Introduction

As one of the most general nitrogen compounds, ammonia (NH3) is widely utilized in the fields of the chemical industry, agriculture, medicine, etc. [1,2,3]. However, the conversion from N2 to NH3 is difficult to accomplish for the high N≡N bond energy and low polarizability of N2 [4,5,6]. At present, it remains to be the most effective technique to synthesize NH3 according to Haber–Bosch process at harsh conditions, meanwhile accompanied with high energy consumption and heavy emissions of CO2 [7, 8]. Therefore, it is essential to develop a both eco-friendly and sustainable method for producing NH3 to replace the Haber–Bosch process.

In recent reports, electrochemical nitrogen reduction reaction (NRR) under ambient conditions has the tendency to be conducted as a novel approach for efficient electrocatalytic N2-to-NH3 conversion [9,10,11,12,13,14,15]. During this process, the selection of suitable electrocatalysts plays an important role in ammonia production efficiency. A large number of the studies reported in the past decade have shown that noble metal materials (Au, Pd, Ru, etc.) can greatly promote N2 reduction reaction [16,17,18,19]. Nevertheless, the defects such as scarcity and high cost of these noble metals could obstruct their application in large-scale ammonia production [20, 21]. Among these noble metal electrocatalysts, Ag is the most promising electrocatalyst due to its lowest price and high catalytic activities, which has been applied in electrochemical hydrogen evolution reaction (HER) or oxygen evolution reaction (OER) [22,23,24]. But the severe self-aggregation of small-size Ag nanoparticles leads to not only the decrease in activity but also reduction in conductivity, resulting in a poor electrocatalytic performance [25,26,27].

Reduced graphene oxide (rGO), a novel two-dimensional carbon material, exhibits super-high surface area and excellent chemical properties, which is an ideal support material for nanoparticles to be well dispersed [28, 29]. Moreover, the unique π–π conjugate sp2 carbon structure of graphene provides an effective path for electron migration, endowing graphene nanocomposites with strengthened conductivity [30, 31]. These advantages of rGO contribute to a tremendous possibility in promoting the electrocatalytic activity of metallic Ag at NRR, which has not been reported before.

In this investigation, Ag nanoparticles-reduced graphene oxide hybrid (Ag NPs-rGO) is synthesized as an improved Ag-based catalyst for electrochemical N2 reduction. In a 0.1 M Na2SO4 electrolyte, the Ag NPs-rGO attains the NH3 yield of 18.86 μg h−1 mg−1 cat. and Faradaic efficiency (FE) of 3.60% at − 0.7 V vs. reversible hydrogen electrode (RHE), evidently higher than that of Ag NPs (9.43 μg h−1 mg−1 cat., 2.25%). In addition, this catalyst shows good stability and excellent selectivity during the NRR experiment.

Results and discussion

Ag NPs-rGO is prepared by a modified method of hydrothermal reaction (see the Supporting Information for preparation details) [32]. Figure 1a shows the X-ray diffraction (XRD) patterns of Ag NPs and Ag NPs-rGO. The XRD pattern for Ag NPs shows diffraction peaks at 38.1°, 44.3°, 64.4° and 77.3° that correspond to the (111), (200), (220) and (311) planes of the Ag phase (PDF No. 87-0597), respectively. Ag NPs-rGO also shows the coincident peaks characteristic of Ag, and the weak diffraction peak at about 23.3° is ascribed to the (002) plane of few-layer rGO. The result indicates that GO was successfully reduced and the precursor of silver was decomposed into metallic silver through the hydrothermal process. The scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images of Ag NPs-rGO are shown in Fig. 1b and c, proving the formation of well-dispersed Ag NPs adsorbed on the surface of rGO through reduction reaction. The high-resolution TEM (HRTEM) image (Fig. 1d) reveals well-resolved lattice fringes with an interplanar distance of 0.232 nm of Ag NPs-rGO, which is matched to the (111) plane of Ag NPs. Selected area electron diffraction (SAED) image (inset in Fig. 1d) shows the polycrystalline structure of Ag NPs-rGO indexed to the (111), (200) and (311) planes of the Ag phase.

Figure 1
figure 1

a XRD patterns for Ag NPs and Ag NPs-rGO. b SEM, c TEM and d HRTEM images for Ag NPs-rGO. Inset: SAED pattern of Ag NPs-rGO

Full size image

Raman spectroscopy is an important method to characterize graphene-based materials and has the advantage of nondestructive testing [33]. The Raman spectra of GO and Ag NPs-rGO are presented in Fig. 2a at a spectral range of 1200–2000 cm−1. Obviously, two feature peaks at ~ 1350 cm−1 and at ~ 1593 cm−1 are attributed to D band (corresponding to the disorder sp3 carbon) and G band (corresponding to the graphitic sp2 carbon) of both GO and Ag-rGO samples [34]. Since D band is related to defects of graphene, the intensity ratio of D band to G band (ID/IG) can be utilized as an index of disorder in GO or rGO. Ideally, the defects will decrease after reduction of GO, with a lower value of ID/IG [33]. From the calculation, the ratio of ID/IG increases from 0.92 of GO to 1.02 of Ag NPs-rGO. The reverse result is probably due to the reason that the oxygen-containing functional groups were reduced, and the conjugated sp2 carbon was reconstructed with smaller size than the original one [35]. Figure 2b displays the X-ray photoelectron spectroscopy (XPS) survey spectrum of Ag NPs-rGO, implying the presence of Ag, N, O and C elements. In the region of C 1s (Fig. 2c), three peaks at the BEs of 284.8 eV, 286.2 eV and 287.8 eV correspond to the sp2-hybridized carbon (C–C/C=C), C–O, and C=O, respectively [36]. The weaker peaks of the carbon–oxygen species signals in Ag NPs-rGO demonstrate a high degree of deoxygenation and successful reduction from the GO to rGO. The Ag 3d spectra (Fig. 2d) shows two peaks of Ag0 3d5/2 (368.2 eV) and Ag0 3d3/2 (374.2 eV) at both Ag NPs and Ag NPs-rGO, indicating that the synthesized Ag is mostly at zero valence state [32, 37].

Figure 2
figure 2

a Raman spectra of GO and Ag NPs-rGO. b XPS survey spectrum for Ag NPs-rGO. XPS spectra in the c C1s, and d Ag 3d regions

Full size image

Under ambient conditions, 0.1 mg Ag NPs-rGO was loaded on a 1 × 1 cm2 carbon paper as the working electrode for NRR experiment with continuous N2 inputting in 0.1 M Na2SO4. All potentials at the analytical process were converted to a RHE scale. Figure S1 shows the linear sweep voltammetry (LSV) curve for Ag NPs-rGO in Ar- and N2-saturated 0.1 M Na2SO4 solution. It is evident that Ag NPs-rGO attains a high current density in the N2-saturated solution, revealing that it is active for the electrocatalytic reaction to N2. As shown in Fig. 3a, the time-dependent current density curves at different potentials keep relative stability under 2 h of electrolysis. The production of NH3 was determined by using the indophenol blue method [38], and the corresponding calibration curve is presented in Fig. S2. Figure 3b shows the UV–Vis absorption spectra of catalytic electrode stained with the indophenol indicator after NRR electrocatalysis for 2 h with potential ranging from − 0.5 to − 0.9 V. Then, calculated by the linear formula of fitting curve, NH3 yields and FEs of Ag NPs-rGO at different potentials are obtained in Fig. 3c. The maximum NH3 yield rate of 18.86 μg h−1 mg−1 cat. and a high FE of 3.60% can be reached at − 0.7 V, while the Ag NPs-rGO shows an excellent performance compared with most reported aqueous-based NRR electrocatalysts listed in Table 1 [10, 11, 39,40,41,42,43,44]. When the applied potential is at − 0.5 V, Ag NPs-rGO achieves a high FE of 4.77% due to its minimal current density. But the inferior NH3 yield of 9.43 μg h−1 mg−1 cat. suggests its unsatisfactory catalytic activity for NRR at this potential. From previous reports, the activity of carbon materials would lead to the possibility of electrocatalytic properties at NRR experiments [10, 11, 22]. To further testify the optimal performance of Ag NPs-rGO, comparative experiments were performed with Ag NPs, rGO and blank CP at − 0.7 V or at open-circuit (OC) potential. As shown in Fig. 3d and Fig. S4a, the NH3 yields and FEs of CP (2.96 μg h−1 mg−1 cat., 0.75%), rGO (3.62 μg h−1 mg−1 cat., 0.77%) and Ag NPs (9.43 μg h−1 mg−1 cat., 2.25%) exhibit deficient catalytic activities for NRR, compared with Ag NPs-rGO. For the unusual NH3 production of blank CP, it can be attributed to the possible activated C, significant conductivity or other catalytic elements. XPS analysis of CP indicates the presence of abundant F that is frequently doped into catalysts for NRR experiments (Fig. S3) [45,46,47]. In this series of experiments, comparative data at a different atmosphere is vital to eliminate background contributions such as dissolved N2, remnant NH4+ absorbed on Nafion membrane or N-containing species in chemicals. And the experiments under an Ar atmosphere were conducted at the same applied potentials. In Fig. S4b, it is clear that the generated NH3 concentrations at various potentials are extremely approaching 0 μg mL−1 by continuously supplying Ar. Thus, the colorimetric data can not only indicate that the background contributions are almost ignorable, but verify that the N atom of NH3 is generated from N2 via NRR. From the colorimetric data of electrolytes in the anodic compartment, NH4+ crossover is inevitable in Nafion membranes that would cause the final calculated NH3 production rate of catalysts to get a smaller value (Fig. S5). Furthermore, electrochemical double-layer capacitance (Cdl) is related to the active surface area of the electrocatalyst [48]. Thus, high Cdl value of Ag NPs-rGO ensures the sufficient active sites for NRR experiments (Fig. S6). The superiority of Ag NPs-rGO is mostly attributed to the enhancement of rGO to nanoparticles at dispersivity and conductivity. The electrochemical impedance spectroscopy (EIS) data show that Ag NPs-rGO has a smaller radius of the semicircle than that of Ag NPs (Fig. S7). The values of charge transfer resistance (Rct) for Ag NPs-rGO and Ag NPs are 14.84 Ω and 17.86 Ω obtained from the equivalent circuit. Thus, a lower Rct of Ag NPs-rGO exhibits preferable electrical conductivity, causing faster NRR kinetics. Moreover, hydrazine (N2H4) may be a possible by-product in the NRR, which is determined using the method of Watt and Chrisp [49]. Figure S8a and S8b presents UV–Vis absorption spectra with various concentrations of N2H4 and the corresponding calibration curves, respectively. Figure S9 shows that no obvious N2H4 can be detected at various potentials, which proves the excellent selective catalytic activity of Ag NPs-rGO.

Figure 3
figure 3

a Time-dependent current density curves for Ag NPs-rGO at different potentials in 0.1 M Na2SO4. b UV–Vis absorption spectra of the electrolytes stained with an indophenol indicator after NRR electrolysis at a series of potentials for 2 h. c NH3 yields and FEs for Ag NPs-rGO at a series of potentials in 0.1 M Na2SO4. d NH3 yields and FEs for different electrodes at − 0.7 V or at OC after 2 h of electrolysis under ambient conditions

Full size image
Table 1 Comparison of the electrocatalytic NRR performance of Ag NPs-rGO with other aqueous-based NRR electrocatalysts at room temperature
Full size table

Furthermore, stability is also an important factor in evaluating the performance of catalysts in long-term utilization. During the six times consecutive cycling tests at − 0.7 V, Ag NPs-rGO has a minimal variation in the NH3 yields and FEs, exhibiting the strong recycling sustainability of the catalyst (Fig. 4a). In Fig. 4b, the time-dependent current density curve at − 0.7 V is maintained at an invariable range all the time. It is inferred that Ag NPs-rGO can hold a stable catalytic activity for at least 24 h, indicating its sustainable stability. For the catalytic mechanism of Ag (111) for NRR, according to previous reports [50, 51], it follows the associative mechanism where the nitrogen molecules are step-by-step reduced with protons and electrons. As illustrated in Fig. 4c, the potential-determining step for NRR on Ag surface is the step of *N2 combined with the first H (where * denotes a surface active site). With the introduced rGO as supporter, the electron transfer reaction will be accelerated and Ag NPs are uniformly distributed on this carbon layers, which maximizes both the exposure of active sites and availability of catalysts.

Figure 4
figure 4

a NH3 yields and FEs at the potential of − 0.7 V with the cycling tests of six times. b Time-dependent current density curve of Ag NPs-rGO at − 0.7 V for 24 h. c The scheme of catalytic mechanism for NRR on AgNPs-rGO

Full size image

Conclusions

In summary, Ag NPs-rGO nanocomposites are experimentally conducted as one modified silver-based NRR electrocatalyst under ambient conditions. With the high specific surface area of rGO, the aggregated Ag NPs are dispersedly anchored on the surface of the rGO sheets, leading to more exposure of active sites. Meanwhile, rGO greatly enhances the conductivity of catalyst, increasing its catalytic performance for NRR. At the potential of − 0.7 V versus RHE, Ag NPs-rGO achieves a high NH3 yield rate (18.86 μg h−1 mg−1 cat.) and FE (3.60%) in 0.1 M Na2SO4, accompanied with excellent selectivity and electrochemical stability. This study not only improves the catalytic activity of Ag NPs at the NRR, but also verifies the enormous effect of rGO on nanocatalysts.