrids MoS2/TMO

MoS2/TS2/TMO hybrids were prepared via a simple low-temperature hydrothermal process. In a typical procedure, 150 mg of TMO or TMO nanorods and certain amount of l-cysteine (0.12, 0.25, 0.36, and 0.48 g) were dispersed in 60 mL of deionized water in a 100 mL Teflon-lined stainless-steel autoclave. The autoclave was heated at 200 °C for 18 h and then naturally cooled to room temperature. The resulting product was collected by a filter and washed by deionized water thoroughly. Finally, the black powder was dried under vacuum. For convenience, the powder was marked as TMS. T symbolizes transition metal, M is the initial letter of molybdenum, and S represents sulfidation.

Synthesis of MoS2 nanosheets and Reference Samples CoS2 and NiS2

MoS2 nanosheets were also prepared via a hydrothermal process. In a typical procedure, 1 mmol sodium molybdate and 5 mmol thioacetamide were dissolved in 60 mL of deionized water in a 100 mL Teflon-lined stainless-steel autoclave. The autoclave was heated at 200 °C for 18 h. Reference samples CoS2 and NiS2 were synthesized using a similar procedure, except that Co(NO3)2·6H2O or Ni(NO3)2·6H2O was used to substitute Na2MoO4.

Physical Characterization

X-ray diffraction (XRD) patterns of the samples were recorded on a Rigaku Miniflex apparatus equipped with a Cu Kα radiation source (λ = 1.5418 Å). Field-emission scanning electron microscopy (SEM) images and energy dispersive spectrometer (EDS) mapping data were performed on JSM-6700F. Transmission electron microscopy (TEM) measurements were carried out on Tecnai G2 S-TWIN F20. Scanning transmission electron microscopy (STEM ) measurements were performed on JEOL-2100F. Specific surface areas of the samples were measured with the Brunauer–Emmett–Teller equation, and pore size distributions were calculated via Barrett–Joynet–Halenda formula by a surface area analyzer ASAP 2020 (Micromeritics). X-ray absorption fine structure data were collected at the 1W1B station in Beijing synchrotron radiation facility.

Electrochemical Characterization

In a typical procedure to prepare the working electrode, each of the sample was dispersed in 1 mL of mix solution (containing 450 μL of H2O, 500 μL of ethanol, and 50 μL of 5% Nafion) via ultrasonic dispersion for 30 min to obtain a homogeneous ink. Then, 10 μL of ink was loaded onto a glassy carbon electrode (GCE) of 5 mm in diameter (loading amount: 0.20 mg cm–2). Prior to the coating of the catalyst, the GCE was polished with an Al2O3 paste (50 nm) and washed ultrasonically with deionized water. Then, the catalyst-modified GCE was dried at room temperature.

Electrochemical measurements were performed in a typical three-electrode cell using a CHI 760E workstation at room temperature. OER and HER were tested under different conditions. For OER, a graphite rod and Hg/HgO electrode were used as the counter electrode and the reference electrode, respectively. Linear scan voltammetry (LSV) was carried out at a scan rate of 5 mV s–1 in 1 M KOH, and all data were revised by Ohmic potential drop (iR) correction. The potentials were displayed versus RHE by RHE calibration: E (RHE) = E (Hg/HgO) + 0.098 + 0.059 × pH. The stability tests were performed in 1 M KOH at room temperature by potential cycling between 1.25 and 1.7 V (vs RHE) at a sweep rate of 50 mV s–1 for 3000 cycles. Cyclic voltammetry (CV) measurements with different scan rates (40, 80, 120, 160, and 200 mV s–1) were used to determine the electrochemical double layer capacitances (EDLC, Cdl). The electrochemical impedance spectroscopy (EIS) measurements were carried out at an overpotential of 300 mV in the frequency range of 10–2 105 Hz. Prior to all the above measurements, oxygen was bubbled into the electrolyte for 30 min.

For HER, a graphite rod and saturated calomel electrode (SCE) were used as the counter electrode and the reference electrode, respectively. LSV was carried out at a scan rate of 5 mV s–1 in 0.5 M H2SO4, and all results were revised by iR correction. The potentials were displayed versus RHE by RHE calibration: E (RHE) = E (SCE) + 0.241 + 0.059 × pH. Prior to all the above measurements, nitrogen was bubbled into the electrolyte for 30 min.

Results and Discussion

MoS2–TS2–TMoO hybrids were initiated to be prepared via a two-step method (shown in Figure ). In step 1, transition metal molybdate hydrates were pyrolyzed under an inert atmosphere to release H2O, and then porous transition metal molybdate nanorods were obtained. In step 2, porous TMO nanorods were sulfided by l-Cys in a hydrothermal environment. During the sulfidation procedure, S2– anions dissociated from l-Cys might react quickly with metal ions (transition metal or Mo ions) on the surface to form a thin layer metal disulfide.[17] This thin layer metal disulfide could play a key role in obstructing further reaction between inner metal ions and outside S2–.

Figure 1

Schematic illustration for the preparation of TMS hybrids via a two-step method.

By optimizing the experiment parameter (Figures S1 and S2), we chose to add 0.36 g of l-Cys in the synthesis of our catalysts. Composition, pore size distributions, and specific surface area of the as-synthesized samples were measured by XRD patterns and nitrogen adsorption–desorption (Figures S3 and S4).

SEM images in Figure illustrate the evolution process from precursors (NCMH, NFMH, and FCNMH) to the corresponding sulfide samples (NCMS, NFMS, and FCNMS). After calcinations at 500 °C, NCMO, NFMO, and FCNMO maintained the nanorod morphology of their precursors, probably due to the slow heating rate. The diameter of NCMO is about 500–1000 nm, whereas the diameter for NFMO and FCNMO is about 0.5–2 μm. By comparing the data in Figure S5, it is found that sizes of NCMO, NFMO, and FCNMO were clearly larger than those of CMO and NMO. That is to say, the size of a nanorod can be tuned by doping a transition metal. Figure c,f,i shows the morphologies of NCMS, NFMS, and FCNMS, respectively. Obviously, after sulfuration, NCMS, NFMS, and FCNMS still showed a rodlike morphology. Moreover, an ultrathin nanosheet array and some cubes (white circle in Figure c,f,i) were grown onto the surface of molybdate. It is reasonable that the nanosheet array is MoS2, whereas the nanocube is TS2 (T = Fe, Co, and Ni), as deduced by the layered structure characters of MoS2 and cubic phase structure of TS2. The composition of TS2 was further verified by STEM and corresponding EDX element mapping (Figure S7).

Figure 2

SEM images for precursors (a) NCMH, (d) NFMH, and (g) FCNMH; for unsulfide (b) NCMO, (e) NFMO, and (h) FCNMO; and for sulfide (c) NCMS, (f) NFMS, and (i) FCNMS. White cycles in (c,f,i) are assigned to TS2.

The morphology and structure of FCNMS, NCMS, and NFMS were further investigated by TEM and HRTEM (Figures and S6). The TEM image (Figure a) shows that the ultrathin nanosheets were wrapped onto an FCNMO nanorod, showing a thickness of less than 5 nm, and that the nanocubes have a lateral size of about 200 nm. Careful data examination indicates that the nanosheets appeared to be stacked loosely and partially curved. The edges were approximately about 5–6 layers. The interlayer distance of nanosheets is about 0.63 nm, which can be indexed to the (002) plane of MoS2. Further, HRTEM of nanocubes shows a lattice distance of 0.25 nm, corresponding to the (210) plane of TS2.

Figure 3

(a) TEM and (b,c) HRTEM of FCNMS.

EDS mapping was used to examine the dispersion of elements for samples FCNMS, NFMS, and NCMS (Figures and S8). After sulfidation, S element is evenly dispersed onto the sample surface. Moreover, it is found that the O element is still uniformly dispersed in the core of FCNMS, which demonstrates a core–shell structure of FCNMS (core FCNMO and shell MoS2/TS2). Also as shown in Figure S8, Co, Ni, Mo, O, S and Ni, Fe, Mo, O, S were uniformly dispersed on nanorods of NCMS and NFMS, respectively.

Figure 4

Elemental mapping images of FCNMS sample: (a) SEM image, the mapping of (b) Fe (cyan), (c) Co (purple), (d) Ni (blue), (e) Mo (yellow), (f) O (red), and (g) S (green), and (h) the mixed image of all elements.

It is worth noting that CMS and NMS could not maintain the rod morphology (Figure S5). There are two possible reasons. First, CMO and NMO nanorods are too thin to maintain the rodlike morphology. Comparatively, once doping other transition metal ions into CMO and NMO, such as FCNMO, the nanorods became thicker (Figures and S5), which may be in favor of maintaining the rod morphology. Second, doping transition metals may change the exposed crystal plane of precursors and thus be conducive to epitaxial growth of MoS2. To verify this, TEM and HRTEM were used to monitor the exposed plane for FCNMO (Figure a,b).

Figure 5

(a) TEM and (b) HRTEM images of FCNMO, (c) schematic diagram for the growth mechanism of FCNMS. The gold and blue arrows in (c) mean that the transition metals (Fe2+, Co2+, and Ni2+) and molybdate ions in FCNMO were sulfided to TS2 and MoS2, respectively.

TEM analysis in Figure a demonstrates the rod characteristic of FCNMO, consistent with the observations by SEM (Figure ). The distance between lattice fringes marked in Figure b is about 0.67 nm, corresponding to the plane (001) of monoclinic CoMoO4 (JCPDF no. 21-0868). Similarly, the lattice distance for NCMO and NFMO was 0.65 nm (Figure S9), very closer to 0.62 nm for the plane (110) of NiMoO4 (JCPDF no. 33-0948). The measured lattice fringes for three samples NCMO, NFMO, and FCNMO matched well with that of 0.63 nm for the plane (002) of MoS2 nanosheets (Figure b). Comparatively, CMO and NMO exhibited a lattice spacing of 0.28 and 0.34 nm, respectively (Figure S10), which could be assigned to the planes (201) and (130). Hence, doping transition metals into molybdate is beneficial for the exposure of planes (001) and (110). Previous literature has established that the well-matched lattice distance could facilitate the epitaxial growth of two different materials.[18] Accordingly, the mechanism of sulfidation can be illustrated in Figure c. That is, during the sulfidation procedure, S2– anions dissociated from l-Cys may react quickly with surface Mo atoms to form MoS2 nanosheets. Then, these MoS2 nanosheets would grow along the plane (001) of FCNMO because the lattice distance of 0.63 nm for the plane (002) of MoS2 is closer to that of 0.67 nm for the plane (001) of FCNMO. On the other side, a part of transition metals (Fe, Co, and Ni) would react with S2– to generate the TS2 nanocubes on the surface of molybdate.

XANES were measured to examine the valence states of Mo ions in FCNMS. As shown in Figure a, the absorption edge of FCNMS shifted toward lower energies relative to that of FCNMO, indicating that partial Mo6+ ions in FCNMO were reduced and reacted with S2– to generate MoS2 nanosheets. The amount of reduced Mo ions can be estimated by linear combination fitting from Mo K-edge XANES of FCNMS, FCNMO, and MoS2. The data fit result was shown in Figure S11, where about 52% of Mo6+ ions were reduced to Mo4+. Hence, the average valence of Mo ions in FCNMS is about 4.96 (Figure b).

Figure 6

(a) Normalized Mo K-edge XANES of FCNMS, FCNMO, and MoS2 and (b) average valence states of Mo ions in FCNMS, FCNMO, and MoS2.

Electrocatalytic activity of FCNMS for OER was studied using a typical three-electrode electrochemical cell in 1 M KOH. Figure a compares their polarization curves (sample loading was set at 0.2 mg cm–2) with a scan rate of 5 mV s–1. FCNMS and FCNMO achieved a current density of 10 mA cm–2 (a current density for 12% efficiency of a solar to fuel conversion device) at 1.53 and 1.56 V, respectively. Further, FCNMS exhibited a large anodic current density of 206 mA cm–2 at 1.7 V, about 2 times larger than that of FCNMO. Apparently, the catalytic performance of FCNMO is substantially enhanced after sulfuration. More importantly, the catalytic activity of FCNMS is even better than that of IrO2, a benchmark noble metal electrocatalyst for OER. Such a performance observed for FCNMS is also better than the previous reports for most of the transition metal-based OER catalysts.[6,13,19−25]

Figure 7

(a) Polarization curves of given samples for OER; (b) corresponding Tafel plots derived from (a); (c) EIS Nyquist plots of the given samples at an overpotential of 320 mV; and (d) initial polarization curves and after 3000 CV scans for the FCNMS sample. Inset of (d) is the time-dependent current density curve at a static overpotential of 300 mV.

To understand better the electrocatalytic performance of the sample FCNMS, one has to take FCNMO, CoS2, NiS2, mixture of FCNMO and CoS2, and mixture of FCNMO and NiS2 as the contrast samples. As shown in Figure a, the catalytic activity of FCNMS is superior to that of other samples, indicating the presence of a strong interaction between TS2 and FCNMO. The data in Figure a show that the activities of mixtures of FCNMO and TS2 (T = Co or Ni), CoS2, and NiS2 do not surpass those of FCNMO. Therefore, it could be reasonable that the molybdate solid solution plays a dominant role in the catalytic reaction of the FCNMS composite. In addition, to gain insights into the oxygen evolution activity, Tafel plots of various catalysts were compared as shown in Figure b. The Tafel slope of FCNMS hybrids is 48 mV dec–1, much closer to the value obtained for IrO2, but apparently lower than those for FCNMO, CoS2, and NiS2, implying that FCNMS hybrids can give rise to a high OER rate in practical applications.[26,27]

The EDLC were measured to compare the active area of electrocatalysts (Figure S12). The values of Cdl for FCNMS are 3.43 mF cm–2, slightly larger than those for other samples, which further indicates that the BET surface area is not a primary factor to catalysis. EIS data of the samples are comparatively studied, as shown in Figure c to understand the electrode reaction kinetics. The equivalent electric circuit of the cell and the fitting results are shown in Figure S13 and Table S1, respectively. The larger semicircle in the low-frequency region corresponds to the electrical models of the anodes. Rct,2 is related to the electrocatalytic kinetics, with a smaller value representing a faster reaction rate.[28] The smallest Rct,2 for FCNMS indicates a faster charge transport among all the studied counterparts.

The long-term stability of the electrocatalysts is also a critical factor for applications in practical devices. Accordingly, the stability of FCNMS-modified electrodes was tested by continuous cycling for 3000 cycles (Figure d). After cycling for 3000 times, the polarization curve was slightly decreased in the anodic current. Although the overpotential at about 10 mA cm–2 was slightly increased from 300 to 318 mV, FCNMS still achieved a current density of 187 mA cm–2 at 1.7 V (vs RHE), and about 91% activity was maintained. Additionally, the current–time plots at a fixed potential were also collected, and the anodic current almost maintained a constant value around 10 mA cm–2 over 10 h, further confirming an excellent stability of the FCNMS-modified electrode.

Meanwhile, the effect of catalyst loading on the OER activity was examined (Figure S14), which indicated that the measured activities were in the regime of intrinsic kinetics. The above results clearly demonstrate that FCNMS possesses an excellent catalytic performance for OER. This observation could be attributed to their distinctive structural advantages: (i) the synergistic effect of multimetal components and strong interaction between FCNMO and TS2 in FCNMS, (ii) the mesoporous structure of FCNMS gives more active sites for the catalytic reaction, as reported elsewhere;[20,29,30] and (iii) the nanosheet array on the surface of nanorods is suitable for desorption of the gas products, thus giving a much improved electrocatalytic performance.[14] It is worth noting that NFMS and NCMS (the vulcanized products of NFMO and NCMO) also exhibited a distinctly improved OER activity (as shown in Figure S15). To further illustrate the advantage of our catalysts, we compared the overpotentials of FCNMS and NFMS at 10 mA cm–2 with the overpotential of the recently reported electrocatalyst (as shown in Figure S16). FCNMS has shown the lowest overpotential, demonstrating a best OER performance.

In addition to the marvelous OER activity, FCNMS has a possibility to show good catalytic activity for HER because a large amount of MoS2 have exposure edges that could become the active sites for HER. Figure a exhibits the polarization curves of various samples for HER. All the vulcanized samples exhibit a better HER activity than MoS2. Especially, FCNMS achieved a cathode current density of 10 mA cm–2 at −0.25 V (vs RHE), superior to that of other samples. Moreover, FCNMS also shows the smallest Tafel slope of 74 mV dec–1 among all samples, indicating the fastest HER rate.[7]

Figure 8

(a) Polarization curves of various samples of HER and (b) the corresponding Tafel plots derived from (a).

Conclusions

In summary, a universal synthetic route has been reported for in situ growth of ultrathin MoS2 nanosheet arrays and transition metal disulfide nanocubes onto the surface of molybdate solid solutions. Such MoS2–TS2–TMoO hybrids have shown merits of strong synergistic effect among components, mesoporous property, and superaerophobia. These merits allow these hybrids to show an impressive anodic and cathodic current density of 10 mA cm–2 at 1.53 and −0.25 V (vs RHE), respectively. The findings reported in this work mark an important step toward the development of non-noble metal electrocatalysts for OER and HER with high efficiency.