Riki
Nakatani
a,
Saikat
Das
*a and
Yuichi
Negishi
*ab
aDepartment of Applied Chemistry, Faculty of Science, Tokyo University of Science, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan. E-mail: saikatdas@rs.tus.ac.jp
bInstitute of Multidisciplinary Research for Advanced Materials, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan. E-mail: yuichi.negishi.a8@tohoku.ac.jp
First published on 6th September 2024
Since the pioneering discovery of metal–organic frameworks (MOFs) a quarter century ago, they have evolved as a new category of porous crystalline extended network structures with atomic/molecular level designability. The distinctive porosity and structural customizability of MOFs have been crucial to their wide spectrum of applications. Among these applications, MOFs have gained prominence as electrocatalysts in the nitrogen reduction reaction (NRR), which has inimitable potential to solve environmental crises. The reticular structures of MOFs with uniformly distributed active sites and their easy accessibility endow them with enhanced catalytic activity. In this review, we provide a comprehensive overview of MOF catalyst design protocols from molecular building blocks to extended networks, and discuss how precisely designed MOF electrocatalyst structures can explicitly control targeted NRR pathways, besides giving an insight into the future prospects and challenges.
Studies in recent years have documented a diverse range of materials harnessed for heterogeneous catalysis research that include supported metal nanoparticles,15 nanoclusters,16 single atoms,17 metal oxides,18 zeolites,19 mesoporous silica,20 MOFs,21 and covalent organic frameworks (COFs),22 among others. Progress notwithstanding, significant gaps still exist, which necessitates addressing several limitations and challenges: (i) the activity, selectivity and durability of functional heterogeneous catalysts for targeted reactions fall well short due to structural degradation and challenges of achieving precise conditions and intricate techniques for the controlled synthesis of catalysts; (ii) the structural and chemical complexity of heterogeneous catalyst systems leads to a poor comprehension of the mechanisms of surface-catalysed reactions and structure–activity correlations, which hinders systematic design approaches to optimal catalysts.23 Regarding the development and optimization of heterogeneous catalysts, we should maintain a proper balance between material-specific properties and properties that are defined by the relationship of the catalyst with its environment. Accordingly, the materials suitable as heterogeneous catalysts must be endowed with the following characteristics: first, the design and synthesis of materials should be in favor of achieving a uniform and dispersed active site distribution; the former ensures high selectivity toward the products of interest while the latter results in high activity. Second, the structure of catalytic materials needs to be sufficiently robust so as to achieve high activity while allowing long-term stability and recyclability over consecutive runs. Third, the tailorability of the material components could provide precise control over catalytic activity and selectivity at the atomic level.
Ammonia (NH3) is necessary to produce nitrogen-based fertilizers and is extensively used as a carbon-free energy source, serving as an alternative to fossil fuels.24 On the other hand, nitrogen occupying the majority of the atmosphere possesses high stability and chemical inactivity because of its triple bond. In early 20th century, Fitz Haber and Carl Bosch developed the breakthrough process, known as Haber–Bosch process, which produced NH3 from N2 and H2 with a single pass conversion of usually 10–15%.25,26 However, the reaction is energy-intensive and entails catalysts operating at elevated temperatures of 400–550 °C and pressures of 15–25 MPa.27 As a result, 2% of energy consumption and 0.5% of green-house gas emission in the world are attributed to this process.28 Therefore, the eco-friendly nitrogen reduction reaction (NRR) utilizing low energy is highly coveted. The electrocatalytic NRR is attracting the attention of the scientific community as a clean and green sustainable approach.29 The electrochemical NRR (eNRR) is capable of producing NH3 from N2 gas and H2O as a hydrogen source at ambient temperature and pressure, consuming much lower energy than conventional processes.30 Herein, we show the reactions occurring at each electrode as follows:31
Anode reaction (in acid electrolyte):
3H2O → 3/2O2 + 6H+ + 6e− |
Anode reaction (in base electrolyte):
6OH− → 3H2O + 3/2O2 + 6e− |
Cathode reaction (in acid electrolyte):
N2 + 6H+ + 6e− → 2NH3 |
Cathode reaction (in base electrolyte):
N2 + 6H2O + 6e− → 2NH3 + 6OH− |
Overall reaction:
N2 + 3H2O → 2NH3 + 3/2O2 |
As the eNRR is an emerging field with increasing research interest, numerous materials have been tested as catalysts for the eNRR. Transition metals such as Ru, Mo, Rh and Fe have been actively studied for improving NRR activity by facilitating the cleavage of NN bonds.32 Transition metal oxides are enticing catalysts for enhancing NRR activity by forming oxygen vacancies and doping heteroatoms or metals, which function as reaction sites.33 However, transition metal oxides face challenges such as low electrical conductivity, poor stability under acidic conditions, and weak adsorption strength, which limit their activity in the NRR.34 Single-atom catalysts (SACs) minimize the amount of metal and allow each individual metal atom to drive the reaction. Nevertheless, they are known to face stability issues, as single atoms tend to aggregate during the eNRR. To overcome this issue, aggregation while maintaining sufficient active sites should be prevented, although this approach is still under investigation.35 In addition to the aforementioned metal catalysts, environment-friendly catalysts that do not use metals, such as graphene, g-C3N4, COFs, etc. have also been developed.36 These non-metal electrocatalysts are suitable for the NRR because their low proton adsorption ability suppresses the HER. However, they exhibit a low nitrogen conversion rate and low current density in NH3 production, resulting in lower NH3 generation efficiency compared to metal catalysts, which makes their practical application difficult. As hybrid materials composed of metal atoms and organic ligands, MOFs provide some remarkable merits as catalysts for the eNRR: (i) MOFs enable atomically precise structural design; (ii) MOFs have the potential to function as molecular sieves, which increases the concentration of specific molecules; and (iii) the ordered lattice structure provides well-defined active sites, which simplifies the consideration of reaction mechanisms.
Regarding the catalysts used for the eNRR, their stability and conductivity are not yet up to the mark and pose challenges.37 Commonly, most MOFs behave as insulators (conductivity <10−10 S cm−1) owing to the absence of energy pathway for charge transport or free charge carriers.38 However, recently some methods have successfully provided MOFs with sufficient conductivity, enabling their utilization in electrochemical applications: implementing specific linkers, introducing guest molecules, and loading with conducting hosts, among others.39 Additionally, the kind of metals used to construct MOFs is also one of the important factors to consider in catalysis, which decides the cost and versatility. Since the first research on MOFs exploited as eNRR catalysts, subsequent studies have enriched this field substantially. Notably, some previous reviews have already provided an overview of MOFs as catalysts for the NRR.40–45 However, there is still a need for in-depth investigation of the correlation between catalytic activities and atomic-level structures of MOFs, as well as recent experimental and theoretical developments in this field. This review provides a unique perspective on understanding the NRR mechanisms and strategies by exploring the interplay of metal sites, functional modules, catalytic species, defects and the confined microenvironment in the MOF crystalline structure. In this review, we illustrate the core concepts of the NRR, avant-garde MOFs harnessed for the NRR, design processes and recent developments in MOF electrocatalysts, as well as chronologically trace the journey of NRR electrocatalysts with MOFs. In terms of unique traits of MOFs, they have ordered uniformity and porosity that enable guest molecules to access active sites readily. Furthermore, we discuss future directions to advance this field in the last section of this review.
In the dissociative pathway, the NN bond in the N2 molecule is first broken and each separated N atom is adsorbed on the surface of the catalyst. Following this, the hydrogenation of N atoms occurs and NH3 molecules are generated and released (Fig. 1a). In this process, the first reaction needs high energy because of the strong NN bond. In practice, the Haber–Bosch process typically follows this pathway.61
Fig. 1 Different pathways of the NRR: (a) dissociative pathway, (b) associative alternating pathway, (c) associative distal pathway, (d) enzymatic pathway, and (e) MvK pathway. |
In the associative alternating pathway, the N2 molecule is initially chemisorbed on the active sites in an end-on configuration. Secondly, hydrogenation takes place and one bond of NN is severed. After that, another hydrogenation reaction occurs and the second bond of NN is severed. Finally, hydrogenation happens and two NH3 molecules are produced in sequence (Fig. 1b).
The associative distal pathway also follows the same procedure as the associative alternating pathway until the halfway point. After being connected with the surface of the catalyst, the preferential hydrogenation of the nitrogen atom that is not adsorbed on the surface and the amputation of the NN bond gradually occur, and NH3 molecules are generated. After this, the nitrogen atom that is adsorbed on the surface also undergoes hydrogenation and a NH3 molecule is formed (Fig. 1c).
In the enzymatic pathway, first, a N2 molecule is attached in a side-on configuration on the catalyst surface. After this, hydrogenation of both N atoms occurs and one bond of NN is cleaved. Furthermore, the same process occurs again. Subsequently, a NH3 molecule is generated and released. Finally, the remaining N atom is converted to NH3 through hydrogenation and released (Fig. 1d).
The MvK pathway unfolds on the surface of TMN catalysts.62 Initially, the N atom which exists on the surface connects H atoms. After reduction of the N atom, the N atom is released as NH3. Following this, vacancies are created on the surface and the incoming N2 molecule is stuffed into the vacancy. Subsequently, the N atom which is not stuffed undergoes hydrogenation and is released as NH3. Finally, the remaining N atom regenerates the surface of the catalyst (Fig. 1e).
Fig. 2 The overall indophenol method reaction between HClO, phenol and NH3 presented by Bolleter and colleagues. Adapted with permission from ref. 63. Copyright 1961, American Chemical Society. |
In 2019, Wei and co-workers employed a self-supported catalyst CuII-MOF, [Cu24L12(H2O)12]·30DMF·14H2O, on carbon cloth (represented as JUC-1000/CC) as the cathode and anode for both the NRR and oxidation of SG to GA.75 First, JUC-1000/CC was utilized as the working electrode in the N2-saturated 1.0 M Na2SO4 solution electrolyte. As a result, the NH3 generation rate and FE were as high as 13.274 μg h−1 mgcat−1 and 1.519% at −0.3 V vs. RHE, respectively. In addition, this electrode was used as both the cathode and anode for the SG oxidation reaction and NRR in a 1.0 M SG solution. The results evinced that the highest NH3 generation rate and FE were 24.7 μg h−1 mgcat−1 and 11.90% at 0.4 V vs. Ag/AgCl, respectively, whereas the SG conversion rate of the oxidation reaction under the same conditions was 100% and the GA selectivity rate was 96.96% for a 12 hour reaction.
In addition to the previously mentioned research study that used transition metal-based MOFs, main group metal-based MOFs have also been explored as NRR catalysts. In the following year, MIL-100(Al) was utilized as a NRR catalyst by the Ma group.71 The performance was evaluated in N2-saturated 0.1 M KOH aqueous electrolytes at ambient temperature and pressure. The amount of produced NH3, rate and FE was 43.15 μg h−1 cm−2 mgcat−1 and 10.9% at −0.7 V vs. RHE, respectively. The stability of the MIL-100 (Al) catalyst was verified, which revealed that upon the completion of five cycles of the NRR at 177 mV, the NH3 production rate and FE were both maintained with insignificant changes. Besides, at a constant overpotential of 177 mV for 10 hours, no observable changes were exhibited by current density.
An approach focusing on functional groups, different from metal sites, was also described. Yin and colleagues utilized both MIL-88B-Fe and NH2-MIL-88B-Fe as electrocatalysts for the NRR.76 0.1 M Na2SO4 aqueous solution was adopted as the electrolyte. The former MOF exhibited an NH3 production rate of 3.575 × 10−11 mol s−1 cm−2 and a FE of 5.59% at −0.25 V vs. RHE. On the other hand, the latter one featured the highest NH3 production rate of 1.205 × 10−10 mol s−1 cm−2 with a FE of 5.66% at −0.45 V vs. RHE, which was higher than that of the former. Additionally, the highest FE of NH2-MIL-88B-Fe was 12.45% at 0.05 V vs. RHE. In terms of the selectivity of NH3, no existing N2H4 was detected during the NRR, which indicated high selectivity.
To achieve more efficient NRR activity, conductive MOFs were explored for this reaction. In a separate study by Jiang, Sun and co-workers, Co3HHTP2 was loaded on CP and employed as the working electrode with a N2-saturated 0.5 M LiClO4 solution as the electrolyte.77 The NRR performances of Co3HHTP2/CP were evaluated at different potentials and the highest NH3 yield rate was 22.14 μg h−1 mgcat−1 with FE of 3.34% at −0.40 V vs. RHE. The stability and durability of Co3HHTP/CP were also evaluated through cycling test and time-dependent current density. All the results revealed that Co3HHTP2/CP has adequate stability during the NRR.
For practical application, Liu and colleagues adopted a typical MOF, HKUST-1 as the NRR catalyst.78 HKUST-1 was dropped on CP and the electrode was named HKUST-1/CP. The highest NRR performance was achieved at −0.75 V vs. RHE with an NH3 yield rate of 46.63 μg h−1 mgcat−1 and a FE of 2.45%. Time-dependent experiments evidenced a nearly linear relation between the amount of NH3 generated and time, and no trace of the by-product N2H4 was detected at any potential.
Furthermore, another investigation of how the amount of metal precursors affects catalytic activity was conducted. In 2021, Jing, Zhang and colleagues reported the Cu-MOF and Ce-MOF grown on the surface of copper mesh (hereinafter Cu@Cu-MOF and Cu@Ce-MOF) and employed as NRR catalysts.79 Regarding Ce-MOFs, three kinds of MOFs with different amounts of Ce precursors were prepared. The amounts of Ce(NO3)3·6H2O used for synthesizing Ce-MOF-1, Ce-MOF-2 and Ce-MOF-3 were 20 mg, 60 mg and 300 mg, respectively. All NRR experiments were conducted by utilizing 0.1 M KOH aqueous solution as the electrolyte. The results indicated that the NH3 yield and FE were 10.23 μg h−1 cm−2 and 5.12%, respectively at −0.2 V vs. RHE. On the other hand, each of the Ce-MOFs exhibited higher activity than Cu-MOFs; Cu@Ce-MOF-2 illustrated the highest NH3 yield (14.83 μg h−1 cm−2) and FE (10.81%) among the three Cu@Ce-MOFs at −0.2 V vs. RHE (Fig. 4b). Additionally, Cu@Ce-MOF-2 possessed high stability during nine cycles. Besides, Cu@Ce-MOF-2 showed selectivity towards NH3, which was indicated by the absence of N2H4 detection.
As part of fundamental research, how solvent conditions affect the NRR catalytic activity of MOFs was evaluated. Chen, Duan and co-workers leveraged the 2D In-MOF for evaluating NRR activity in various pH electrolytes.80 The highest NH3 yield rates at −0.5 mA cm−2 were 24.70, 25.80, 64.73, 79.20 and 71.58 μg h−1 mg−1 corresponding to pH values of 1, 2, 7, 12 and 14, respectively. The FEs were calculated to be 6.72, 7.50, 12.23, 14.98 and 13.54% at pH 1, 2, 7, 12 and 14, respectively, in which the former two were measured at −0.25 mA cm−2 and others were measured at −0.5 mA cm−2. The results unveiled the best NRR performance at pH 12 among the various pH experimental conditions. Furthermore, the In-MOF exhibited superb stability even when it was exposed to acid and alkaline solutions. The theoretical investigation showed that the reaction occurred on In atoms, and the process followed an enzymatic pathway, which was accomplished by the Gibbs free energy changes. Furthermore, the calculation results showed that the rate-determining step in the NRR was the first NH3 generation step.
Several studies have demonstrated the improvement in catalytic activity upon creating defects in the MOF structure. In 2022, the Yin group harnessed defective UiO-66-NH2 to adsorb O2 radicals for the NRR.81 The best NRR performance in terms of the NH3 yield rate and FE with UiO-66-NH2 was ∼2.071 × 10−10 mol s−1 cm−2 and ∼85.21%, respectively, at −0.39 V vs. RHE in N2-saturated 0.1 M Na2SO4. Besides, UiO-66-NH2 was used in the NRR for 6 cycles, indicative its high stability during the course of this reaction. The high value of FE is attributable to the Zr–OO˙ sites which are formed by exposed Zr atoms, adsorbed O2 radicals and –NH2 functional groups stabilizing the Zr–OO˙ sites.
A study on defective MOFs was reported by the Yin group as well. They also utilized the same MOF, UiO-66, that was made defective by implementing HPMo (hereinafter UiO-66-xHPMo, where x represents the ratio of Zr and Mo: 0.5, 1, 2 and 3) as an electrocatalyst for the NRR.82 UiO-66-xHPMo presented a four-fold enhancement in NRR activity with a maximum NH3 yield rate of 36.61 μg h−1 mgcat−1 relative to pristine UiO-66. However, regarding FE, UiO-66 displayed the best results with a FE of 35.76%, which suggested that the MOF defects decreased the proton conductivity thereby enhancing the charge transfer resistance. Regarding UiO-66-xHPMo, UiO-66-2HPMo illustrated the best NRR activity with an NH3 yield rate of 36.61 μg h−1 mgcat−1 and FE of 31.09% at −0.3 V vs. RHE (Fig. 4d). The N2H4 by-product yield during the NRR was also measured to assess the selectivity, and the highest N2H4 yield was 0.04 μg h−1 mgcat−1 at −0.1 V vs. RHE. Furthermore, the current density at −0.3 V vs. RHE remained almost unchanged for 48 h, and UiO-66-2HPMo showed no obvious changes after five catalytic runs, demonstrating the adequate durability of UiO-66-2HPMo.
Recently, another defect-based MOF study for the NRR has been carried out. Liu, Yan and colleagues introduced a defective Al-Fum MOF, which was deposited on CP for the NRR.83 The concentration of K2SO4 as the electrolyte was selected as 0.4 M. Under this condition, the highest NH3 yield rate and FE of Al-Fum MOF at −0.15 V vs. RHE were 53.9 μg h−1 mgcat−1 and 32.9%, respectively. To corroborate the importance of defective sites in the Al-Fum MOF, the NRR performances of pristine and defective Al-Fum MOF were compared under the same conditions, which indicated noticeably lower activity of the former (NH3 yield rate: 13.2 μg h−1 mgcat−1, FE: 11.8%) than the latter. To validate the stability of the defective Al-Fum MOF, ten NRR cycles were conducted which demonstrated that the NH3 yield rate and FE were maintained, suggestive of its superb stability. The DFT calculations deciphered the pivotal role of defective Al sites in activating N2, which facilitated the NRR process.
Research studies also focused on exploring the relationship between the symmetry of connections between metal sites and linkers and catalytic activity. Zhou, Lee and co-workers decoded the effect of symmetry-breaking in three MOFs on the NRR activity by enabling the adsorption of N2.84 The Zn-N2S2-MOF presented the highest NH3 yield rate of 25.07 ± 1.57 μg h−1 cm−2 and a FE of 44.57 ± 2.79% at −0.3 V vs. RHE (Fig. 4e) and it was further characterized by XPS, X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS). The study highlighted that increasing the number of S atoms around Zn-sites enhances the Zn electron density because of delocalized electrons donated by S, which resulted in high NRR activities by suppressing the HER. Additionally, the importance of symmetry-breaking for the ORR and CRR in the three MOFs was explored.
The study on bimetallic MOFs containing Fe was further investigated in another study by Chen, Zhao and co-workers employing a zero-dimensional Ni and Fe-based bimetallic MOF, NiFe-MOF, as the NRR catalyst.86 The NRR performance of the NiFe-MOF was investigated with 0.1 M NaHCO3 aqueous electrolyte at ambient temperature and pressure. The NH3 production rate was obtained to be 9.3 μg h−1 mgcat−1 and the FE was 11.5% at −347 mV vs. RHE. In order to examine the reaction pathway in detail, density functional theory (DFT) calculations were conducted. The nitrogen molecules were physically adsorbed onto the metal sites, which worked as active sites, and the reaction proceeded through the associative pathway. Furthermore, the analysis of the minimum-energy pathway indicated that the Ni-MOF and Fe-MOF showcased the highest Gibbs free energy during the release of the first NH3 molecule, making it the rate-limiting step. On the other hand, in the case of the NiFe-MOF, the energy during NH3 desorption decreased as compared to the previous stage. This result demonstrated that MOFs constructed using multiple metals allow for the control of Gibbs free energy during the reaction step.
The Yin group reported the UiO-Zr-Ti MOF, prepared by post-synthetic metal exchange on parent UiO-66-NH2 which led to substitution of some Ti clusters for Zr clusters, which displayed NRR activity in 0.1 M Na2SO4 electrolyte.87 UiO–Zr–Ti synthesized in an oven for 5 days showed the highest NH3 production rate of 1.16 ± 0.058 × 10−10 mol s−1 cm−2, and the highest FE was 80.36 ± 4.23% at −0.3 V vs. RHE. From the viewpoint of product selectivity, N2H4 which is the by-product of the NRR was only 5.8 × 10−13 mol s−1 cm−2, which was appreciably lower than the NH3 production rate.
There is also a further study that changed the ratio of the number of metals of polyoxometalate (POM)-based MOFs. Wang, Ma, Kan and co-workers reported POM-based MOFs, namely, FexCoyMOF-P2W18 which were employed as catalysts for the NRR.88 FexCoyMOF-P2W18 was prepared by changing the Fe/Co ratio and named accordingly as CoMOF-P2W18, FeCo2MOF-P2W18, FeCoMOF-P2W18, Fe2CoMOF-P2W18 and FeMOF-P2W18. The results demonstrated that FeCoMOF-P2W18 outperformed other MOFs with an NH3 yield rate and FE of 47.04 μg h−1 mgcat−1 and 31.76% at −0.4 V vs. RHE, respectively. The NRR performance of FeCoMOF-P2W18 was also compared with that of pristine FeCoMOF and the former was found to be notably higher than the latter (Fig. 4c). Furthermore, pure P2W18 did not show NRR activities, which indicated that the NRR activities of bimetallic MOFs were greatly improved. During the NRR, N2H4 was not identified, which indicated the high selectivity of NH3. To substantiate the recyclability of FexCoyMOF-P2W18, cycling tests were performed and the current densities during the NRR process were measured. The results of cycling tests revealed that NH3 yield rates and FEs remained stable over five cycles. No detectable changes in current densities were found during the 12 hours of the NRR experiment.
CoFe-MIL-88A was in situ grown on 2D V2CTx MXene (hereinafter CoFe-MIL-88A/V2CTx) by Li, Zhang and co-workers for use as a NRR electrocatalyst.89 CoFe-MIL-88A/V2CTx displayed a NH3 yield rate and FE of 29.47 μg h−1 mgcat−1 at −0.3 V vs. RHE and 28.86% at −0.1 V vs. RHE, respectively. These results were far superior to those for pristine CoFe-MIL-88A and V2CTx, and the NH3 was entirely derived from N2. The selectivity of NH3 was also validated by assessing the amount of N2H4 production and during eNRR measurement; no N2H4 was identified. The high NRR activity can be ascribed to the coating of CoFe-MIL-88A by V2CTx, thereby rendering the surface hydrophobic which inhibited the competitive HER pathway. The MOF/MXene composites also indicated high stability during electrocatalytic tests.
Research focusing on metal single atoms within the organic linker of multiple metal MOFs was also reported. Gao, Zhang and colleagues reported the NRR activities of Fe-TCPP, Co-TCPP and Zn-TCPP MOFs that changed with differing metal ions residing inside the porphyrin macrocycle.90 0.1 M HCl was utilized as electrolyte. The NH3 production performance and FE of Fe-TCPP, Co-TCPP and Zn-TCPP MOFs were 44.77 μg h−1 mgcat−1/16.23%, 28.3 μg h−1 mgcat−1/11.58% and 19.59 μg h−1 mgcat−1/6.37%, respectively. Each performance was higher than that of CP, and Fe-TCPP showcased the best NRR performance among all MOFs reported before this study. DFT calculation results suggested the interaction strength and bond distance between nitrogen molecules and each metal atom, revealing that the activation state of nitrogen molecules is favorable in the order of Fe-TCPP, Co-TCPP, and Zn-TCPP. In terms of the NRR pathway, the calculation also indicated that only Fe-TCPP followed the associative distal pathway, and the others proceeded via the associative alternating pathway. This difference was caused by the bond distance between nitrogen molecules and active sites and the degree of activation. The reaction rate was determined by the initial step of hydrogeneration of the nitrogen molecules, and the Gibbs free energy for this step is the lowest for Fe-TCPP and highest for Zn-TCPP, which was consistent with the experimental results.
Another research study on MOFs including porphyrin linkers by Solla-Gullón, Tatay, Marti-Gastaldo and colleagues probed the NRR activities with the PCN-224 family of Zr cluster-based MOF electrocatalysts.91 PCN-224 indicated brilliant stability under NRR conditions. Furthermore, the MOFs constructed with Co2+, Ni2+ and Cu2+-metalated porphyrin linkers exhibited the FEs of 11.4, 32.3 and 21.7% at −0.04 V vs. RHE, respectively. However, PCN-224-Ni was found to elute during the NRR, which was suggested by ICP-analysis. Hence, PCN-224-Ni faced a stability issue. Moreover, fluorine was introduced into PCN-224-Ni to compare the importance of functional groups in the porphyrin linker. As a result, FE was improved from 32.3% to 34.5% at −0.04 V vs. RHE and the NH3 yield rate was 0.9 μg h−1 mgcat−1. From the viewpoint of stability, the durability of PCN-224-Ni(F) was enhanced and during the electrocatalytic reaction for 24 h, the morphology of the crystal did not show any obvious change.
Fig. 3 Methods for introducing hydrophobicity: (a) schematic illustration of Ag nanocubes deposited on the Au electrode followed by covering with ZIF-71. Adapted with permission from ref. 92. Copyright 2018, The American Association for the Advancement of Science. (b) M@ZIF-Oam electrocatalyst prepared by utilizing ZIF-71 for coating the metallic electrocatalysts followed by modifying OAm to impart hydrophobicity. Adapted with permission from ref. 93. Copyright 2020, American Chemical Society. (c) HT Au@MOF prepared by inserting AuNPs as guests inside the MOF pores followed by covering the MOF surface with organosilicon. Adapted with permission from ref. 94. Copyright 2021, Elsevier B.V. (d) PdCu@UiO-S@PDMS prepared by encapsulating PdCu NPs into the pores of UiO-S followed by coating with hydrophobic PDMS. Reprinted with permission from ref. 95. Copyright 2023, Wiley-VCH GmbH. |
In the following year, Li, Du and colleagues utilized a different zeolite imidazole framework, ZIF-8, to encapsulate NPG.96 The performance of this composite catalyst (termed NPG@ZIF-8) was examined by using a N2-saturated 0.1 M Na2SO4 water solution as the electrolyte. When the experiment was carried out at −0.8 V vs. RHE, the NH3 production was as high as 28.7 ± 0.9 μg h−1 cm−2. On the other hand, the highest FE obtained was 44% and the selectivity of ammonia was 98% at −0.6 V vs. RHE (Fig. 4a). NPG@ZIF-8 also displayed brilliant stability even if it was exposed to a constant potential of −0.6 V vs. RHE for 12 h, showing only a marginal decrease in current density.
Fig. 4 Catalytic performances of different electrocatalysts in terms of the NH3 yield rate and FE: (a) NPG encapsulated into ZIF-8 (NPG@ZIF-8). Adapted with permission from ref. 96. Copyright 2019, John Wiley and Sons. (b) Cu-MOF and Ce-MOF grown on the surface of copper mesh (Cu@Cu-MOF and Cu@Ce-MOF). Adapted with permission from ref. 79. Copyright 2021, American Chemical Society. (c) Polyoxometalate-based MOFs (FexCoyMOF-P2W18 and FexCoyMOF). Adapted with permission from ref. 88. Copyright 2023, Wiley-VCH GmbH. (d) UiO-66 that was made defective by introducing HPMo (UiO-66-xHPMo). Adapted with permission from ref. 82. Copyright 2024, Elsevier B.V. (e) Zn-N2S2-MOF, Zn-N3S-MOF, and Zn-N4-MOF. Adapted with permission from ref. 84. Copyright 2023, Wiley-VCH GmbH. (f) xAu/MIL-101(Fe), x = 5, 10, 15 or 20%. Adapted with permission from ref. 97. Copyright 2024 Wiley-VCH GmbH. |
ZIF-71 composite catalysts with different catalysts incorporated inside have also been presented. A study by Ling et al. reported the use of ZIF-71 for encapsulating PtNSs on Au electrodes (referred to as Pt/Au@ZIF).98 The electrolyte was an anhydrous THF solution containing LiCF3SO3 with 1 wt% ethanol serving as a proton source. First, PtNS size dependence was confirmed, and the results showcased that the smaller PtNSs ranging from 60 nm to 231 nm exhibited higher FE and NH3 yield rates. Second, the NRR performance was evaluated by utilizing 60 nm Pt/Au@ZIF at various potentials from −1.7 V to −3.3 V vs. Ag/AgCl. Furthermore, the results revealed that the highest FE and NH3 yield rate were 44.8 ± 4.2% and 161.9 ± 16.7 μg h−1 mgcat−1, respectively under ambient conditions at −2.9 V vs. Ag/AgCl. Pt/Au@ZIF also demonstrated a FE of over 40% and an ammonia yield rate of over 130 μg h−1 mgcat−1 across a wide range of potentials. They also compared Pt/Au@ZIF with Pt/Au, Au@ZIF and dcim-Pt/Au, which was covered with monolayer of linker molecules (dcim) and the results displayed that Pt/Au@ZIF possesses the best NRR activities.
Apart from MOFs, another method that enhanced hydrophobicity was also endeavored. A study by Ling et al. utilized ZIF-71 for coating the metallic electrocatalysts and then modified OAm to impart hydrophobicity (denoted as M@ZIF-OAm) (Fig. 3b).93 Additionally, butanol molecules were added into ZIF-71 to limit the amount of accessible water molecules. First, CV measurements of M@ZIF-OAm with butanol, M@ZIF with butanol, and M@ZIF-OAm without butanol were conducted with dry THF, trifluoromethanesulfonate and 1 wt% butanol as electrolytes including 0.25 wt% water under N2 and Ar-bubbling. M@ZIF-OAm with butanol showed a reduction peak at −2.5 V vs. Ag/AgCl, indicating NRR activity. On the other hand, other two electrodes exhibited no peak at −2.5 V, and the peak at −1.3 V confirmed the occurrence of the HER. The results revealed that both coating with OAm and adding butanol were crucial for adsorbing N2 and adjusting the interaction of water molecules with the electrocatalyst. Second, the NRR of M@ZIF-OAm and M@ZIF was performed at −2.9 V for 6 hours under N2-bubbling upon varying the amount of added water. The highest ammonia yield rate and FEs in the presence of water between 0 wt% and 0.25 wt% were 45.3 ± 2.9 μg h−1 mg−1 and 18%, respectively, in the presence of 0.1 wt% water. Both values were by far higher than those observed in the absence of water. In contrast to these results, the NRR performance of uncoated M@ZIF illustrated the best NH3 production rate and FE were 21 ± 7.2 μg h−1 mg−1 and 5 ± 0.2%, respectively, under no-water conditions, which substantiated the necessity of hydrophilicity. On another note, the stability of M@ZIF-OAm was confirmed by monitoring the values of current density at −2.9 V during 80 cycles, and there was relatively no change in this value.
In 2022, the Du group prepared a 3D MOF by using Zn(II) ions and 3,3′-bis(mercapto)biphenyl-p,p′-dicarboxylic acid, followed by inserting AuNPs as guests inside the MOF pores.94 Moreover, to enhance hydrophobicity, the surface of the MOF was covered with organosilicon (PMX-200–500CS), and this substance was named HT Au@MOF (Fig. 3c). The NRR performance of HT Au@MOF achieved the highest NH3 yield and FE of 49.5 μg h−1 mgcat−1 and 60.9%, respectively, at −0.3 V vs. RHE. The high selectivity of NH3 was also established by the absence of N2H4 detection. The performance of HT Au@MOF surpassed that of Au@MOF, and the LSV tests under Ar-saturation also indicated that the current density of HT Au@MOF was strikingly lower than that of Au@MOF, which highlighted that coating organosilicon could suppress the HER. Furthermore, they implemented other three MOFs (DTU-67, 2,5-TP and MIL-101-SO3) in the same reaction. The experimental findings revealed that there is a negative correlation between the size of AuNPs and the NRR activities. The size of AuNPs that were encapsulated into MOFs in ascending order was HT Au@MOF, HT Au@DTU-67, HT Au@2,5-TP and HT Au@MIL-101-SO3. Meanwhile, the order of NRR activities in these materials followed HT Au@MOF > HT Au@DTU-67 > HT Au@2,5-TP > HT Au@MIL-101-SO3. The reaction was carried out on the surface of AuNPs, and the mechanism confirmed that the reaction progressed via the associative distal pathway because no N2H4 was observed. The stability of HT Au@MOF was also confirmed for six cycles of the NRR at −0.3 V vs. RHE with virtually no change in the NH3 yield rate and FE after the sixth catalytic run.
The Jiang group encapsulated PdCu NPs into the pores of UiO-S followed by coating with PDMS (hereinafter PdCu@UiO-S@PDMS) and used it as catalyst for the NRR.95 The PDMS coating endowed PdCu@UiO-S with hydrophobicity (Fig. 3d). PdCu@UiO-S@PDMS exhibited the highest NH3 yield rate of 20.24 μg h−1 mgcat−1 and a FE of 13.16% at −0.25 V vs. RHE. Comparative analyses of the performances of PdCu@UiO-S@PDMS, UiO-S@PDMS and Pd@UiO-S@PDMS substantiated that PdCu NPs functioned as active sites for the NRR. The study revealed that the optimal NRR performance was achieved in the presence of a 10% molar ratio of Cu in CuPd NPs. On another note, the performances of PdCu@UiO-S@PDMS and PdCu@UiO@PDMS were compared to evince the importance of –SO3H functional groups, showing the superiority of PdCu@UiO-S@PDMS over PdCu@UiO@PDMS. To assess hydrophobicity, the contact angles of PdCu@UiO-S and PdCu@UiO-S@PDMS were determined to be 2.2° and 137.3°, respectively, which illustrated that coating PDMS enhanced the hydrophobicity, which eventually impeded the HER. Furthermore, PdCu@UiO-S@PDMS featured high selectivity and stability.
The Yin group loaded Au NPs on MIL-101(Fe) to boost its electrocatalytic activity for the NRR.97 The ratio of AuNPs and MIL-101(Fe) was changed and the compounds were termed xAu/MIL-101(Fe), x = 5, 10, 15 or 20%. Among xAu/MIL-101 (Fe), 15%Au/MIL-101 (Fe) exhibited the best performance with the highest NH3 yield rate of 46.37 μg h−1 mgcat−1 and a FE of 39.98% at −0.4 V vs. RHE (Fig. 4f). The NRR activity was also inspected by changing the potential, which revealed that the NRR performance was highest at −0.4 V owing to the strongest absorbance intensity at −0.4 V. Au NPs demonstrated high HER activity but due to the increased N2 concentration provided by MIL-101(Fe), the NRR activity was dramatically enhanced by hindering the HER pathway.
MOF composites combining MOFs and carbon materials were also explored. A study on MOFs with CNTs for the NRR was reported by Ding et al. in the same year.99 In this contribution, four MOFs (UIO-66, BIT-58, CAU-17 and MIL-101(Fe)) were utilized for inserting CNTs and N-doped CNTs (hereinafter CNT/NCNT@MOF), and all electrical measurements were done in 0.05 M H2SO4. All of the MOF compounds featured markedly superior NRR activities compared to pristine CNTs and NCNTs. From these results, MOFs contributed to hydrophobic suppression of the HER and enrichment of N2. On another note, CNT@MOFs preserved their crystallinity after six cycles, which suggested that CNT@MOFs have notable stabilities.
Other MOF and CNT compounds are also utilized as NRR catalysts. Yin and co-workers prepared UiO-66-NH2 deposited on CNTs (UiO-66-NH2/CNT-X, where X denotes the amount of CNTs used to synthesize the composite: 0.05, 0.10, 0.15 or 0.20 g) and used them as NRR catalysts.100 A 0.1 M Na2SO4 solution was used as the electrolyte and MOF/CNT composites served as working electrodes. All results of the NRR showed that UiO-66-NH2/CNT-0.15 achieved the highest NH3 yield rate of 1.854 × 10−10 mol s−1 cm−2 (FE: 37.38%), whereas UiO-66-NH2/CNT-0.05 showed the best FE of 45.51% (NH3 yield rate: 1.091 × 10−10 mol s−1 cm−2), both at −0.3 V vs. RHE. The amount of N2H4 as a by-product of the NRR was also measured by utilizing UiO-66-NH2/CNT-0.15, and the highest yield of N2H4 was 2.84 × 10−13 mol s−1 cm−2 at −0.3 V vs. RHE, which was significantly lower than the amount of NH3. To ascertain stability, UiO-66-NH2/CNT-0.15 was recycled in chronoamperometric tests 5 times and the MOF/CNT composites demonstrated impressive stability (the relative NH3 yield rate and FE were 92.97% and 94.89%, respectively, compared with the first measurement). To monitor the reaction pathway, in situ ATR-SEIRAS was utilized, and the results confirmed the formation of hydrogenated nitrogen species (N2H1–4) on the MOF/CNT composite surface. From this fact, it was revealed that the pathway follows the associative alternating pathway.
Efforts to improve conductivity have also been made. Ren, Liu and colleagues utilized ZIF-67 that was supported on Ti3C2 (denoted as ZIF-67@Ti3C2) for the NRR.101 The performance of ZIF-67@Ti3C2 was measured in 0.1 M KOH and the highest NH3 production rate and FE were 6.52 μmol h−1 cm−2 and 20.2%, respectively, at −0.4 V vs. RHE. For comparative analysis, control experiments were also performed with individual ZIF-67 and Ti3C2, which revealed that the NH3 yields were 1.61 μmol h−1 cm−2 and 2.77 μmol h−1 cm−2 for ZIF-67 and Ti3C2, respectively.
Moreover, composites including MOFs and sulfide materials were also adopted as catalysts for the NRR. A study by the Wang group adopted ZIF-71 for encapsulation of the MoS2 nanoflower (hereinafter MoS2@ZIF-71), which enhanced the NRR activity in comparison to both of the pristine materials.102 CP covered with MoS2@ZIF-71 achieved the highest NH3 production rate of 56.69 μg h−1 mgMoS2−1 and the FE was 30.91% at −0.2 V vs. RHE. These results also highlighted the contributory role of the MOF coating as a hydrophobic barrier to restrain the HER.
Lang and co-workers prepared the MIL-101(Fe)/MoS3 composite for use as an NRR catalyst.103 The best NRR performance was achieved with an NH3 yield rate and FE of 25.7 μg h−1 mgcat−1 and 36.71%, respectively, at −0.1 V vs. RHE, which exceeded those of MIL-101(Fe) (9.38 μg h−1 mgcat−1 and 13.39%) and MoS3 (0.75 μg h−1 mgcat−1 and 0.03%). From the standpoint of the conversion rate from N2 to NH3, MIL-101(Fe)/MoS3 furnished almost no N2H4 by-product from the NRR. Additionally, MIL-101(Fe)/MoS3 showed outstanding stability during five cycles and electrolysis for 12 h at −0.1 V vs. RHE, which was assessed by X-ray photoelectron spectroscopy (XPS) measurements.
In addition to the composites mentioned before, composites based on MOFs and oxides were also leveraged as electrocatalysts for the NRR. Li, Lin and colleagues prepared ZIF-8 and combined it with CeO2 nanorods, which were synthesized by calcining Ce-MOF, (hereinafter CeO2-ZIF-8), and the CeO2-ZIF-8 dispersion was applied to CP and subsequently used as the NRR electrode.104 The NRR performance showcased the highest NH3 yield rate and FE of 2.12 μg h−1 cm−2 and 8.41% at −0.5 V vs. RHE, respectively. The stability of CeO2-ZIF-8 was validated through recycling tests. After the sixth cycle, the NH3 yield rate and FE were 2.10 μg h−1 cm−2 (99.06% compared with the first cycle) and 8.21% (97.62% compared with the first cycle), which implied exceptional stability. Furthermore, when the NRR activity was measured, no N2H4 by-product could be traced, which confirmed the high selectivity towards NH3.
In the preceding discussion, numerous MOFs that have been employed hitherto as electrocatalysts for the NRR have been highlighted. The performance of different MOF electrocatalysts in the NRR is summarized in Table 1. Striving for better NRR activities, scientists not only adopted a diverse range of metals and linkers to construct MOFs but also introduced defects to facilitate accessibility of the active sites. Besides, other materials like CNTs, MXenes, etc. have been used to prepare composites with MOFs for synergistic enhancement of electrocatalytic activities. On another note, MOFs have also been utilized as coatings on other electrocatalysts to impart hydrophobicity, thereby suppressing the HER, which is one of the factors impeding the NRR.
Year | Electrocatalysts | NH3 yield rate | Potential/current density | Ref. |
---|---|---|---|---|
2017 | MOF (Fe) | 2.12 × 10−9 mol s−1 cm−2 | −1.2 V vs. Ag/AgCl | 74 |
MOF (Co) | 1.64 × 10−9 mol s−1 cm−2 | −1.2 V vs. Ag/AgCl | 74 | |
MOF (Cu) | 1.24 × 10−9 mol s−1 cm−2 | −1.2 V vs. Ag/AgCl | 74 | |
2018 | Ag–Au@ZIF | ∼10 pmol s−1 cm−2 | −2.9 V vs. Ag/AgCl | 92 |
2019 | NPG@ZIF-8 | 28.7 ± 0.9 μg h−1 cm−2 | −0.8 V vs. RHE | 96 |
JUC-1000/CC | 24.7 μg h−1 mgcat−1 | 0.4 V vs. Ag/AgCl | 75 | |
2020 | Co3Fe-MOF | 8.79 μg h−1 mgcat−1 | −0.2 V vs. RHE | 85 |
NiFe-MOF | 9.3 μg h−1 mgcat−1 | −345 mV vs. RHE | 86 | |
MIL-100(Al) | 43.15 μg h−1 mgcat−1 | −0.7 V vs. RHE | 71 | |
NH2-MIL-88B-Fe | 1.205 × 10−10 mol s−1 cm−2 | −0.45 V vs. RHE | 76 | |
Pt/Au@ZIF | 161.9 ± 16.7 μg h−1 mgcat−1 | −2.9 V vs. Ag/AgCl | 98 | |
Co3(HHTP)2/CP | 22.14 μg h−1 mgcat−1 | −0.40 V vs. RHE | 77 | |
HKUST-1/CP | 46.63 μg h−1 mgcat−1 | −0.75 V vs. RHE | 78 | |
M@ZIF-Oam | 45.3 ± 2.9 μg h−1 mgcat−1 | −2.9 V vs. Ag/AgCl | 93 | |
2021 | Cu@Cu-MOF | 10.23 μg h−1 cm−2 | −0.2 V vs. RHE | 79 |
Cu@Ce-MOF | 14.83 μg h−1 cm−2 | −0.2 V vs. RHE | 79 | |
CNT@UiO-66 | 3.811 μg h−1 mgcat−1 | −0.55 V vs. RHE | 99 | |
NCNT@UiO-66 | 6.081 μg h−1 mgcat−1 | −0.6 V vs. RHE | 99 | |
CNT@BIT-58 | 4.135 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
NCNT@BIT-58 | 8.108 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
CNT@CAU-17 | 11.92 ± 0.083 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
NCNT@CAU-17 | 13.30 ± 0.01 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
CNT@MIL-101(Fe) | 5.514 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
NCNT@MIL-101(Fe) | 6.97 μg h−1 mgcat−1 | −0.45 V vs. RHE | 99 | |
Fe-TCPP | 44.77 μg h−1 mgcat−1 | −0.3 V vs. RHE | 90 | |
Co-TCPP | 28.3 μg h−1 mgcat−1 | −0.3 V vs. RHE | 90 | |
Zn-TCPP | 19.59 μg h−1 mgcat−1 | −0.3 V vs. RHE | 90 | |
UiO–Zr–Ti | 1.16 ± 0.058 × 10−10 mol s−1 cm−2 | −0.3 V vs. RHE | 87 | |
ZIF-67@Ti3C2 | 6.52 μmol h−1 cm−2 | −0.4 V vs. RHE | 101 | |
MoS2@ZIF-71 | 56.69 μg h−1 mgMoS2−1 | −0.2 V vs. RHE | 102 | |
2D In-MOF | 79.20 μg h−1 mgcat−1 | −0.5 mA cm−2 | 80 | |
2022 | HT Au@MOF | 49.5 μg h−1 mgcat−1 | −0.3 V vs. RHE | 94 |
Defective UiO-66-NH2 | ∼2.071 × 10−10 mol s−1 cm−2 | −0.39 V vs. RHE | 81 | |
MIL-101(Fe)/MoS3 | 25.7 μg h−1 mgcat−1 | −0.1 V vs. RHE | 103 | |
CeO2-ZIF-8 | 2.12 μg h−1 cm−2 | −0.5 V vs. RHE | 104 | |
2023 | FeCoMOF-P2W18 | 47.04 μg h−1 mgcat−1 | −0.4 V vs. RHE | 88 |
PdCu@UiO-S@PDMS | 20.24 μg h−1 mgcat−1 | −0.25 V vs. RHE | 95 | |
UiO-66-NH2/CNT-0.15 | 1.854 × 10−10 mol s−1 cm−2 | −0.3 V vs. RHE | 100 | |
2024 | UiO-66-2HPMO | 36.61 μg h−1 mgcat−1 | −0.3 V vs. RHE | 82 |
Defective Al-Fum | 53.9 μg h−1 mgcat−1 | −0.15 V vs. RHE | 83 | |
Zn-N2S2-MOF | 25.07 ± 1.57 μg h−1 cm−2 | −0.3 V vs. RHE | 84 | |
15% Au/MIL-101(Fe) | 46.37 μg h−1 mgcat−1 | −0.4 V vs. RHE | 97 | |
CoFe-MIL-88A/V2CTx | 29.47 μg h−1 mgcat−1 | −0.3 V vs. RHE | 89 | |
PCN-224-Ni(F) | 0.9 μg h−1 mgcat−1 | −0.04 V vs. RHE | 91 |
NiRux-BDC | Ru-introduced Ni2(OH)2(C8H4O4); |
BTC | Benzene-1,3,5-tricarboxylate; |
NiCo-UMOFN | Ni and Co ultrathin MOF nanosheet; |
CoBDC | Co2(OH)2(C8H4O4); |
Fc | Carboxyferrocene; |
Cu-BTC | Cu3(BTC)2(H2O)x; |
Cu-bipy-BTC | [Cu2(OH)(bipy)2(BTC)3·2H2O]n; |
Bipy | 2,2′-Bipyridine; |
HITP | 2,3,6,7,10,11-Hexaiminotriphenylene; |
BmimBF4 | 1-Butyl-3-methylimidazolium tetrafluoroborate; |
BmimOTf | 1-Butyl-3-methylimidazolium trifluoromethanesulfonate; |
BmimPF6 | 1-Butyl-3-methylimidazolium hexafluorophosphate; |
BmimClO4 | 1-Butyl-3-methylimidazolium perchlorate; |
MvK | Mars-van Krevelen; |
TMN | Transition metal nitrides; |
EDTA | Ethylenediaminetetraacetic acid; |
LiCF3SO3 | Lithium trifluoromethanesulfonate; |
RHE | Reversible hydrogen electrode; |
NPG | Nanoporous gold; |
SG | Sodium gluconate; |
GA | Glucaric acid; |
NiFe-MOF | Ni0.81Fe0.19((C12H6O4)(H2O)4); |
PtNS | Pt nanosphere; |
HHTP | 2,3,6,7,10,11-Hexahydroxytriphenylene; |
OAm | Oleylamine; |
CNT | Carbon nanotube; |
AuNPs | Au nanoparticles; |
FexCoyMOF-P2W18 | [FexCoy(Pbpy)9(ox)6(H2O)6][P2W18O62]·3H2O; |
UiO-S | UiO-66-SO3; |
PDMS | Polydimethylsiloxane; |
Al-Fum MOF | Aluminum-fumarate MOF; |
HPMo | Phosphomolybdic acid |
This journal is © The Royal Society of Chemistry 2024 |