Facile Synthesis of the Single-Atom Decorated Co_x-MoS₂/RGO Catalysts by Thermal-Annealing Vacancy-Filling Strategy for Highly Efficient Hydrogen Evolution

Abstract

A “thermal-annealing vacancy-filling” synthesis strategy was developed to engineer cobalt single-atom catalysts (Co-MoS₂/RGO) for exceptional hydrogen evolution reaction (HER) performance. By anchoring atomic Co onto Frenkel defect-engineered MoS₂ nanosheets supported by reduced graphene oxide (RGO), we achieved simultaneous optimization of catalytic stability, electrical conductivity, and active site accessibility. The optimized Co₃-MoS₂/RGO hybrid demonstrates remarkable alkaline HER activity, requiring only 94.0 mV overpotential to achieve 10 mA cm⁻² current density while maintaining excellent durability over extended operation. The atomically dispersed Co promoted HER kinetics through electronic structure modulation of MoS₂ basal planes, creation of catalytic active centers, and defect-mediated synergies. The RGO further contributed to performance enhancement by preventing nanosheet aggregation, facilitating charge transfer, and exposing active sites. This defect engineering strategy provides a facile method for developing cost-effective, stable, and high-performance electrocatalysts for sustainable hydrogen production.

1. Introduction

Green hydrogen, generated through renewable-powered water electrolysis with zero carbon emissions, has emerged as a pivotal eco-friendly alternative in sustainable energy solutions, particularly for decarbonizing industries such as steelmaking, ammonia synthesis, and long-haul transportation [1,2,3]. However, the sluggish kinetics of water decomposition, especially in alkaline conditions, present a significant bottleneck due to the additional energy required to overcome the high activation barrier for breaking H-OH bonds [4]. While precious metal-based catalysts, such as platinum (Pt) and Pt-based materials, have demonstrated high efficiency in water splitting, their high cost and limited availability hinder widespread practical applications [5,6]. To overcome these issues, researchers are increasingly focusing on alternative, cost-effective catalysts that offer high electrocatalytic performance [7].

Transition metal dichalcogenides (TMDs) have emerged as a highly investigated material class in electrocatalysis research, attributed to their cost-effectiveness, natural abundance, and superior catalytic performance in critical clean energy applications such as hydrogen evolution reactions (HER) [8,9,10]. Among TMDs, MoS₂ has shown promising electrochemical stability and catalytic performance [11,12,13]. The active sites for electrocatalysis in MoS₂ are typically found at the edge sites, where the coordination of atoms is unsaturated, whereas the basal planes remain catalytically inert due to their fully coordinated Mo-S bonds. As a result, pure MoS₂ often exhibits suboptimal HER activity [14]. This limitation has spurred innovative strategies to activate the basal planes, including defect engineering [15], phase transition [16], surface modification [17], and hybridization with other materials [18], aimed at optimizing the electrocatalytic efficiency of MoS₂ by strategically increasing the catalytic active surface sites.

Defect engineering, particularly the creation of point defects such as vacancies and interstitial atoms, has proven effective in enhancing the electrocatalytic properties of MoS₂ [19,20,21]. For instance, the introduction of Frenkel defects in monolayer MoS₂ has been shown to significantly improve HER performance, achieving a reduced overpotential of 164 mV at 10 mA cm⁻² [22]. Moreover, doping MoS₂ with heteroatoms such as Pd single atoms can activate inert atoms in the basal plane, further improving HER efficiency [23]. However, traditional single-atom synthesis methods are challenged by limitations such as low atomic loading density. Furthermore, the use of precious metal dopants often complicates synthesis and increases costs. Thus, there is growing interest in exploring non-precious metals for defect engineering in MoS₂ as a more cost-effective and scalable strategy.

In this work, we propose a “thermal-annealing vacancy-filling” strategy to synthesize single-atomically dispersed Co-doped MoS₂ catalysts. By creating Frenkel defects and filling Co atoms into MoS₂ matrix, we achieved targeted augmentation of the catalytically active sites of MoS₂ basal planes. To address nanoparticle aggregation, reduced graphene oxide (RGO) is employed as a support material, which simultaneously enhances charge transfer kinetics and surface dispersion stability [24]. Aberration-corrected scanning transmission electron microscopy (AC-STEM) coupled with synchrotron X-ray absorption spectroscopy (SXAS) confirmed the atomic dispersion of Co in MoS₂ matrix. Electrochemical tests in alkaline electrolytes show that the HER performance of the obtained Co-MoS₂/RGO catalyst is significantly improved.

2. Results and Discussion

Scheme 1 illustrates the key steps for synthesizing single-atom dispersed Co_x-MoS₂/RGO catalysts using the “thermal-annealing vacancy-filling” strategy. The process begins with the hydrothermal synthesis of MoS₂ flower-like nanoparticles (NPs), followed by thermal annealing to induce Frenkel defects in the basal plane of the 2H-MoS₂ sheets [22]. During thermal annealing under an argon atmosphere, Mo atoms migrate from their lattice positions, creating interstitial Mo atoms and Mo vacancies. Single-atomic Co atoms are then introduced to Frenkel defects and substitute the interstitial Mo atoms via the formation of S-Co-S bonds in the subsequent hydrothermal treatment in CoCl₂ solution. The concentration of the Co precursor solution is adjusted to control the amount of Co loading. An additional annealing step further promotes the creation of Frenkel defects and enhances Mo vacancies. This approach not only activates previously inert in-plane atoms, transforming them into active sites, but also stabilizes the Frenkel defects, which are crucial for boosting the overall catalytic performance [22,23]. Finally, the Co_x-MoS₂ catalyst is supported on RGO to prevent aggregation of Co-MoS₂ nanosheets, ensuring maximal exposure of active edges and basal planes. The resulting 3D structure of Co_x-MoS₂/RGO not only modulates the electronic and geometric structures but also prevents Co atoms from directly contacting acidic or basic solutions, thus protecting it from oxidation. Simultaneously, electron conductivity is maintained as electrons can efficiently travel through the MoS₂ sheets, resulting in outstanding electrocatalytic performance for water splitting. The Co doping levels were controlled by varying the concentration of CoCl₂ (0.25, 0.5, 0.75, and 1.0 mM) to yield Co₁-MoS₂, Co₂-MoS₂, Co₃-MoS₂, and Co₄-MoS₂, respectively. It should be pointed out that the same sample was synthesized more than three times, with no significant variations observed in the color, yield, or structural properties.

Figure 1A,B present representative scanning electron microscopy (SEM) images of MoS₂ and Co₃-MoS₂ NPs, revealing a flower-like spherical structure with an average diameter of approximately 313 nm, showing that Co doping does not significantly alter the morphology. Transmission electron microscope (TEM) images in Figure 1C further confirm that the Co₃-MoS₂ NPs consist of interconnected thin sheets. These petal-like sheets not only provide a high specific surface area but also expose abundant MoS₂ edges, which are typically the active sites for catalysis. Figure 1D shows the SEM image of Co₃-MoS₂/RGO NPs, where the wrinkled planar RGO supports the Co₃-MoS₂ NPs, preventing them from aggregating and restacking. No Co-related NPs or clusters are observed. To confirm the atomic dispersion of Co atoms, aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC-HAADF-STEM) was employed. The AC-HAADF-STEM image in Figure 1E shows faintly colored Co atoms distributed across the MoS₂ matrix; further, the corresponding intensity of the line scan confirms the single-atom dispersion of Co (Figure 1F). The atomic dispersion of Co was quantitatively analyzed using inductively coupled plasma (ICP), revealing a Co content of 3.21 wt% in Co₃-MoS₂. Table S1 in Supplementary Materials listed the weight percentages of the as-prepared Co_x-MoS₂ samples quantified by ICP analysis. When the Co loading increased to 4.80 wt% in Co₄-MoS₂, some Co clusters and aggregated particles were observed (Figure S1), indicating that excessive Co loading leads to aggregation. Thus, a Co loading of 3.21 wt% is optimal for achieving uniform Co atom dispersion. Figure 1G shows the energy dispersive X-ray spectrometer (EDS) mapping of Co₃-MoS₂/RGO, confirming the existence of Co and the homogeneous distribution of Mo, S, Co, and C elements across the Co₃-MoS₂/RGO. Additional EDS analysis (Figure S2 and Table S2) provides further details on the elemental composition of the as-prepared Co_x-MoS₂/RGO samples, confirming the expected molar percentages of Mo, S, Co, and C, which are consistent with the ICP analysis result.

Powered X-ray diffraction (XRD) patterns of the synthesized samples are presented in Figure 2A. The diffraction peaks of Co_x-MoS₂ are similar to those of pristine MoS₂, indicating that Co doping does not significantly alter the crystalline structure. The main peaks correspond to the (100), (111), and (122) characteristic planes, which match the hexagonal MoS₂ (2H-MoS₂, JCPDS card no. 37-1492). Slight shifts in the peak positions, such as a left-shift of the (002) peak at 14.1° and a weakening of the (103) peak at 39.7°, suggest that Co doping introduces minor distortions to the MoS₂ lattice, reducing crystallinity [25]. Importantly, no distinct diffraction peaks corresponding to cobalt containing phases are observed, indicating that Co element is present in an atomically dispersed form rather than as metallic clusters or compounds.

Raman spectroscopy was used to examine structural changes in the MoS₂ and Co_x-MoS₂ samples. Figure 2B shows that MoS₂ exhibits characteristic peaks at 375 cm⁻¹ (E_2g) and 402 cm⁻¹ (A_1g), confirming the existence of the 2H-MoS₂ phase [26,27]. The Raman profiles of the Co-doped MoS₂ samples remain largely unchanged, indicating that Co doping does not affect the crystalline structure of MoS₂. Notably, two characteristic peaks centered at 1346 and 1587 cm⁻¹ (Figure 2C), corresponding to RGO’s defect-induced (D band) and graphitic (G band) vibrational modes, are observed. With the increase of Co doping in the Co_x-MoS₂/RGO samples, the notably enhanced D/G intensity ratio (I_D/I_G) signifies increasing structural disorder within the graphene lattice, a critical feature that facilitates active site exposure through defect-mediated interfacial interactions while preserving charge transport pathways [21].

Figure 2D presents Fourier-transform infrared (FTIR) spectra of MoS₂ and various Co_x-MoS₂/RGO samples. The FTIR spectrum of MoS₂ shows a broad peak near 3435 cm⁻¹, attributed to the stretching vibration of adsorbed water molecules. In addition, two peaks at 1621 and 1189 cm⁻¹ correspond to the bending of adsorbed water and the asymmetric stretching of S=O bonds, respectively. These features suggest that water is present in the interlayer spaces of Co_x-MoS₂/RGO samples, which likely aids in its electrochemical activity [14].

X-ray photoelectron spectroscopy (XPS) was further employed to confirm the elemental composition and valence states of the as-synthesized Co₃-MoS₂/RGO. Figure 3A shows the XPS survey spectrum of the Co₃-MoS₂/RGO, confirming the presence of Mo, S, Co, C, and O elements across the sample. The high-resolution Mo 3d spectrum (Figure 3B) shows peaks corresponding to Mo⁴⁺ and Mo⁶⁺ states [28,29], while the high-resolution S 2p spectrum (Figure 3C) indicates the presence of S ions in MoS₂. The high-resolution Co 2p spectrum (Figure 3D) shows peaks for Co ions and their associated satellite peaks, without an observable zero valence Co-Co peak [30], confirming the formation of Co single atoms within the MoS₂ matrix. The high-resolution C 1s and O 1s spectra were shown in Figure S3, which indicates that a small amount of oxygen-containing groups are still retained on the surface of RGO, which is beneficial for the stable loading of Co_x-MoS₂ NPs.

The electronic configuration of Co species in Co₃-MoS₂/RGO catalyst was elucidated through SXAS spectroscopy. As depicted in Figure 4A, the normalized Co K-edge X-ray absorption near edge structure (XANES) profile of Co₃-MoS₂/RGO exhibits an absorption threshold energy of 7710.6 eV, positioned intermediately between Co foil (7705.4 eV) and CoO (7714.0 eV) reference standards. This characteristic pre-edge feature (1s→3d transition at 7700–7715 eV) demonstrates a 5.2 eV positive shift relative to Co foil while maintaining a 3.4 eV negative displacement compared to CoO, suggesting a mixed oxidation state regime spanning Co⁰ to Co²⁺. Figure 4B shows Fourier-transformed k²-weighted extended XAFS (EXAFS) spectra of Co₃-MoS₂/RGO in R-space that show notable differences compared to Co foil, CoO and Co₃O₄. A clear peak at ~1.64 Å in the R-space spectrum corresponds to the Co-S bond, while the absence of peaks at 2.19 Å (Co-Co) and 1.71 Å (Co-O) confirms that Co atoms are predominantly single-atomically within the MoS₂ matrix via Co-S bonds. Wavelet transforms (WT) of the Co K-edge EXAFS (Figure 4D) further support this, showing a signal center at 1.64 Å in R-space, corresponding to the Co-S scattering path. Quantitative EXAFS fitting was performed to determine the Co-S coordination structure, with the best-fitting model shown in Figure 4C. The results reveal a Co-S coordination number of approximately 3.5 (Table S3), suggesting Co atoms occupy trigonal prismatic sites within the MoS₂ lattice, consistent with theoretical predictions for highly active Co-MoS₂ configurations [25].

The electrocatalytic HER performance of the as-synthesized Co_x-MoS₂/RGO samples was systematically evaluated through comprehensive electrochemical characterization. All measurements were conducted using a standard three-electrode system in 1 M KOH electrolyte, with a saturated calomel electrode (SCE) as reference, a graphite rod as a counter electrode, and a catalyst-modified glassy carbon electrode (GCE) as working electrode. As shown in Figure 5A, the linear sweep voltammetry (LSV) curves demonstrate that Co-doped MoS₂/RGO hybrids exhibit significantly enhanced HER activity compared to pristine MoS₂. Particularly, the optimized Co₃-MoS₂/RGO catalyst requires only 94.0 mV overpotential to achieve a current density of 10 mA cm⁻², representing a 41% reduction compared to undoped MoS₂ (160 mV) and approaching the commercial Pt/C catalyst (~45 mV) [31]. The Co₃-MoS₂/RGO displayed a competitive HER performance compared with the recently reported Co-doped MoS₂ or MoS₂/RGO hybrids (Table S4) [32,33,34,35,36,37,38,39]. This remarkable enhancement could be attributed to the synergistic effects of single-atom Co doping-induced electronic structure modulation and RGO’s superior charge transport properties [14,22]. Replicate LSV scans were conducted three times to confirm the consistency and accuracy in representing the corresponding LSV curves. Further kinetic analysis through Tafel plots (Figure 5B) reveals that Co₃-MoS₂/RGO exhibits the lowest Tafel slope of 63.6 mV dec⁻¹, suggesting fast electron transfer kinetics during the HER process. Electrochemical active surface area (EASA) evaluation via cyclic voltammetry (CV) demonstrates that Co₃-MoS₂/RGO possesses the largest double-layer capacitance (C_dl = 166.4 mF cm⁻²) among the Co_x-MoS₂/RGO series (Figure 5C,D and Figure S4), indicating the high density of active sites for HER. The Co₃-MoS₂/RGO catalyst demonstrates the smallest overpotential and Tafel slope and the highest C_dl value, showing a faster electron transfer rate and suggesting that the rate-determining step is dominated by the Volmer-Heyrovsky pathway [40]. Notably, the Co₃-MoS₂/RGO catalyst demonstrates exceptional operational stability, retaining 88% of its initial activity after 3000 accelerated CV cycles (Figure 5E) and maintaining stable potential during 10 h chronopotentiometric testing (Figure 5F). This durability enhancement stems from RGO’s structural reinforcement and single-atom Co doping-induced stabilization of the MoS₂ matrix [14,23]. The systematic improvement in both activity and stability positions Co₃-MoS₂/RGO as a promising non-precious metal catalyst for alkaline water electrolysis applications.

The improved HER performance of Co_x-MoS₂/RGO can be attributed to several synergistic mechanisms: Firstly, the doping of Co induces band structure reconfiguration in MoS₂, narrowing the electronic bandgap and optimizing hydrogen adsorption free energy (ΔG_H*) to thermoneutral hydrogen adsorption free energy [41]. These states enhance charge transfer kinetics and reduce the energy barrier for proton adsorption and H₂ desorption; Secondly, the thermal-annealing vacancy-filling strategy can induce Frenkel-like defects (S vacancies and Mo interstitials), and then Co atoms replace Mo or occupy interstitial positions, which expose under-coordinated coordination sites, further increasing active site density. Thirdly, RGO’s high electrical conductivity mitigates the inherently poor conductivity of semiconducting 2H-MoS₂, facilitates rapid electron transfer between Co_x-MoS₂/RGO and the electrode, lowering overpotential, and improving HER kinetics. Moreover, RGO can disperse the Co_x-MoS₂ NPs and prevent their restacking and Co_x-MoS₂ corrosion. This maximizes the exposure of coactive sites, increasing the density of catalytically active centers.

3. Experimental

3.1. Materials

Reagents including cobalt chloride hexahydrate (CoCl₂·6H₂O, 99%), ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, 99%), thiourea (CS(NH₂)₂, 99%), and potassium hydroxide (KOH, 85%) were procured from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Natural graphite powder with a particle size of 44 µm was supplied by Qingdao Zhongtian Graphite Co., Ltd. (Qingdao, China). A 5% Nafion perfluorinated resin solution was acquired from Alfa Aesar (Shanghai, China). All chemical reagents were utilized as received without undergoing additional purification. In all experimental procedures, Milli-Q water with a resistivity of 18.2 MΩ·cm⁻¹ was used.

3.2. Synthesis of Flower-like Spherical MoS₂

Flower-like MoS₂ nanospheres were synthesized using a combined hydrothermal synthesis and thermal annealing approach. A 1:5 molar proportion of (NH₄)₆Mo₇O₂₄·4H₂O and CS(NH₂)₂ was first dissolved in 40 mL of ultrapure water under continuous stirring. The resulting homogeneous solution was transferred into a 50 mL polytetrafluoroethylene-lined autoclave, which was then subjected to hydrothermal treatment at 200 °C for 36 h. Following natural cooling to room temperature, the reaction mixture underwent centrifugation, followed by repeated washing with ultrapure water. The collected solid was dried under vacuum at 60 °C to obtain as-synthesized MoS₂. Subsequently, the MoS₂ product was annealed at 350 °C for 2 h in an argon atmosphere to induce the formation of Frenkel defects.

3.3. Synthesis of Co_x-MoS₂

To synthesize Co-doped MoS₂, the Frenkel defect MoS₂ was dispersed in a 40 mL aqueous solution of cobalt chloride (0.75 mM) and sonicated for 30 min. The mixture was then subjected to hydrothermal treatment at 200 °C for 8 h. After cooling, the product was collected by centrifugation, washed, dried, and annealed at 350 °C under argon. The Co doping levels were controlled by varying the concentration of cobalt chloride (0.25, 0.5, 0.75, and 1.0 mM) to yield Co₁-MoS₂, Co₂-MoS₂, Co₃-MoS₂, and Co₄-MoS₂, respectively.

3.4. Synthesis of Co_x-MoS₂/RGO

GO was synthesized using potassium permanganate oxidation of natural graphite following Hummers’ method [42]. GO (10 mg) was dispersed in 15 mL of water to form a colloid. The above Co_x-MoS₂ powder was dispersed in 15 mL of water and then added to the GO colloid and sonicated, followed by magnetic stirring for 1 h. 1 mL of NaBH₄ (0.15 M) was added to reduce the GO and form Co_x-MoS₂/RGO, which was then purified by centrifugation/wash cycles and dried under vacuum.

3.5. Characterization

The as-synthesized materials were analyzed using various techniques, including powder XRD (D8 Advanced, Bruker, Billerica, MA, USA), SEM (FEI NanoSEM450, Thermo Fisher Scientific, Hillsboro, OR, USA), TEM (JEM-2100, JEOL, Tokyo, Japan), HAADF-STEM (FEI Themis Z, Thermo Fisher Scientific, Hillsboro, OR, USA), FTIR (Nicolet Nexus 470, Thermo Fisher Scientific, Madison, WI, USA) spectroscopy, XPS (Thermo ESCALAB250, Thermo Fisher Scientific, Waltham, MA, USA), and ICP (VISTA-MPX, Agilent, Santa Clara, CA, USA) spectrometry.

3.6. Raman Measurements

Raman spectra were acquired using a Thermo Scientific DXR confocal microscope with a 532 nm diode-pumped excitation laser source (Thermo Fisher Scientific, Waltham, MA, USA). The excitation beam was focused onto the samples using an Olympus LMPlanFL 50× objective (0.50 NA, long-working-distance), yielding a spot size of approximately 2 µm. The laser power was maintained at 2 mW, and a grating with 1800 lines mm⁻¹ (2 cm⁻¹ spectral resolution) was employed for dispersion. Each spectrum was accumulated over three acquisitions (10 s integration time per scan).

3.7. Electrochemical Measurements

Electrochemical experiments were performed on a CHI 660E electrochemical workstation in 1 M KOH electrolyte (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). A glassy carbon electrode (GCE) loaded with the catalyst served as the working electrode, while a graphite rod and a saturated calomel electrode (SCE) functioned as the counter and reference electrodes, respectively. LSV and CV measurements were employed to assess HER activities.

4. Conclusions

In summary, we demonstrate a defect-engineering strategy, thermal-annealing vacancy-filling, to synthesize single-atomically dispersed Co-MoS₂/RGO catalysts. This approach enables precise anchoring of Co atoms at Frenkel defect sites within the basal planes of 2H-MoS₂, synergistically coupled with the conductive RGO. The optimized Co₃-MoS₂/RGO configuration achieves exceptional alkaline HER activity, requiring only 94.0 mV overpotential at 10 mA cm⁻², rivaling state-of-the-art noble metal benchmarks. More importantly, this methodology offers a facile, scalable synthesis route to develop cost-competitive, high-performance electrocatalysts for sustainable hydrogen energy systems.