Pengju
Guo
,
Fengxiang
Yin
* and
Jiahui
Liang
Jiangsu Key Laboratory of Advanced Catalytic Materials and Technology, School of Petrochemical Engineering, Changzhou University, Changzhou 213164, China. E-mail: yinfx@cczu.edu.cn; Tel: +86-519-86330253
First published on 7th July 2025
Low NH3 yield and faradaic efficiency (FE) are the main technical bottlenecks of the electrocatalytic nitrogen reduction reaction (NRR) to synthesize ammonia. Herein, Bi/ZIF-8 series catalysts were successfully designed and synthesized through the polyol reduction method. The NH3 yield and FE of Bi/ZIF-8 catalysts were found to be greatly improved. Among them, the 8%Bi/ZIF-8 catalyst obtained the highest NH3 yield (34.53 μg h−1 mgcat.−1) and FE (23.27%), which were better than those of most of the reported Bi-based NRR catalysts. Experiments and density functional theory (DFT) calculations revealed the following important roles of ZIF-8 on the enhanced NRR performance of Bi/ZIF-8 catalysts: (I) the high specific surface area of ZIF-8 promoted the high dispersion of Bi active sites, thereby exposing more Bi sites; (II) the abundant and ordered porous structure of ZIF-8 facilitated the mass transfer diffusion of N2-/N-related intermediates; (III) ZIF-8 facilitated full contact of the highly dispersed Bi with highly exposed Zn, resulting in more charge transfer from Zn to Bi. This effective charge transfer not only improved the adsorption capacity of Bi/ZIF-8 catalysts for N2 but also enhanced the p-electron feedback ability of Bi active sites and improved the Bi 6p-N2 π* interactions, thereby promoting the transfer of more occupied Bi 6p orbital electrons to π* of N2, which weakened the strength of NN bonds and reduced the rate-determining step (RDS) (*N
N → *N
N–H, 0.16 eV); and (IV) the introduction of ZIF-8 appropriately improved the charge transfer efficiency and increased the hydrogen adsorption free energy (ΔGH*: 0.35 eV), which effectively suppressed the hydrogen evolution reaction (HER).
In recent years, transition metal-based NRR catalysts, such as transition metal compounds, transition metal-based alloy materials, and transition metal-based single-atom materials,7–10 have been developed to enhance the adsorption of N2 molecules and promote the cleavage of NN bonds, thereby reducing the activation energy barrier. However, the d-orbital electrons of these materials are prone to form metal–H bonds and lead to occurrence of the HER,11 which reduces the FE of NRR process. Many studies have shown that most main group metals have poor hydrogen adsorption capacity.12 In particular, bismuth can not only effectively weaken the combination of charge and adsorbed proton owing to its limited surface electron accessibility,13,14 but also effectively activate N2 molecules owing to its unique p-electron structure.15,16 Various Bi-based materials, including Bi nanosheets,13 Bi nanodendrites17 and PdBi alloys,18 have been explored as NRR catalysts in recent years. All these research results have proven that Bi-based catalysts can effectively inhibit the HER and enhance the activation of N2 owing to the presence of a large number of Bi active sites. However, many Bi-based catalysts still encounter the problem of weak N2 adsorption capacity.19–21 Therefore, improving the adsorption capacity of Bi-based catalysts for N2 is expected to synthesize high-performance NRR catalysts.
It is worth noting that metal–organic framework (MOF) materials usually have excellent N2 adsorption capacity and have been significantly used in NRR in recent years.22,23 For example, some pure MOFs, such as ZIF-8,24 MIL-100 (Al)25 and HKUST-1,26 have been used as NRR electrocatalysts. In particular, ZIF-8 combined with active materials, such as Au/ZIF-8,24 WS2/ZIF-8,27 ZIF-8@Ti3C2 (ref. 28) and CeO2/ZIF-8,29 have shown efficient NRR performance. Among these, AuCu/ZIF-8 catalyst has been reported to achieve an excellent NH3 yield (63.9 μg−1 mgcat.−1) and FE (14.2%).30 Thus, the above studies have showed that ZIF-8 combined with metal active materials can exhibit higher NH3 yield and FE. This is mainly owing to the high porosity and high specific surface area of ZIF-8, which could enhance the adsorption and mass transfer diffusion of N2-/N-related intermediates;31,32 however, the crucial role of ZIF-8 in improving NRR performance has not been thoroughly and systematically studied yet. Herein, Bi/ZIF-8 series catalysts were successfully synthesized via a simple one-step reduction method. These catalysts significantly improved the NRR performance (NH3 yield and FE) compared with Bi NPs. Among them, 8%Bi/ZIF-8 exhibited the highest activity and selectivity, with an NH3 yield of 34.53 μg h−1 mgcat.−1 and FE of 23.27%, which were much higher than those of the latest reported Ti–Bi2WO6 (NH3 yield: 23.14 μg h−1 mgcat.−1, FE: 11.44%)33 and Bi2MoO6 (NH3 yield: 20.46 μg h−1 mgcat.−1, FE: 8.17%).34 The important role of ZIF-8 in improving the NRR performance of Bi/ZIF-8 series catalysts was studied in-depth and systematically via experiments and DFT calculations.
The prepared Bi/ZIF-8 catalysts all show obvious ZIF-8 and Bi diffraction peaks (Fig. 2a), indicating that ZIF-8 is preserved during the synthesis process of Bi/ZIF-8 catalysts. The diffraction peak intensity of Bi gradually increases as the amount of Bi precursor increases. Among them, most of the Bi nanoparticles in 8%Bi/ZIF-8 are uniformly dispersed on the ZIF-8 surface (Fig. 2b), and the majority of particle sizes are distributed between 12.3 and 16.7 nm (Fig. 2c), with an average particle size of about 13.58 nm. Compared with Bi NPs (Fig. 1d), Bi nanoparticles supported on ZIF-8 have better dispersibility and a smaller average particle size, which indicates that the introduction of ZIF-8 can fully expose more of the Bi active sites. Fig. 2d shows that the Bi nanoparticles in 8%Bi/ZIF-8 exhibit a 0.319 nm lattice fringe, which is attributed to the (012) crystal face of Bi. Elemental mappings of 8%Bi/ZIF-8 (Fig. 2f–i) show that the C, N, Zn and Bi elements are uniformly distributed.
![]() | ||
Fig. 2 (a) XRD patterns of Bi/ZIF-8 series catalysts; (b) TEM image; (c) particle size distribution; (d) HRTEM image; (e) HAADF-STEM image and (f–i) elemental mappings of 8%Bi/ZIF-8. |
Fig. 3a shows the N2 adsorption–desorption isotherms of ZIF-8, Bi NPs and Bi/ZIF-8 series catalysts. The isotherm curves of Bi NPs, ZIF-8 and Bi/ZIF-8 series catalysts belonged to the type I isotherm, indicating that they possess a typical microporous structure.38,39 Bi NPs exhibited the smallest specific surface area (BET SSA: 134.61 m2 g−1). After the introduction of ZIF-8 (BET SSA: 812.37 m2 g−1), the BET SSA of Bi/ZIF-8 series catalysts were greatly improved. Among them, the BET SSA of 8%Bi/ZIF-8 was 687.42 m2 g−1, which was higher than that of 4%Bi/ZIF-8 (556.74 m2 g−1) and 12%Bi/ZIF-8 (483.72 m2 g−1). The high specific surface area of ZIF-8 can well disperse Bi nanoparticles and inhibit Bi nanoparticle agglomeration, thereby exposing more Bi active sites. Table S1† shows that the introduction of ZIF-8 in Bi results in a higher pore volume and mean pore diameter of Bi/ZIF-8 catalysts compared with Bi NPs, which facilitates the full contact of catalysts and electrolyte and accelerates the mass transfer diffusion of N-related intermediates during the NRR process. Therefore, the introduction of ZIF-8 with an abundant and ordered porous structure improves the NRR performance of Bi/ZIF-8 catalysts.
In ZIF-8 and 8%Bi/ZIF-8, C 1s can be deconvoluted into two peaks, which are attributed to the C–N peak (285.8 eV) and C–C/C–H peak (284.6 eV) (Fig. S1a†).40,41 Similarly, the N 1s spectra show obvious N–C and NC bonds, corresponding to binding energies of 398.9 and 400.3 eV, respectively (Fig. S1b†).40,42 Compared with ZIF-8, the binding energy of C 1s and N 1s in 8%Bi/ZIF-8 is negatively shifted, indicating that the introduction of bismuth leads to the redistribution of surface charges.
The Zn 2p spectra exhibited two strong peaks at 1021.4 and 1044.5 eV (Table S2†), which were attributed to Zn 2p3/2 and Zn 2p1/2, respectively (Fig. 3c).40,42 The Zn 2p binding energy of 8%Bi/ZIF-8 showed a positive shift compared with Zn 2p binding energy of ZIF-8 (Table S2†). The introduction of Bi caused the decrease in the electron cloud density of Zn elements on the 8%Bi/ZIF-8 catalyst surface. Meanwhile, the introduction of NaBH4 degraded a part of the pore structure of ZIF-8, thus exposing more Zn elements. An effective exposure to Zn led to an effective contact with highly dispersed Bi sites. As for the Bi 4f spectra of Bi NPs, the peaks at 164.2 and 158.9 eV belonged to Bi3+, and the peaks at 162.1 and 156.8 eV belonged to the metal Bi0 (Fig. 3d).43,44 However, Bi 4f of 8%Bi/ZIF-8 only exhibited a Bi0 peak and showed an obvious negative shift compared with Bi NPs. This was mainly attributed to the fact that the introduction of ZIF-8 enabled full contact of highly dispersed Bi with highly exposed Zn, causing a part of the charge transfer from Zn to Bi, thereby promoting the formation of electron-rich Bi on the 8%Bi/ZIF-8 catalyst surface.
![]() | ||
Fig. 4 (a) LSV curve; (b) NH3 yields and (c) FEs of 8%Bi/ZIF-8; (d) Optimal NRR performance of the catalysts at −0.5 V vs. RHE; (e) N2H4 yields; and (f) NH3 yields of the control experiments. |
Fig. 4d and Table 1 show that the NH3 yields and FEs of Bi/ZIF-8 are much higher than those of Bi NPs and ZIF-8 at −0.5 V vs. RHE, indicating that the introduction of ZIF-8 can significantly improve the NRR catalytic activity of Bi/ZIF-8 catalysts. In the Bi/ZIF-8 series catalysts, the NH3 yield and FE show the trend of first increasing and then decreasing with the gradual increase of Bi. Among them, the 8%Bi/ZIF-8 catalyst has the best performance, and its NH3 yield and FE are 34.53 μg h−1 mgcat.−1 and 23.27%, respectively, which are superior to most Bi-based NRR catalysts (Table S5†), such as OVs-BiVO4 (NH3 yield: 8.61 μg h−1 mgcat.−1; FE: 10.03%),45 Au nanoparticles on Bi nanosheets (NH3 yield: 20.39 μg h−1 mgcat.−1; FE: 15.53%),46 2D mosaic bismuth nanosheets (NH3 yield: 13.23 μg h−1 mgcat.−1; FE: 10.46%).13
Samples | NH3 yields (μg h−1 mgcat.−1) | Faradaic efficiencies (%) |
---|---|---|
ZIF-8 | 6.87 ± 0.14 | 5.37 ± 0.12 |
Bi NPs | 8.86 ± 0.14 | 9.73 ± 0.11 |
4%Bi/ZIF-8 | 26.14 ± 0.31 | 20.29 ± 0.30 |
6%Bi/ZIF-8 | 30.68 ± 0.47 | 21.65 ± 0.35 |
8%Bi/ZIF-8 | 34.53 ± 0.52 | 23.27 ± 0.43 |
10%Bi/ZIF-8 | 27.65 ± 0.41 | 19.83 ± 0.21 |
12%Bi/ZIF-8 | 21.73 ± 0.25 | 16.26 ± 0.17 |
Fig. S8† and 4e show that the highest N2H4 yield is 0.04 μg h−1 mgcat.−1, indicating that the 8%Bi/ZIF-8 catalyst exhibited good selectivity for the NRR. To prove that the produced NH3 comes from the introduced N2, a series of control experiments were performed (Fig. S9a and b†), including (i) 8%Bi/ZIF-8 tested in the N2-filled state at OCP; (ii) 8%Bi/ZIF-8 tested in the Ar-filled state at −0.5 V; (iii) blank carbon paper (CP) tested in the N2-filled state at −0.5 V and (iv) electrolyte (0.1 M Na2SO4) placed in air for 24 h. Fig. 4f shows that the highest NH3 yield is only 0.08 μg h−1 mgcat.−1 under the above control experimental conditions, which proves that the detected NH3 was from the N2 introduced during the NRR process. Fig. S9c† shows that the current density of 8%Bi/ZIF-8 remains stable at −0.5 V vs. RHE during the 24 h i–t test. Fig. 5a and S9d† show that during ten consecutive cycle tests, current densities and the corresponding absorbances remain stable, and the NH3 yield and FE of 8%Bi/ZIF-8 have almost no decline, and its NH3 and FE losses are only 1.62% and 1.93%, respectively (Fig. 5b). These results show that the 8%Bi/ZIF-8 catalyst possess excellent electrochemical stability.
Fig. S11† shows that the N2-TPD results of Bi NPs, ZIF-8 and 8%Bi/ZIF-8. Bi NPs have N2 desorption peaks at 267 °C. Meanwhile, 8%Bi/ZIF-8 also has N2 desorption peaks at 309 and 370 °C. Compared with Bi NPs, 8%Bi/ZIF-8 has a higher N2 desorption temperature and larger N2 adsorption area, which further indicates that the introduction of ZIF-8 can enhance the N2 adsorption capacity of 8%Bi/ZIF-8.
The previous experiments confirmed that the NRR process of 8%Bi/ZIF-8 and Bi NPs follows the same alternating pathway, but 8%Bi/ZIF-8 exhibited a better N2 adsorption capacity than Bi NPs. Therefore, DFT simulation calculations were performed to further study the effect of ZIF-8 on the N2 adsorption capacity of the catalysts. Fig. S12† shows the optimal structural models of ZIF-8, Bi (110), and Bi/ZIF-8. The N2 adsorption Gibbs free energy of Bi/ZIF-8 (*N2, −0.31 eV) was significantly lower than that of Bi (110) (*N2, 0.13 eV) (Fig. 7), which indicated that the introduction of ZIF-8 was beneficial for improving the N2 adsorption capacity of Bi/ZIF-8. Meanwhile, Fig. S13† shows that the N–N bond lengths of *NN increase from 1.47316 Å (ZIF-8) and 1.51524 Å (Bi (110)) to 1.58463 Å (8%Bi/ZIF-8). Similarly, Bi/ZIF-8 exhibited longer N–N bond lengths thanZIF-8 and Bi (110) (Fig. S11†), which indicated that Bi/ZIF-8 better adsorbed and activated NxHy, thereby effectively weakening N
N and activating N2. As shown in Fig. 7, the thermodynamic rate-determining step (RDS) of ZIF-8, Bi (110) and Bi/ZIF-8 was the first hydrogenation process (*N
N → *N
N–H). After the first protonation of N2, the free energy values of the alternating pathway intermediates (*NHNH, *NHNH2 and *NH2NH2) were lower than those of the distal pathway intermediates (*NNH2, *N and *NH) (Fig. 7), indicating that ZIF-8, Bi (110) and Bi/ZIF-8 catalysts followed the alternating pathway. Meanwhile, 8%Bi/ZIF-8 exhibited a lower RDS (0.16 eV) than Bi (110) (0.47 eV) and ZIF-8 (0.64 eV). From the above results, it is clear that the introduction of ZIF-8 promotes Bi/ZIF-8 catalyst to better adsorb N2 and cleave N
N bonds, thereby reducing the RDS.
![]() | ||
Fig. 7 Free energies on (a) ZIF-8, (b) Bi (110) and (c) Bi/ZIF-8 via alternating and distal pathways. |
Generally, the N2 adsorption/activation ability of the active sites on the catalyst surface is closely related to the electronic structure of the active species. Fig. 8a and b show that the p-band center of Bi in Bi/ZIF-8 (−0.138 eV) is closer to the Fermi level compared with Bi (110) (p-band center: −1.157 eV), which is mainly attributed to the fact that highly dispersed Bi in Bi/ZIF-8 can effectively contact with highly exposed Zn in ZIF-8, resulting in a partial charge transfer from Zn to Bi, thereby promoting the formation of electron-rich Bi on the Bi/ZIF-8 surface. This proper adjustment of the p-band center effectively enhanced the adsorption capacity for N2- and N-related intermediates on Bi active sites in Bi/ZIF-8 series catalysts. Moreover, the projected state density (PDOS, Fig. 8c and d) of *NN showed that the formation of electron-rich Bi on the Bi/ZIF-8 surface further enhanced the p-electron feedback ability of Bi active sites and improved the Bi 6p-N2 π* interaction, thereby transferring more occupied Bi 6p orbital electrons to the π* antibonding orbitals of N2 (Fig. 9a). This process effectively weakened the strength of NN bonds, thus reducing the RDS (*N
N → *N
NH, 0.16 eV). Specifically, the Bi 6p state and N2 π* orbitals of Bi/ZIF-8 overlapped below and above the Fermi level (Fig. 8d), and the p–π* interactions led to the increased feedback of the occupied Bi 6p orbital electrons to the N2 π* orbital. This p-electron feedback effectively cleaved the N
N bonds and strengthened the metal–nitrogen bonds. Compared with Bi/ZIF-8, the overlap of 6p states and π* orbitals of N2 on Bi (110) below/above the Fermi level was significantly lower (Fig. 8c), indicating that Bi/ZIF-8 exhibited a stronger N2 activation ability than Bi (110), which further proved that the effective charge transfer between Bi sites and Zn in Bi/ZIF-8 can enhance the p-electron feedback ability of Bi active sites and weaken N
N bonds, thereby reducing the RDS.
![]() | ||
Fig. 8 Density of states (DOS) diagram of (a) Bi (110) and (b) Bi/ZIF-8 and PDOS of N2 adsorbates on (c) Bi (110) and (d) Bi/ZIF-8. |
Meanwhile, HER calculations were performed to further study the effect of ZIF-8 on the NRR selectivity of Bi/ZIF-8 catalysts. Fig. 9b shows that the ΔGH* values of ZIF-8, Bi (110) and Bi/ZIF-8 are 0.04, 0.20 and 0.35 eV, respectively, indicating that 8%Bi/ZIF-8 better suppress the proton adsorption, thereby limiting the HER process. Moreover, ZIF-8, Bi (110) and Bi/ZIF-8 were located in the N2 dominant region (ΔGH* > ΔGN2*) (Fig. 9c). Especially for Bi/ZIF-8, this indicated that it is more prone to the desired NRR process than the HER side reactions.
According to the above analysis results, ZIF-8 played a very important role in enhancing the NRR performance of Bi/ZIF-8 series catalysts. The introduction of ZIF-8 promoted the Bi/ZIF-8 series catalysts to better adsorb/activate N2 molecules, thereby reducing the RDS (Fig. 9d), significantly increasing the hydrogen adsorption free energy and inhibiting the hydrogen adsorption, which improved the selectivity for NRR (Fig. 9d). Therefore, the NH3 yield and FE of Bi/ZIF-8 catalysts were much higher than those of Bi NPs.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta02526e |
This journal is © The Royal Society of Chemistry 2025 |