Two-Dimensional Molybdenum Diselenide Tuned by Bimetal Co/Ni Nanoparticles for Oxygen Evolution Reaction.

Herein, we report fabrication of MoSe2 functionalized with bimetal Co/Ni particles, which shows promising electrochemical performance in oxygen and hydrogen evolution reactions (OER and HER) due to its physicochemical properties such as electronic configuration and great electrochemical stability. We propose functionalization with two transition metals, cobalt and nickel, expecting a synergic effect in electrocatalytic activity in a water splitting reaction. These electrocatalytic reactions are essential for efficient electrochemical energy storage. The thin flakes were obtained by exfoliation of bulk molybdenum diselenide. Next, after deposition of metals, precursors were carbonized. Electrochemical data reveal that the presence of Ni and Co particles boosts electrocatalyst performance, providing a great number of active sites due to their conductivity. Interestingly, the material exhibited great evolution potential and good stability in long-term tests.

Introduction

Nowadays, we should place more and more emphasis and commitment to research related to energy storage and environmental protection. The production and use of hydrogen as an energy source seem to be one possible path for development. Hydrogen in the near future may replace the fossil fuels used so far and become the basis of a whole new industry related to clean energy. One of the most promising methods to obtain hydrogen and oxygen is the decomposition of water as a result of its electrolysis. The water electrolysis reaction consists of an oxygen evolution reaction (OER) along with a hydrogen evolution reaction (HER). The only limitation of these reactions in a potential application is the highly available efficiency of their performance. Currently, the most efficient catalysts for HER/OER are those based on platinum and its modifications with other materials.[1] Other, still very good examples are catalysts based on iridium oxide[2] and ruthenium oxide.[3] However, there are two obstacles to their commercial use: price and availability. Therefore, to generate cost-effective processes, it is necessary to develop new catalysts that will be efficient, cheap, and robust.[4] Potential candidates for such catalysts are sought among transition metals (like Ni, Fe, Co, and Cu),[5−7] their oxides, phosphates, sulfates[8,9] and dichalcogenides, first-row transition metal spinels,[10] perovskites,[11] and metal–organic frameworks (MOFs).[12,13]

Transition metal dichalcogenides (TMDCs) (like MoS2, WS2, MoSe2, WSe2, and MoTe2) are a promising group of electrocatalysts due to their layered structure, unique electronic configuration, and electrochemical stability. Particularly for selenides, there are special expectations because heavier chalcogens provide better electrical conductivity, which is very important for kinetics of HER/OER electrocatalysis.[14,15] Furthermore, the TMDC family of materials is characterized by the presence and availability of catalytic active sites at the edges of these structures.[16] Therefore, one of the methods to improve HER/OER performance is to create more catalytic active sites. A particularly interesting strategy for increasing the catalytic activity of the material is doping with other transition metals.[17,18] Such doping with homogeneous metal particles can improve the hydrogen adsorption free energy and the water adsorption/dissociation capabilities.[19]

To the best of our knowledge, molybdenum diselenide decorated with both cobalt and nickel nanoparticles as cocatalysts has not been reported yet. However, various samples with similar structures based on single transition metals and dichalcogenides or with similar architectures have been described as promising materials for different catalytic processes and in electrochemical systems for energy conversion. A few examples of the samples and their practical applications are summarized in Table . Xu et al.[20] have synthesized rose-like particles created with hexagon flakes as a catalyst in the process of thermal decomposition of ammonium perchloride. Another group (Zhao et al.)[21] has reported nanosheets created by bimetallic dichalcogenides (FeCoS2) examined as highly active catalysts in the dye photodegradation process. Furthermore, Caihua et al.[22] have studied CuS/graphene composites as an anode material in lithium ion-batteries. Other heterostructures based on dichalcogenides have been published by Zhao et al.[23] In this work, WS2–MoS2 composites have exhibited enhanced photocatalytic performance. These are only a few examples indicating the multifunctionality of TMD-based composites. Therefore, there is still plenty of room to reveal the full potential of the studied materials in a wide range of applications.

Table 1

Comparison of Applications and Geometry of Different Transition Metal Structures

sample	architecture	application	result	ref
Co3O4	nanoboxes	OER	overpotential, 358 mV (10 mA/cm2); Tafel, slope 56 mV/dec	(24)
Co3O4 (from β-Co(OH)2)	rose-like	catalyst	lower the thermal decomposition temperature of ammonium perchlorate from 450 to 245 °C	(20)
FeCoS2	nanosheets	(1) photodegradation; (2) Na-ion batteries	(1) degradation of methylene blue (MB) (95% degraded after 150 min UV irradiation); (2) first discharge capacity of 806 mAh/g at 50 mA/g	(21)
CuS/graphene	sheets with metal particles between	lithium storage	specific capacities of 568 mAh/g after 100 cycles at 50 mA/g and 143 mAh/g at 1000 mA/g	(22)
WS2/MoS2	flower-shaped structure (assembled by nanoplates)	photocatalytic	degradation of MB (complete degradation after 150 min); difference value of the photocurrent, 26 μA	(23)

In this contribution, we focus on the synergy of the two above-mentioned strategies, dichalcogenide (MoSe2) and transition metals (Co and Ni), to explore the synergic effect in their electrocatalytic performance. It could be expected due to the small size of metal particles, which are in close contact with MoSe2 sheets supporting electron and mass transport during the electrochemical reaction. As an effect, we expect boosting of electrical properties and exposure of more active sites for the oxygen evolution reaction.[25] Therefore, we tested the efficiency of the exfoliated molybdenum diselenide and their hybrids with cobalt/nickel, cobalt, and nickel as electrocatalysts in the OER reaction.

Characterization

Morphology, geometry, and structure of the materials were analyzed with transmission electron microscopy (FEI Tecnai G2 F20 based at 200 kV accelerating voltage) equipped with energy dispersive X-ray spectroscopy (EDS) and scanning electron microscopy (VEGA3 Tescan). Atomic force microscopy (AFM; Nanoscope V MultiMode 8, Bruker) provides information on the thickness and number of layers of exfoliated MoSe2. The crystallographic information regarding the as-synthesized samples was established by X-ray diffraction (XRD), performed using an AERIS PANalytical X-ray diffractometer with Cu-Kα radiation. Raman spectroscopy was applied (inVia Renishaw) for the characterization of chemical composition in the samples (laser, 785 nm).

Electrochemical Measurement

Electrochemical tests were performed by a potentiostat station (BioLogic VMP-3, France) in a three-electrode cell, and the temperature of the set was thermostated by a temperature control bath (Hubner KISS 6, Germany) at 25 ± 0.05 °C. Mercury oxide Hg|HgO with KOH (1 M solution) (MOE) was used as the reference electrode. A platinum wire with a surface area of ∼5 cm2 as the counter electrode was used. The working electrode was a 10 × 10 mm graphite foil (99.8%, GoodFellow) of 125 μm in thickness. On the surface of the working electrode, 30 μL of the active material with the concentration of 10 mg/mL was mixed with the solution of isopropanol/water containing 0.05% Nafion. Such a mixture was drop-casted on the working electrode and left to dry for 12 h in air. The measurements were performed in an alkaline electrolyte (1 M KOH) deaerated by purging argon (5.0 class) for 20 min. All measurements are acquired according to the reversible hydrogen electrode (RHE) potential calculated by the equation ERHE = EMOE + 0.128 + 0.059 × pH(V).

Results and Discussion

Successful preparation of few-layered MoSe2 samples was confirmed using AFM (Figure ). The number of layers was determined through analysis of the height profile. The flakes in bulk material were typically 21.37 nm in height, which corresponded to ∼25 layers. In the exfoliated sample, the height of the flakes was ∼4.89 nm, which is ∼5 layers of MoSe2 (assuming that the average thickness of a single layer is ca. 0.84 nm[26]).

Figure 1

AFM images of MoSe2: (a) bulk and (b) exfoliated; height profile of MoSe2: (c) bulk and (d) exfoliated.

The morphology and distribution of metal particles on MoSe2 were observed via electron microscopy tools. The results are present in TEM (Figure a–d) and SEM (Figure e–h) images of samples at different magnifications. All of them clearly show flake-like structures with irregular hexagon-shaped particles. The detailed analysis allowed us to estimate (i) the interplanar distance from the adjacent lattice fringes (0.32 nm), (ii) the average particle size distribution, which was 25.00 and 35.00 nm for Co and Ni, respectively, and (iii) the diameter of the flakes (indicated in SEM images with arrows), which was about 8.00 μm.

Figure 2

TEM images of MoSe2: (a) bulk, (b) exMoSe2, (c) exMoSe2@Co, and (d) exMoSe2@Ni; SEM images of MoSe2: (e) exMoSe2, (f) exMoSe2@Co, (g) exMoSe2@Ni, and (h) exMoSe2@CoNi; (i, j) TEM images of exMoSe2@CoNi (i) with elemental mapping (j).

Successful functionalization with metal oxide particles was also confirmed by EDS mapping, which is present in Figure j. It shows Mo, Se, Co, and Ni elements in mapping obtained from the exMoSe2@CoNi sample. It confirmed the successful modification of MoSe2 with metal particles. It is evident that they are distributed homogeneously on the surface.

The vibrionic properties of the samples were analyzed by Raman spectroscopy. The main modes of MoSe2 are in the range of 100 and 650 cm–1.[27] All of them are observed in each spectrum but with different intensities, which proves that there is a decrease in van der Waals forces between the layers in the exfoliated samples.[28] The vibrational modes of 2D-layered MoSe2 are observed at 143 cm–1 E1g modes, at 241 cm–1 A1g modes, at 286 cm–1 E2g mode, and at 450 and 586 cm–1 A1g modes, confirming that exfoliation by ultrasonication does not change the chemical structure of MoSe2.[29−31] The bulk sample exhibits high crystallinity indicated by highly intense vibration modes at 820 and 990 cm–1 representing A1g.[32]

The crystallographic composition of the analyzed samples is provided by X-ray diffractometry. Figure b,c shows the XRD patterns obtained from the pristine and exfoliated MoSe2. The peaks were referenced with JCPDS no. 29-0914.[28,33,34] XRD patterns of each sample contain peaks at 2θ values of 13.6°, 27.5°, 31.37°, 34.3°, 37.9°, 41.9°, 47.5°, 55.9°, 56.8°, 57.8°, 65.5°, 66.4°, 69.5°, 72.2°, 76.3°, and 83.8° corresponding to (002), (004), (100), (102), (103), (006), (105), (110), (008), (112), (114), (108), (203), (116), (205), and (118) in MoSe2, respectively. The lower intensity of the peaks in exfoliated samples confirms the efficiency of the exfoliation process, providing a larger interlattice distance in the few-layered structure. This is followed by peak broadening, meaning a decrease in vertical size of the structure.[35] The diffractograms of samples with metal particles in Figure c show characteristic peaks for Co and Ni,[32] which confirm successful functionalization. The peaks corresponding to metal nanoparticles are not intensive. It is due to the very small size of the Co and Ni nanoparticles.[36] The XRD pattern of exMoSe2@CoNi presents the highest intensity peaks that are related to the presence of both metals with a larger size.

Figure 3

(a) Raman spectra of all samples; XRD patterns of (b) MoSe2 bulk and exMoSe2 and (c) exMoSe2@Co, exMoSe2@Ni, and exMoSe2@CoNi.

The XRD patterns and Raman spectra confirm the Co, Ni, and Co/Ni functionalization of MoSe2. This effect decreases the overpotential value compared to pristine samples due to favoring of the oxygen evolution reaction.[37] The impact of the functional particles depends on their number and type,[38] which leads to different electroactivities.

Electrocatalytic performance strongly depends on the microstructure of the materials. The morphology and chemical composition have an impact on electrochemical properties. All samples, especially with both metal particles, have shown a distinct (002) peak, which indicates the existence of a few layers of sheet MoSe2, which is in agreement with AFM results. Intense peaks are due to the existence of a greater number of vacancies, which lead to an enhanced effect on the oxygen evolution, which is due to better electrochemical accessibility of reactive ions.[39]

The efficiency of the prepared samples has been tested as catalysts in OER processes performing well-defined measurements. First, the oxygen evolution reaction was investigated. The electrocatalytic performance of the obtained samples was examined by linear sweep voltammetry (LSV) measurement performed in a three-electrode cell in a 1 M KOH electrolyte with 5 mV/s sweep speed and corrected by an IR drop factor. The main purpose of the measurement is to determine the parameter describing the electrocatalyst activity, i.e., overpotential η value (mV). The overpotential value is described as the difference between the applied potential and theoretical potential that arises from the thermodynamics. Figure a presents the results of the LSV test for a commercial RuO2 catalyst and obtained materials. The evolution potential is most often measured at 1 and 10 mA/cm2 current density values, designated as η1 and η10, and investigated for comparison. The results for a commercial RuO2 catalyst and obtained materials are presented in Table . All samples of exfoliated MoSe2 modified with metal particles present better overpotentials at 1 and 10 mA/cm2 than bulk MoSe2 (444 and 556 mV, respectively) and pristine exfoliated MoSe2 (431 and 574 mV, respectively). The best results among the synthesized samples at a current density of 10 mA/cm2 were obtained for exMoSe2@Ni and exMoSe2@CoNi. They have even lower overpotential values at 10 mA/cm2 than the RuO2 commercial catalyst (380 mV). For a current density of 1 mA/cm2, all samples of exfoliated MoSe2 modified with metal particles present better overpotentials than the RuO2 commercial catalyst. The best result was obtained for the bimetal functionalized sample of exMoSe2@CoNi (−103 mV). The determination of the overpotential for the exMoSe2@Co sample was impossible due to the very strong oxidation.

Figure 4

(a) LSV plots, (b) Tafel plots, (c) Nyquist plots, and (d) chronopotentiometry plots of the obtained materials (OER).

Table 2

Overpotential at 1 and 10 mA/cm2 Current Densities, Tafel Slope, and R2 Equivalent Resistance Values of the Obtained Materials (OER)

	evolution potential at 1 and 10 mA/cm2
sample name	η1 (mV)	η10 (mV)	Tafel slope (mV/dec)	R2 equivalent resistance (Ω)
RuO2	278	380	136	6.4
MoSe2	444	556	144	5.3
exMoSe2	431	574	145	5.9
exMoSe2@Co		391	109	2.8
exMoSe2@Ni	180	367	227	3.1
exMoSe2@CoNi	103	378	170	4

The next measurement is steady-state chronopotentiometry, which is used to represent the Tafel dependency of η = f(log(j)). The lower Tafel slope value means a higher electrochemical reaction rate. In Figure b, Tafel plots of a commercial RuO2 catalyst and prepared catalysts are presented. Here, the exfoliated MoSe2 modified with Co particles presented a lower slope value (109 mV/dec) than bulk MoSe2 (144 mV/dec) and exfoliated MoSe2 (145 mV/dec). This value was even lower than for the RuO2 commercial catalyst (136 mV/dec).

Furthermore, the electrochemically active surface area (ECSA) of each catalyst was estimated by determining the double-layer capacitance of the system from CV measurement. They were performed at scan rates of 20, 50, and 100 mV/s in a non-Faraday potential range. This is typically a 0.1 V window of the open-circuit potential (OCP). The ECSA of the catalyst can be calculated by dividing the electrochemical double layer capacitance (CDL) by the specific capacitance (CS) of the sample as in the equation . The slopes of the differences at double-layer charging currents versus scan rate are equal to 0.0022, 0.0016, 0.0036, 0.0030, and 0.0025 mF/cm2 for MoSe2, exMoSe2, exMoSe2@Co, exMoSe2@Ni, and exMoSe2@CoNi samples, respectively. The double-layer specific capacitance of the flat surface in 1 M KOH electrolyte is 0.04 mF/cm2. Finally, the estimated ECSA values are 0.06, 0.04, 0.09, 0.08, and 0.06 of the studied samples, respectively. The lower value of ECSA indicates lower roughness of the outer surface of the catalyst[40] and the thinner layer is the catalyst layer. The thick layer of active material can cause local turbulences, which can disturb pure laminar flow. This can occur due to the non-uniform thickness of the layer. This disturbance is because the molecules of the products get to be part of the locally occurring turbulence and results in being yet again adsorbed at the catalyst active sites. That means that the thinner diffusion layer leads to a diffusion-limited value of current that is higher than the predicted one.[41]

Furthermore, to assess the quality of the catalyst activity, the electrochemical impedance spectroscopy (EIS) method was applied. The obtained results, presented as Nyquist plots, were fitted to equivalent circuit R1R2Q by a Z-fit subprogram of BioLogic potentiostat software. Figure c shows the EIS Nyquist plots for all tested materials. They have the comparable shape of the semicircles and the main difference is the size of the semicircle, which indicates a change in the R2 value. EIS analysis confirms the high activity of the synthesized materials. All samples functionalized with metal particles exhibit high activity with lower R2 values than the RuO2 commercial catalyst. The R2 equivalent resistance value is the smallest for exMoSe2@Co and is equal to 2.8 Ω, while the commercial catalyst RuO2 has a resistance value of 6.4 Ω.

Finally, despite high activity, the OER catalyst should have as well good stability at higher current densities and in a long time period. To reveal their performance, the chronopotentiometry method was used. The measurements were conducted at 10 mA/cm2 for 10 h followed by 20 mA/cm2 for the next 5 h. The chronopotentiometry plots are presented in Figure d. All the synthesized samples have similar stability as commercial RuO2 catalysts in the first 10 h. However, exMoSe2@CoNi has the lowest potential at this current density. At a current density of 20 mA/cm2, all the synthesized samples have a lower potential than the commercial RuO2 catalyst. Here, the exMoSe2@CoNi exhibited the best stability in chronopotentiometry long tests.

Hydrogen evolution reaction results were also investigated. All figures and values are placed in the Supplementary Information. All samples of exfoliated MoSe2 modified with metal particles present better overpotentials at 1 and 10 mA/cm2 than exfoliated MoSe2 without modification (343 and 488 mV, respectively). The best results among the synthesized samples were obtained for exMoSe2@Ni: 198 and 331 mV for η1 and η10, respectively. The exfoliated MoSe2 modified with metal particles presented lower slope values than bulk MoSe2 (173 mV/dec) and exfoliated MoSe2 (180 mV/dec). The exMoSe2@Ni has the lowest slope value among the synthesized samples (142 mV/dec), indicating the highest reaction rate. The exMoSe2@Ni has a similar R2 value (6 Ω) to the RuO2 commercial catalyst (5.4 Ω), while exMoSe2@Co presents the lowest R2 value (4.8 Ω). The most stable catalyst was based on Co and Ni particles present simultaneously on exfoliated MoSe2 (exMoSe2@CoNi). For both current densities, it has the lowest potential.

Conclusions

In conclusion, we successfully obtained MoSe2 with metal nanocomposites having a flake-like morphology, which exhibits great activity as an electrocatalyst, showing potential for application in electrochemical systems for energy conversion. Presented results make them great candidates to replace the expensive ruthenium oxide as an OER electrocatalyst. Molybdenum diselenide serves as both a matrix and conductive substrate. The reaction system has been investigated to compare the influence of metal particles and their combination on electrocatalytic activity. We have showed complete tests of different samples with MoSe2 (bulk, exfoliated, and with Co or/and Ni nanoparticles) toward a variety of electrochemical measurements in water splitting reactions. The experiments demonstrated a significant increase in electroactivity in samples with metal particles compared to the pristine material. Also, it can be concluded that the Co/Ni-functionalized sample is a superstable electrocatalyst outperforming commercial RuO2 catalysts. It can be explained by the synergy of the Co and Ni particles and the tubular structure, which lead to a higher density of the active sites on the surface of the catalysts. This composite can be practically applied due to its abundance, low cost, and stability in chronopotentiometry long tests. By combining MoSe2 with metal nanoparticles, a conductive network is obtained that enabled high accessibility of the active material in the OER process.

Experimental Section

Exfoliation of MoSe2

All chemical reagents were of analytical grade and were obtained commercially. Typically, MoSe2 powder (1 g, Sigma-Aldrich) was dispersed in 190 mL of N-methyl-2-pyrrolidone (NMP CHEMPUR) and 10 mL of H2O2 (CHEMPUR). Then, the dispersion was sonicated continuously for 2 h using a horn probe sonic tip. After ultrasonic treatment, the dark dispersion was centrifuged four times at 10000 rpm for 30 min and the supernatant was collected. The sample was dried 24 h at 60 °C.

Preparation of MoSe2/NiO, MoSe2/CoO, and MoSe2/NiO/CoO

Three samples of molybdenum selenide modified by metal oxide nanoparticles (MoSe2/NiO, MoSe2/CoO, and MoSe2/NiO/CoO) were prepared according to the following procedure: 150 mg of MoS2 and 150 mg of nickel (II) acetate tetrahydrate (product referred to as exMoSe2@Ni), cobalt(II) acetate (product referred to as exMoSe2@Co), and the mixture of both sources of metal (product referred to as exMoSe2@CoNi) were dispersed in 250 mL of ethanol and sonicated for 2 h. Afterward, the mixture was stirred for another 24 h. Finally, the sample was dried in high vacuum at 440 °C for 3 h. MoSe2/NiO and MoSe2/CoO were fabricated for comparative study.