Interfacial engineering of cobalt sulfide/graphene hybrids for highly efficient ammonia electrosynthesis

Significance

Ammonia is one of the most important chemical raw materials with an annual production exceeding 200 million tons. The Haber–Bosch process is still the dominant route for industrial ammonia synthesis, which consumes 1∼3% of global annual energy production and represents a significant contributor to climate change. Electrocatalytic N₂ reduction reaction is an attractive alternative candidate for carbon-free and sustainable NH₃ production, but often suffers from low efficiency. Here, we developed an interfacial engineering strategy for preparing a class of strongly coupled hybrid electrocatalysts for N₂ fixation. The hybrids exhibit superior N₂ reduction reaction activity with a high NH₃ Faradaic efficiency of 25.9% under ambient conditions. This strategy provides an approach to design advanced materials for ammonia production.

Abstract

Electrocatalytic N₂ reduction reaction (NRR) into ammonia (NH₃), especially if driven by renewable energy, represents a potentially clean and sustainable strategy for replacing traditional Haber–Bosch process and dealing with climate change effect. However, electrocatalytic NRR process under ambient conditions often suffers from low Faradaic efficiency and high overpotential. Developing newly regulative methods for highly efficient NRR electrocatalysts is of great significance for NH₃ synthesis. Here, we propose an interfacial engineering strategy for designing a class of strongly coupled hybrid materials as highly active electrocatalysts for catalytic N₂ fixation. X-ray absorption near-edge spectroscopy (XANES) spectra confirm the successful construction of strong bridging bonds (Co–N/S–C) at the interface between CoS_x nanoparticles and NS-G (nitrogen- and sulfur-doped reduced graphene). These bridging bonds can accelerate the reaction kinetics by acting as an electron transport channel, enabling electrocatalytic NRR at a low overpotential. As expected, CoS₂/NS-G hybrids show superior NRR activity with a high NH₃ Faradaic efficiency of 25.9% at −0.05 V versus reversible hydrogen electrode (RHE). Moreover, this strategy is general and can be extended to a series of other strongly coupled metal sulfide hybrids. This work provides an approach to design advanced materials for ammonia production.

Ammonia is not only regarded as a promising chemical hydrogen storage candidate due to its high energy density but also an important raw material for the fertilizer, pharmaceutical, and light industries (1–3). Currently, industrial NH₃ synthesis still relies on the traditional Haber–Bosch process, which must be performed at high temperatures (350∼550 °C) and high pressures (150∼350 atm) with high-purity streams of nitrogen and hydrogen as raw materials (4, 5). This process accounts for 1∼3% of global annual energy production and represents a significant contributor to climate change (6, 7). Thus, there is an urgent need to explore new technologies for green and efficient NH₃ synthesis. The electrochemical N₂ reduction reaction (NRR) enables NH₃ synthesis using N₂ and H₂O under ambient conditions, making it a highly promising candidate for clean, carbon-free, and sustainable NH₃ production (8, 9). However, the sluggish kinetics of N₂ absorption and the permanent dipole of the triple bond seriously impede its catalytic efficiency (10). Therefore, the discovery of low-cost and highly efficient NRR catalysts for NH₃ electrosynthesis is of great significance.

To date, a series of noble metal [Au (11, 12), Pd (13, 14), Ru (15, 16)] and nonnoble metal electrocatalysts [metal oxides (17, 18), sulfides (19), nitrides (20, 21), and heteroatom-doped carbon materials (22–24)] has been pursued for NRR. However, these catalysts still suffer from low NH₃ yield and selectivity due to the small number of catalytically active sites and poor charge-transport capacity (11, 25, 26) Recently, interfacial engineering by combining electrocatalysts with hierarchical substrates such as TiO₂, graphene, and CeO₂ nanofibers has been regarded as an efficient way to protect growing nanoparticles from agglomeration and supply more active sites, thereby enhancing NRR catalytic activity (11, 14, 25). However, most of these hybrid catalysts lack strongly coupled bonds at the interface between the substrates and electrocatalysts, which still severely limits their interfacial charge-transfer capacity, restricting the scope of promoting the electrocatalytic activity for NRR.

Recently, numerous regulating strategies, such as designing 2D/2D hybrid materials to improve the contact interface (27) or directly constructing strong bridging bonds between nanoparticles and supports (28), have been applied in developing highly active catalysts for oxygen evolution and reduction. Among these strategies, the construction of strong bridging bonds at an interface would be a more effective way to improve catalytic activity because these bonds can directly accelerate reaction kinetics by functioning as an electron transport channel. Herein, we highlight an interfacial engineering strategy for strongly coupled hybrid materials as highly active NRR electrocatalysts. Remarkably, the strong Co–N/S–C bridging bonds enable a controllable interfacial coupling effect between the cobalt sulfide nanoparticles and graphene. As expected, the strong-coupled CoS₂/NS-G catalyst exhibits excellent NRR performance with a high NH₃ yield rate (25.0 μg h⁻¹⋅mg⁻¹_cat at −0.2 V vs. RHE) and Faradaic efficiency (25.9% at −0.05 V vs. RHE) under ambient conditions. Moreover, a series of other strongly coupled hybrid materials (MS_x/NS-G; M = Ni, Fe, and Sn; x = 1, 2) was also developed for NRR by using this general approach. This work paves a route for developing advanced electrocatalysts for ammonia synthesis.

Results and Discussion

Characterizations of CoS_x/NS-G Hybrid Materials.

In this work, strong-coupled CoS_x/NS-G hybrids were synthesized by an in situ annealing method that enables the uniform growth of cobalt sulfides nanoparticles on a graphene sheet. The stoichiometric ratio of CoS₂ to CoS can be easily regulated by controlling the amount of thiourea powder added (SI Appendix, Fig. S1). The detailed structural features of the products were investigated by X-ray diffraction (XRD). All of the different diffraction peaks can be well indexed to standard cubic CoS₂ [Joint Committee on Powder Diffraction Standards (JCPDS) card no. 70–2866] and hexagonal CoS (JCPDS card no. 70–2864) without any peaks for impurities, indicating the successful preparation of different phases of CoS₂ and CoS samples (SI Appendix, Fig. S3). Moreover, transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) were further conducted to evaluate the morphology and composition of the CoS₂/NS-G and CoS/NS-G samples. As shown in Fig. 1A and SI Appendix, Fig. S4A, both the as-prepared CoS₂ and CoS sample have morphologies of small nanoparticles uniformly grown on the graphene sheet. The size distribution of the resulting CoS₂ nanoparticles was also characterized. As shown in SI Appendix, Fig. S17 C and D, most sulfide nanoparticles are between 20 and 60 nm. The metallic nanoparticles can also been tuned by employing different amounts of cobalt salt as raw material. SI Appendix, Fig. S17 shows TEM images and particle-size distribution of CoS₂/NS-G hybrids synthesized by different amounts of cobalt salt. CoS₂/NS-G synthesized by 0.1, 0.2, and 0.4 mmol CoCl₂ exhibit particle size of 10–30, 20–60, and 100–250 nm, respectively. The above result reveals that the particle size of the cobalt sulfide increases with the increasing of the adding amount of cobalt salt, and the composition of metallic nanoparticles can be tuned by controlling the synthetic raw material. Moreover, the HRTEM image shows a lattice fringe of 3.2 Å (Fig. 1A, Inset), which is consistent with the (111) lattice plane of cubic CoS₂, while the (002) plane of the hexagonal CoS sample can also be identified by clear lattice fringes of 2.6 Å (SI Appendix, Fig. S4B). Furthermore, energy-dispersive X-ray spectroscopy elemental mapping analysis of CoS₂/NS-G and CoS/NS-G shows the homogeneous distribution of Co, C, N, and S in the as-prepared hybrids (Fig. 1B and SI Appendix, Fig. S6B), indicating the successful introduction of N and S into the graphene.

Fig. 1.

Characterization of CoS_x/NS-G hybrids. (A) TEM image and corresponding HRTEM image (Inset) of CoS₂/NS-G. (B) Elemental mapping images of the CoS₂/NS-G hybrid. High-resolution XPS spectra of Co 2p (C) and S 2p (D) for the CoS₂/NS-G and CoS/NS-G hybrid products.

X-ray photoelectron spectroscopy (XPS) was carried out to investigate the compositions and valence states of CoS₂/NS-G and CoS/NS-G. The core-level scan XPS spectra of Co 2p, as illustrated in Fig. 1C, comprise two main peaks that result from the spin–orbit splitting of the p orbital and are assigned to Co 2p_3/2 and Co 2p_1/2 (29, 30). The two satellite peaks located at ∼786.8 and 803.7 eV can be ascribed to the shakeup type peaks of Co (31). In the high-resolution XPS spectra of S 2p shown in Fig. 1D, the CoS₂/NS-G sample exhibits two main peaks located at 162.9 and 164.1 eV that can be attributed to S 2p_3/2 and S 2p_1/2, indicating the presence of (S₂)²⁻-type species bonded to Co²⁺ (32, 33). In contrast, these two peaks are slightly shifted to lower binding energies of 161.9 and 163.2 eV for the CoS/NS-G sample, which is consistent with the bonding energies of S²⁻ species (34). The different valence bonds in the sulfur spectra further demonstrate the successful preparation of two phases of cobalt sulfide. The peaks at 165.3 eV for the two hybrids can be ascribed to the C–S–C bond, indicating the introduction of S into the graphene (35, 36). Moreover, SI Appendix, Fig. S7 shows that the core-level scan spectrum of N 1s can be divided into several characteristic peaks located at 398.6, 400.0, and 401.6 eV, which can be attributed to the peaks of pyridinic-N, pyrrolic-N, and graphitic-N, respectively (37, 38).

Characterizations of Interfacial Interaction.

To gain insight into the interfacial interaction, synchrotron soft X-ray absorption near-edge spectroscopy (XANES), which is sensitive to local chemical configuration and partial electronic state, was also performed. As shown in Fig. 2A, the cobalt L-edge XANES spectra of CoS₂/NS-G and CoS/NS-G exhibit two main peaks located at energies of 772.5 and 787.5 eV, respectively, corresponding to the L₃- and L₂ edges. The enhanced absorption intensity and the increased resolution of the XANES features of CoS₂/NS-G are in accordance with the typical characteristics of standard CoS₂ and CoS samples (39, 40). Furthermore, the sulfur L-edge XANES spectra (Fig. 2B) of CoS₂/NS-G and CoS/NS-G show obvious peaks in the region of 163∼168 eV corresponding to C–S–C coordination species, which are consistent with the result of the NS-G sample (35, 41). This result suggests that there is substantial S doping into the graphene in both hybrid materials.

Fig. 2.

Local chemical configuration and electronic state of CoS_x/NS-G hybrid catalysts. (A) Co L-edge XANES spectra of the CoS/NS-G and CoS₂/NS-G hybrids. S L-edge (B), C K-edge (C), and N K-edge (D) XANES spectra of CoS/NS-G, CoS₂/NS-G, and NS-G materials.

In Fig. 2C, the carbon K-edge spectra of the three materials all are dominated by three strong resonances with peak centers of ∼284.2 eV (peak C1), 286.9 eV (peak C2), and 291.2 eV (peak C3). These resonances can be ascribed to the dipole transition of the 1s core electron of carbon into the π*_{C = C}, π*_{C–O/N/S–C}, and σ*_C–C antibonding states, respectively, consistent with previous reports (42). The preservation of the main features indicates that the cobalt sulfide nanoparticles did not change the framework of the NS-G support. However, the different types of cobalt sulfide subtly modified the electronic state and the chemical bond of the corresponding hybrid materials, as indicated by the intensity changes in peaks C1 and C2. The intensity decrease in peak C1 suggests that the π*_{C = C} unfilled state of the aromatic ring was doped by exotic electrons, while the sharp intensity increase in peak C2 confirms that additional strong chemical bonds form out of the graphene layer and hence introduce more sp³ interactions (28, 43). This result provides solid evidence to confirm that numerous Co–N/S–C bonds existed at the interface between the nanoparticles and graphene of the CoS₂/NS-G and CoS/NS-G hybrids and that charge transfer from cobalt to the N,S-decorated graphene occurred. This can be further confirmed by the double peaks within the energy range from 294 to 300 eV, which are characteristic of an sp³ interaction in carbon-based materials. Moreover, to further characterize the extent of the interfacial interactions, the normalized intensities of peak C1 and C2 relative to C3 are shown in SI Appendix, Table S1. Compared with CoS/NS-G, CoS₂/NS-G exhibits an increased intensity ratio of C2/C3 and a more obvious sp³ interaction, which demonstrates that a stronger interfacial interaction was constructed in CoS₂/NS-G and would enable superior charge transfer ability for electrocatalysis. In addition, the cobalt sulfide nanoparticles not only regulate the electronic state of the aromatic ring but also selectively control the nitrogen configurations in the hybrid materials, which can be further confirmed by the nitrogen K-edge XANES spectra (Fig. 2D). According to previous research works (28), nitrogen XANES features arise from two types of nitrogen chemical configurations: C-N-C and N-3C. The former, which features in the energy region of 395–400 eV, indicates the chemical states of the nitrogen from pyridinic ring and pyrrolic ring; the latter, a peak at ∼405 eV, denotes graphitic nitrogen. The almost identical N-3C features of the CoS₂/NS-G and NS-G hybrids and the obviously varying C-N-C features suggest that the introduction of CoS₂ nanoparticles maintained the frame structure of the N,S-doped graphene and modified the nitrogen configurations in the edge cutting. Therefore, a strong interfacial interaction was successfully constructed in the as-prepared hybrid materials, which favors the electrocatalytic activity for NRR.

Electrocatalytic Activity for NRR.

To study the NRR catalytic activity of CoS₂/NS-G and CoS/NS-G, electrocatalytic N₂ reduction measurements were carried out in an Ar- or N₂-saturated 0.05 M H₂SO₄ solution under ambient conditions. As shown in Fig. 3A, NRR is initiated at −30 mV versus RHE under N₂ saturation at room temperature and atmospheric pressure, which is different from that in Ar. The obvious difference in current density between N₂- or Ar-saturated environments can also be observed at a more positive potential, suggesting the high NRR catalytic activity of the CoS₂/NS-G catalysts. Moreover, the polarization curves of the CoS₂/NS-G hybrid exhibit a more positive onset potential (−30 mV) and higher current density than those of CoS/NS-G (−45 mV) and NS-G (−95 mV), which can be attributed to the CoS₂/NS-G hybrid having the strongest interfacial coupling effect (Fig. 3B). Electrochemical impedance spectroscopy analysis of the CoS₂/NS-G catalyst also exhibits that its charge-transfer resistance is smaller than that of CoS/NS-G and NS-G (Fig. 3C), indicating that strong Co–N/S–C bonds lead to faster reaction kinetics by accelerating electron transfer. In addition, the current density at different potentials shows good stability (Fig. 3D). As shown in SI Appendix, Fig. S18, the current density exhibits little degeneration for CoS₂/NS-G after chronoamperometric test of 10 h. These result confirms that CoS₂ hybrid shows good stability in N₂-saturated 0.05 M H₂SO₄ during NRR test, which might be attributed to the good dispersion of the cobalt sulfide nanoparticles on NS-G support and the strongly coupled interaction between them.

Fig. 3.

Electrochemical characterization of CoS_x/NS-G hybrids. (A) Linear sweep voltammetry tests of CoS₂/NS-G in Ar- and N₂-saturated 0.05 M H₂SO₄ under ambient conditions. Polarization curves (B) and corresponding Nyquist plots (C) of different catalysts in N₂-saturated 0.05 M H₂SO₄ solution. (D) Chronoamperometric results of the CoS₂/NS-G hybrid at the different potentials.

To further verify the NRR activity of the CoS₂/NS-G and CoS/NS-G hybrid catalysts, a chronoamperometric method was employed to investigate the NH₃ yield rate and Faradaic efficiency. As shown in Fig. 4A, the CoS₂/NS-G hybrid shows a high yield rate and superior Faradaic efficiency for NH₃ production. The highest Faradaic efficiency of 25.9% was achieved at −0.05 V vs. RHE, while the highest ammonia yield of 25.0 μg h⁻¹⋅mg⁻¹_cat was obtained at −0.2 V vs. RHE. Moreover, the NRR performance of CoS/NS-G also is comparable with a high Faradaic efficiency of 15.6% at −0.05 V vs. RHE and a NH₃ yield rate of 15.7 μg h⁻¹⋅mg⁻¹_cat at −0.2 V vs. RHE (SI Appendix, Fig. S10). As shown in SI Appendix, Fig. S26, both of the as-obtained strongly coupled cobalt sulfide hybrids show a higher UV-vis absorption intensity at 655 nm than NS-G, suggesting a higher yield rate of NH₃ (SI Appendix, Fig. S11). The XPS spectra and XRD patterns of CoS₂/NS-G hybrid before and after NRR tests are shown in SI Appendix, Fig. S20. And, there was no change of XPS spectra XRD pattern for CoS₂/NS-G after NRR tests. Chronoamperometric experiments under Ar gas flow were also implemented and no product of NH₃ has been detected in this case (SI Appendix, Fig. S26). To further confirm the N source of the NH₃ during NRR process, an isotopic labeling experiment using ¹⁵N₂ as the feeding gas was performed. As shown in Fig. 4B, there are two peaks corresponding to ¹⁵NH₄⁺ by using ¹⁵N₂ as the feeding gas, indicating that ammonia is produced via electrochemical nitrogen reduction. Moreover, three peaks corresponding to ¹⁴NH₄⁺ are observed when using ¹⁴N₂ as the feeding gas and there is no detectable amount of ammonia in blank electrolyte, which further confirms that the N source of ammonia is N₂ gas. The electrolyte after different NRR durations by using ¹⁵N₂ as feeding gas was also tested by ¹H NMR. As shown in SI Appendix, Fig. S25, the NMR signal integrations for 40 h are almost two times the signals obtained at test for 20 h. This result demonstrated good stability of the electrocatalyst.

Fig. 4.

Catalytic performance of CoS_x/NS-G during the electrocatalytic N₂ reduction process. (A) NH₃ yield rate and Faradaic efficiency of CoS₂/NS-G at each given potential. (B) NMR spectra of ¹H for the electrolytes after NRR test by using ¹⁵N₂ and ¹⁴N₂ as feeding gas. (C) Comparison of NH₃ yield rate at −0.2 V and particle size for CoS₂/NS-G hybrids which was synthesized by different amounts of cobalt salt. (D) Faradaic efficiency of well-developed NRR electrocatalysts at room temperature and atmospheric pressure (SI Appendix, Table S2).

As shown in SI Appendix, Fig. S11, the strongly coupled cobalt sulfide hybrids exhibit superior NH₃ yield rate than that of the physical mixture of CoS₂ and NS-G. The above results reveal that the strongly coupled interface induced by Co–N/S–C bonds play the crucial role in enhanced NRR performance (SI Appendix, Fig. S21). Moreover, the overall NRR activity was also affected by the composition and distribution of cobalt sulfide nanoparticles. As shown in Fig. 4C, the average particle size can be tuned by using different amounts of cobalt salt. With the increasing added amount of CoCl₂, the particle size of CoS₂ increased. And, CoS₂/S-G exhibited the highest catalytic activity when the added amount of cobalt salt was 0.2 mmol. Notably, the strongly coupled hybrids exhibit an outstanding Faradaic efficiency and NH₃ production yield, which are superior to some reported non-noble-metal electrocatalysts (Fig. 4D and SI Appendix, Table S2). Based on the above results, the CoS₂/NS-G and CoS/NS-G hybrids have proven to be highly active NRR electrocatalysts, presenting promising potential for application in the field of N₂ fixation.

Universality of the Interfacial Engineering Strategy.

The interfacial engineering strategy is general and phase selective and it can be extended to develop a series of other strongly coupled metal sulfides and graphene hybrid materials, including NiS₂/NS-G, NiS/NS-G, FeS₂/NS-G, FeS/NS-G, SnS₂/NS-G, and SnS/NS-G. Detailed characterization of these hybrid materials was performed by XRD, TEM, and XPS (SI Appendix, Figs. S12–S15). All of the strong coupled sulfide hybrids can be well indexed to standard metal sulfide phases, and these hybrids show the morphology of small nanoparticles homogeneously grown on a graphene nanosheet, which is similar to the morphology of the CoS₂/NS-G and CoS/NS-G materials. Moreover, according to the XPS spectra, nitrogen and sulfur have been successfully introduced into graphene nanosheet and the strong interaction between metal sulfide and graphene substrate has been constructed by the in situ annealing method. As shown in SI Appendix, Fig. S16, the metal disulfide hybrids exhibit a higher NH₃ yield rate than monosulfide hybrids, which suggests higher NRR electrocatalytic activity. This result reveals that the strong bridging-bonds–induced interface effect can accelerate electron transfer and plays an important part in improving electrocatalytic N₂ reduction. Thus, all of the above results confirm the universality of this interfacial engineering strategy, and this general and phase-selective synthetic method will be a promising approach for producing highly active and low-cost electrocatalysts.

Conclusion

In conclusion, we have developed an interfacial engineering strategy to design two strongly coupled cobalt-based hybrids as NRR electrocatalysts. Detailed characterization confirms that many strong Co–N/S–C bridging bonds are constructed at the interface between the metal sulfide nanoparticles and graphene support. The strongly coupled chemical bonds enable controllable interfacial effects and can accelerate the reaction kinetics by acting as an electron transport channel. As expected, the CoS₂/NS-G and CoS/NS-G hybrids exhibit excellent electrocatalytic performance for NRR with an ultrahigh Faradaic efficiency and NH₃ yield rate. This work provides a strategy to design strongly coupled hybrid materials as highly active electrocatalysts for ammonia production.

Materials and Methods

The synthetic strategy and experimental methodologies used in this work are elaborated in SI Appendix. These methods include the synthesis of MS₂/NS-G, MS/NS-G, and NS-G hybrid catalysts, structural characterization, and electrochemical measurements. The calculation method of NH₃ yield rate and Faradaic efficiency, product quantification, as well as ¹⁵N₂ isotope labeling experiment are also are provided in SI Appendix.