Structure-activity relationship of defective electrocatalysts for nitrogen fixation

Yusheng Wang Nan Yang Xue Xin Yingjie Yu Yuao Wei Baoli Zha Wenjing Liu

Citation:  Yusheng Wang, Nan Yang, Xue Xin, Yingjie Yu, Yuao Wei, Baoli Zha, Wenjing Liu. Structure-activity relationship of defective electrocatalysts for nitrogen fixation[J]. Chinese Chemical Letters, 2023, 34(7): 107841. doi: 10.1016/j.cclet.2022.107841 shu

Structure-activity relationship of defective electrocatalysts for nitrogen fixation

English

  • Ammonia (NH3) is not only a basic chemical used as fertilizer in agricultural production, but also an important chemical composition used in chemical production such as refrigerants and explosives [1,2]. More than 80% of the ammonia is obtained by Haber-Bosch reaction which requires both high temperature and high pressure [3]. In addition, this reaction consumes energy which accounts for about 1% of the world's annual energy consumption and is not conducive to environmental protection and sustainable economic development [4]. Moreover, most of the hydrogen sources for ammonia synthesis are obtained through multi-step reaction of natural gas, resulting in a large amount of carbon dioxide emissions [5]. Considering these problems of energy consumption and gas emission, it is meaningful to design eco-friendly and low-carbon methods for synthesizing ammonia to enhance N2 fixation under mild conditions [6,7].

    So far, some methods such as nitrogenase enzymes technology [8], photocatalysis [9] and plasma technology [10] have been explored for nitrogen fixation under mild conditions. For example, Brown et al. used nanocrystalline CdS as photosensitizer to realize the nitrogen fixation by driving MoFe protein to synthesize ammonia simulating the biological nitrogen fixation process [8]. In terms of photocatalytic synthesis of ammonia, Feng et al. have recently synthesized novel surface oxygen vacancies (OVs) modified Bi2O2CO3 (BOC/OVs) with micronanosheet structure which exhibited good performance in photocatalytic NRR [11]. Plasma-assisted technology involving plasma reactor has been considered as a fast and convenient method to realize the collection of ammonia at ambient conditions, but its development is hindered by the unsatisfactory efficiency [12]. Therefore, it is still necessary to explore new nitrogen fixation technologies with high efficiency and high ammonia yield [1,13].

    In recent years, the electrocatalytic synthesis of ammonia by utilizing water as the hydrogen source under ambient temperature and pressure has received unprecedented attention [14]. By this means, it is convenient to use renewable electricity in a distributed way for NRR through mature fuel cells and water electrolyzer technology. However, the N2 molecule not only possess a strong non-polar N≡N triple bond and the high ionization energy (940.95 kJ/mol), but also the absence of permanent dipoles in its molecules leads to a low proton affinity (493.8 kJ/mol), making hydrogenation of N2 becomes a challenge [15]. Electrocatalytic reduction of nitrogen to ammonia is a multi-step reaction, which involves the transfer of six electrons and six protons. The equilibrium potential of nitrogen and hydrogen evolution reaction (HER) at acid conditions during electrocatalytic reduction is as follows [16]:

    (1)

    (2)

    The potentials of these two equilibrium reactions, Eqs. 1 and 2, are similar. In principle, when a fairly negative potential is applied to the electrode, the N2 reduction reaction would happen at room temperature and atmospheric pressure [16]. Compared with the potential E1 of two electrons transfer in hydrogen evolution reaction, E0 of nitrogen reduction to ammonia represents the average value of multiple steps in this process [17]. Therefore, in the actual electrocatalytic reaction, hydrogen evolution reaction will dominate and lead to a low Faradaic efficiency (FE) [18]. When the current density is very low, there will be a relatively high FE and consequently low ammonia production rate [19]. As a result, designing high-efficient, high-active and high-selective electrocatalysts to enhance the nitrogen fixation with high ammonia yield and FE is desirable [20].

    Defect engineering, through precisely controlling the defect features (concentration, morphology, size, etc.), can affect the performance of electrocatalysts [21,22]. The advantages of such strategy are as follows: (1) The electrocatalysts with larger surface areas can be obtained by creating defects or modifying the shape structure to speed up electron transfer from the electrode to the reaction interface to further accelerate the electrocatalytic reaction rates [23-25]. For example, Wang et al. [26] fabricated porous gold film (pAu/NF) directly loaded on nickel foam as electrocatalysts. The three-dimensional porous structure of pAu/NF possesses a large specific surface area and abundant active sites to enhance the adsorption of nitrogen and accelerate the electron and proton transfer in the catalytic reaction, thus greatly enhancing the catalytic performance of the reaction. (2) The introduction of doping heteroatoms or vacancies on the surface of catalysts could adjust the electron density of active sites to increase the nitrogen adsorption capacity and effectively reduce the overpotential during the catalytic process [27]. For example, Fan et al. [28] achieved the crystalline phase transition of Mo2C from hexagonal to cubic by B doping into Mo2C phase. The B-doped cubic Mo2C has stronger hybridization with the 2p orbitals of the active adsorbed nitrogen atom in the density state near the Fermi energy level, which indicates that the B doping enhances the bonding between the catalytic material and the reacting species, thus favoring the stabilization of the adsorbed species and lowering the NRR reaction energy barrier. In this review, we summarize the latest progress about the applications of defects in the electrocatalytic synthesis of ammonia under mild conditions. We first introduce the concept, classifications and characterization of defects. Secondly, the approaches of introducing defects into various electrocatalytic nanomaterials are analyzed. Thirdly, a summary of these defects commonly used in NRR is provided. These defects are divided into five types (vacancy, dopant, amorphism, edge/corner, and porousness) and their influence on the performances of the electrocatalysts is explained in detail (Scheme 1). Finally, our points of view on challenges and outlook in the development of NRR electrocatalysts are proposed.

    Scheme 1

    Scheme 1.  Schematic diagram of defects engineering for NRR (vacancy, dopant, amorphism, edge/corner, and porousness). The advantages of defects towards NRR can be summarized into the following four points: (1) adjusting electronic structure; (2) modifying physical structure; (3) increasing reaction active site; and (4) enhancing nitrogen adsorption.

    Defects play an important role in tuning the electronic states of electrocatalysts and enhancing reaction active sites, influencing the performances of electrocatalysts. Therefore, it is necessary to understand the definition of defects and classify the synthetic routes to create them. The term "defect" is generally defined as the distortion of periodic potential filed in crystal lattice structure [29]. According to the definition of geometry, defects are commonly divided into point defects, line defects, plane defects, and bulk defects, which fall into the category of zero-dimensional (0D), one-dimensional (1D), two-dimensional (2D) and three-dimensional (3D) defects, respectively [30,31]. The creation of defects in the catalysts depends on external conditions, synthetic methods or post-treatment methods, etc.

    Among the reported defect catalysts affecting NRR electrocatalytic activity, the defect species could be divided into vacancies, dopants, amorphisms, edges/corners, and porousness.

    Vacancies are formed because the lattice atoms at the lattice nodes break free from the binding of the surrounding atoms and deviate from the original position, thus changing the physicochemical properties of crystal materials. These atomic scale defects can be intrinsic or introduced into the surface of catalytic materials, which can effectively enhance the catalytic activity. For example, anion vacancy, especially OVs on metal oxide electrocatalyst have been widely reported, including MnO2 [32], NiCo2O4 [33], CeO2 [34] and TiO2 [35], which have been proven to be efficient active sites for the adsorption and activation of reactants. Consistent with the OVs, other anionic electrocatalysts such as nitrogen vacancies for Ni3N [36], g-C3N4 [37], and sulfur vacancies for CoS@FeS [38], CdS [39], and cationic vacancies in electrocatalytic materials such as iron vacancies [40], molybdenum vacancies [41] and copper vacancies [42] have also been reported with reliable performances.

    Heteroatom doping is formed by introducing foreign atoms [43]. Foreign atoms enter the lattice to either replace the original atoms into the normal lattice points, or occupy interstitial positions where there are no atoms. According to the different types of dopants, they can be divided into metal doping, such as Cu [44], Co [45], Ru [46] and non-metal doping such as N [47], P [48] and B [49]. In addition, the direct confinement of atoms or ions in the crystal lattice is also a type of doping. Ma et al. [50] designed atomic Pt catalysts with crystal lattice confinement in metal carbides and demonstrated that the confined atoms exhibited near-zero valence states and matched metal Pt electronic structure, thus providing a catalytic behavior similar to that of metal Pt adsorbed on *H on the surface of metal Pt, which is much better than that of Pt sites adsorbed on WCx. This provides a new idea for NRR to effectively improve the activation and adsorption of N2. In general, dopants can not only adjust the electron density distribution around active sites, but also may lead to structure transformation to improve the activity of electrocatalysts [51-54].

    Amorphism refers to materials lacking the characteristics of long-range order and periodicity. The accumulation of vacancies or doping defects may produce disordered crystal structure with exposed active sites, which can enhance the transport ability of electrons and protons during the catalytic process. Some recent reports show that amorphous catalysts have advantages in catalytic performance over crystalline catalysts such as Ni-P alloy [55] and NiOx-C [56], possibly due to their long-range disordered structure and abundant unsaturated active sites.

    The edge and corner atoms of nanocrystals possess high surface energy and activity due to the presence of unsaturated coordination bonds, which have been shown to significantly reduce the activation energy barriers of inert molecules such as N2 [57,58]. For example, the edge defects of catalysts such as MoS2 [59] and MXene [60] are demonstrated to tune electron density, which further enhances their electrocatalytic performance. Porous structures are usually caused by the loss of templates or metastable surface atoms during the post-synthesis of catalytic nanomaterials (pyrolysis, reduction, etching, etc.), which may not only provide abundant active sites but also enhance the enrichment of N2 and intermediates, prolong N2 residence time, accelerate the transformation of specific species, and shorten the electron transport path [61]. In summary, the defect sites have the ability to change the coordination environment and charge distribution of surrounding atoms, thus regulating the adsorption capacity of key intermediates, and further affecting the electrocatalytic activity.

    The characterization of defects at the atomic level is still challenging. Some advanced characterization techniques have been developed to understand the existence species and formation mechanism of defects. High angle annular dark field scanning transmission electron microscope (HAADF-TEM) can be used for visual defect detection. As shown in Figs. 1a and b, the white dot [62] and black hole [63] represent the existence of metal vacancies and doping sites, respectively, because the imaging intensity is proportional to the square of the atomic number. Other defects such as twin boundaries can also be captured (Fig. 1c) [64]. X-ray photoelectron spectroscopy (XPS), as an effective non-destructive surface measurement technology, can well display the electronic effect caused by surface defects through the shift of binding energy. As shown in Fig. 1d, the Rh 3d peak of a/c-Rh shifts slightly to a higher binding energy due to the existence of disordered atoms in the amorphous state [65]. Electron paramagnetic resonance (EPR) advances in detecting unpaired electrons, thus becoming an effective method to characterize vacancies. For instance, CoMnNiOx-S-A shows stronger EPR signal at g = 2.003 than that of CoMnNiOx-S indicating that rich OVs have been introduced into CoMnNiOx-S-A (Fig. 1e) [66]. Transient photoluminescence (PL) spectroscopy can qualitatively detect the vacancy concentration produced by photon exciton. For example, NbOx exhibits stronger photoluminescence intensity at 588 nm than Nb2O5, indicating its higher OVs concentration (Fig. 1f) [67]. X-ray diffraction (XRD) is also an effective method to define the crystal structure and change the law of catalytic materials. Usually, there are obvious peaks in the diffraction pattern of crystalline materials, while large envelope peaks in that of amorphous materials (Fig. 1g) [68]. Positron annihilation spectroscopy (PAS) provides direct information about the types and relative concentrations of defects and is a powerful technique for studying defects in materials. For instance, Liu et al. employed PAS to identify defect changes in CoSe2 nanosheets, with the distinct lifetime and intensity being slightly extended in Fig. 1h. This suggested improved defect dimension and content compared with CoSe2 solid materials [69]. X-ray absorption spectroscopy (XAS) including X-ray absorption near edge structures (XANES) and extended X-ray absorption fine structures (EXAFS) has been generally used to characterize the electronic and geometric structures of defects at the atomic scale. For example, by utilizing EXAFS spectra, Li et al. reported that atomically dispersed Fe atoms interact with N through coordination bonds (Fig. 1i) [70].

    Figure 1

    Figure 1.  Characterizations of defects in nanomaterials. (a) HAADF-STEM images of SA-Fe-NHP highlighting isolated Fe atoms. Reproduced with permission [62]. Copyright 2020, John Wiley and Sons. (b) HAADF-STEM images of sulfur vacancies on MoS2. Reproduced with permission [63]. Copyright 2020, American Chemical Society. (c) HRTEM image of defect rich SnO2, and the arrows indicate the grain-boundary surface terminations. Reproduced with permission [64]. Copyright 2021, Elsevier. (d) XPS spectra of a/c-Rh nanosheets (NSs) and c-Rh NSs. Reproduced with permission [65]. Copyright 2021, John Wiley and Sons. (e) EPR spectra of CoMnNiOx with different contents of oxygen vacancies. Reproduced with permission [66]. Copyright 2022, Elsevier. (f) PL spectra of NbOx and Nb2O5. Reproduced with permission [67]. Copyright 2022, Royal Society of Chemistry. (g) XRD patterns of amorphous and crystalline Ru/CNTs catalyst. Reproduced with permission [68]. Copyright 2022, American Chemical Society. (h) Positron lifetime spectra of ultrathin CoSe2 nanosheets and bulk CoSe2. Reproduced with permission [69]. Copyright 2014, American Chemical Society. (i) EXAFS spectra of Fe foil, Fe-PPy and Fe2O3 with different doping sites. Reproduced with permission [70]. Copyright 2021, Royal Society of Chemistry.

    In summary, different types of defects can be obtained by controlling some conditional factors via various approaches and have been proven to improve electrocatalysts' performances [71]. In the following sections, popular methods about how to introduce defects, for example, by etching with different etchants, by using plasma technology at different atmospheres and plasma concentrations and by pyrolyzing catalysts under different temperatures and pressures, will be introduced in detail.

    2.4.1   Etching

    Etching is a technology that removes materials by chemical reaction or physical impact, which is generally considered to be an effective and direct approach to induce defects on nanoparticles [72]. Various etching approaches (chemical etching, base etching, acid etching and plasma etching) have been developed in a more facile and convenient way, enabling a deliberate control over the concentrations and types of defects in electrocatalysts for NRR.

    As a well-known approach, chemical etching introduces etching agent to react with the catalysts and obtain target materials. The etchants such as halide ions and oxygen are commonly used in the preparation of precious metal catalysts [73]. For example, Wang et al. reported Pt3Sn catalyst with rich defects for formic acid oxidation utilizing chloridion and oxygen as etchants, platinum acetylacetonate (Pt(acac)2), SnCl2·2H2O, polyvinylpyrrolidone (PVP), and N, N-dimethylformamide (DMF) as reagents at high temperature and high pressure. With the increase of precursor concentration, the reduced Sn was diffused into the lattice of Pt due to the existence of high heat energy, and thus a large number of cubic Pt3Sn nanocubics with rich defects were formed during the reaction [74]. Furthermore, Xia et al. synthesized Pt-based icosahedral nanocages with twin boundaries defects by chemical etching Pd@Pt4.5 L nanomaterials as precursors for electrocatalytic reactions. The introduction of the etchants FeCl3 and KBr by controlling the etching dose and etching time not only selectively etched Pd out, but also increased the surface area of the material and the abundance of three-phase boundary [75]. As a result, the as-synthesized Pt icosahedral nanotubes had a better electrocatalytic performance than commercial Pt/C catalyst due to the existence of large surface, twin boundaries and {111} facets.

    Alkali etching and acid etching are relatively facile methods by introducing base or acid for inducing defects. In terms of alkali etching, Zhu et al. found a new nanomaterial with rich-carbon pentagonal defects (PD-C), whose precursor was prepared by pouring potassium hydroxide solution into fullerene suspension (Fig. 2a) [76]. Density functional theory (DFT) and experiments have proven that the inherent pentagonal defects of carbon nanomaterials can be used in electrocatalysis due to the good electrochemical reaction activity, strong oxygen binding ability and stability. MXene is a two-dimensional inorganic nanomaterial composed of transition metal carbides, nitrides or carbonitrides, with multiple atomic layer thicknesses, rich components and large specific surface area [77,78]. In terms of using hydrofluoric acid as an etchant, Meshkian et al. synthesized the vacancy-rich W1.33C nanosheets by selectively etching away the A(Al) layer and interlayer Sc/Y from (W2/3M1/3)2AlC (M = Sc and Y) nanosheets (Fig. 2b) [79]. Ma et al. successfully fixed the defective PdZn nanoparticles on the N-doped hollow carbon polyhedrons (NHCP) as NRR electrocatalysts by etching PdZn/NHCP with sulfuric acid [80]. PdZn/NHCP was synthesized by reacting Pt-based compounds with ZIF-8 during heat treatment. Accordingly, due to the rich exposed defects and porous carbon matrix, the surface area of the catalysts and the electron transport capacity were significantly improved.

    Figure 2

    Figure 2.  (a) Illustration of the KOH etching synthetic method of PD-C. Reproduced with permission [76]. Copyright 2019, John Wiley and Sons. (b) W1.33C MXene with ordered vacancies after selective etching and delamination. Reproduced with permission [79]. Copyright 2018, John Wiley and Sons. (c) Schematic illustration of the preparation process of SG-P and the plasma system. Reproduced with permission [83]. Copyright 2017, Elsevier. (d) Schematic illustration of the preparation procedure for Ni-implanted N-doped nanoporous carbons. Reproduced with permission [87]. Copyright 2020, John Wiley and Sons. (e) Synthetic procedure of Ce-Bi2WO6. Reproduced with permission [96]. Copyright 2021, American Chemical Society.

    Plasma etching, which utilizes the plasma to produce a high energy ionized gas and strong ultraviolet radiation, can be used to create defects by means of ion penetration, covalent bond destruction, ion impact, etc. [81]. After nanomaterials (graphene, metal compounds, etc.) were etched, not only abundant active sites can be obtained to increase the surface area of catalysts, but also more channels will be provided on the surface of catalysts to accelerate the mass transport of materials [82]. For example, using plasma etching technology to treat S-doped graphene (SG), Tian et al. created a large number of topological defects on graphene. The SG after plasma (SG-P) etching exhibited an obviously improved electrocatalytic activity (Fig. 2c) [83]. In another example, Zhang et al. reported a OVs-rich TiN electrocatalyst for NRR by plasma etching under Ar atmosphere for 10 min via using modified TiN which was ball milled for 8 h. The etching method exposes abundant vacancies in the material, allowing it to exhibit a higher NRR performance than pure TiN without OVs. Wang et al. introduced rich defects on (110) crystal face nanoplates of Bi through the plasma bombardment of Bi2O3 in a 30 Mega Hertz Radio Frequency excited quartz tube at a power density of 300 W for about 5 min [30]. Notably, the plasma-treated Bi showed much higher performance than those Bi samples without treatment. These above progresses demonstrate the plasma technology as an effective route to modify the electronic structure of Bi in order to get more (110) crystal faces, thus improving the active site of NRR and enhancing the catalytic performance.

    In summary, etching is an effective method to introduce defects in catalysts which is applicable to various kinds of materials (noble metals, transition metals, nonmetals, etc.). Chemical etching, alkali etching and acid etching belong to wet-etching, which is an etching method that peels off the etched material through the chemical reaction between the chemical etching solution and the etched material. In contrast, dry etching etches materials mainly by plasma, in which way, the types of defects can be controlled by reasonable selection of gas concentration, atmosphere type, etching time, pressure, etc. Generally speaking, the conditions required for etching are relatively mild, rendering possibility of the large-scale production.

    2.4.2   Pyrolysis treatment

    Pyrolysis treatment is a convenient and rapidly developed technology to introduce rich defects and various vacancies into carbides, nitrides and oxides to enhance the activity of catalysts [84,85]. The precursors such as Metal Organic Frameworks (MOFs), biomass, metal salts and metal free materials under different annealing treatment (e.g., temperature and time) in disparate gas flow (e.g., NH3, H2, N2 and Ar) could create diverse defects in nanostructures [43]. For instance, MOFs are composed of metal ions and organic connecters forming porous structures with the merits of large surface area, adjustable structure and high porosity. MOFs as precursors can be decomposed into more stable conductive carbon/metal based porous materials (such as metal oxide/phosphide) with rich defects by means of pyrolysis treatment [86]. Jiang et al. reported a single-atom Ni electrocatalyst by pyrolysis of pre-treated bimetallic precursor MgNi-MOF-74. MgNi-MOF-74 was first introduced with Mg ions to increase the spacing between adjacent Ni atoms. Moreover, the introduction of polypyrrole (PPy) as a ligand could effectively stabilize Ni atoms, and then the ppy@MgNi-MOF-74 was transformed into Ni-implanted N-doped porous carbon through pyrolysis treatment which showed good electrocatalytic activity (Fig. 2d) [87]. Chen et al. reported MoFe-phosphorus doped carbon (PC) microspheres with various vacancies as NRR electrocatalysts via carbonization of MoFe-MOFs in Ar at 300 ℃ for 2 h [88]. Another defective electrocatalyst CuPt/NC fabricated by pyrolyzing Cu-tpa MOF along with Pt precursor under Ar-H2 atmosphere at 700 ℃ for 3.5 h exhibited highly efficient NRR activity [89]. Huang et al. reported N-doped carbon aerogel with large amounts of defects via carbonization of zinc 2-methylimidazolate (ZIF-8) in N2 at 1000 ℃ for 3 h. Specifically, ZIF-8 precursor was first obtained by solvothermal treatment of Zn ions, and then pyrolyzed with carboxymethylcellulose composite gel (CMCC) to obtain N-doped porous carbon framework, which presented good electrical conductivity, thermal stability and chemical stability [90].

    Besides MOF materials, biomass fuel is also a common source of carbon. Bhuvanendran et al. prepared a highly efficient active CeO2 on biochar carbon (CeO2/BC700) through pyrolyzing CeO2/BC at 700 ℃ for 2 h [91]. CeO2, as a stable rare earth oxide, is not only rich in storage but also capable of transferring electrons, and therefore usually serve as a high-quality supporter to combine with matrix materials in the field of electrocatalysis [92]. The experimental results showed that the doping of nitrogen and phosphorus and the OVs of CeO2 contributed to high electrocatalytic activity. Moreover, the pyrolysis treatment of metal-free materials is also a common way to introduce defects. For example, Chen et al. successfully prepared amorphous carbon based nanosheets (BCN) with B-N substitutions and rich defects by annealing the organic precursors composed of PEG-2000, boric acid, and urea under Ar atmosphere at 900 ℃ for 2 h [93]. The theoretical and experimental results showed that B-N doping can not only provide the active trigger point and active center, but also effectively inhibit HER in NRR. In this way, more active centers and defects can be obtained by adjusting doping parameters reasonably to produce highly efficient catalysts under the same environmental conditions.

    In summary, pyrolysis treatment is a facile approach to anneal different nanomaterials (MOFs, metal compounds, nonmetal compounds, etc.) and introduce defects whose parameters (concentration, morphology, types, etc.) can be managed by controlling the anneal time, temperature, atmosphere, etc. Considering the wide varieties of nanomaterials, more catalysts with rich defects will be prepared via pyrolysis treatment for electrocatalytic reaction.

    2.4.3   Other alternative approaches

    In addition to the approaches described above, there are alternative approaches reported to induce defects such as chemical reduction method, "stacking and carving strategy", hydrothermal method, and Prussian-blue-analogue-sacrificed strategy.

    The chemical reduction method adopts reducing agents (NaBH4, or NaH2PO4, etc.) to reduce the metal ions, at the same time, the B or P and the reduced metals are converted to a monomeric state while precipitating to form an alloy under reducing conditions. This method has the advantage of simple equipment and easy operation, so it is usually used for the synthesis of amorphous alloys such as M-B or M-P (M = Fe, Co, Ni, Ru, etc.) and further in the study of electrocatalytic properties. Wang et al. [94] reported amorphous Ni0.5Fe0.5B nanosheets by utilizing NaBH4 as the reducing agent to reduce the mixed salt solution of FeCl3 and NiCl2 in one step at room temperature. This amorphous nanosheet exposes abundant vacancies and unsaturated coordination sites to make it exhibit excellent electrocatalytic performance in NRR.

    "Stacking and carving strategy", as an ideal approach to improve the performance of catalyst, can introduce defects on Pt-based nanomaterials via controlling the carving path and stacking route of crystal growth. In the "stacking" route, the nanoparticles formed in situ are arranged regularly in the process of crystal growth, thus resulting in a large number of defects in the junction region. While in the "carving" path, the particles formed in situ are etched regularly during the crystal growth process, and a large number of defects are produced in the eroded area. These two routes encouraged Wang's team to synthesize Pt-Fe and Pt-Zn nanoparticles with distinct 3D defects nanostructures [95].

    Hydrothermal method is an efficient and common way to prepare defective electrocatalysts. For example, the defective Ce-doped Bi2WO6 with nanoflower morphology was synthesized by mixing Bi(NO3)3·5H2O, Na2WO4 and Ce(NO3)3·6H2O under Teflon-lined autoclave at the temperature of 160 ℃ for 24 h (Fig. 2e) [96]. The as-synthesized catalyst Bi2WO6 has the crystal defects, which could facilitate N2 adsorption and further enhance NRR activity. Polyoxometalate based MOFs (POMOFs) are another typical example of excellent precursors, which combine the excellent redox activity of POMs with the porosity of MOFs [97]. Ma et al. achieved an efficient NRR electrocatalyst Fe1.89Mo4.11O7/FeS2@C by using PMo12@MIL-100(Fe)@PVP as the precursor via hydrothermal technique. The high activity during electrocatalytic process came from the formation of Mo-N bond, which could be inclined to adsorb *N-NH and be apart from *NH2 species thus reducing the reaction energy barrier. Manganese oxides are considered as a promising material because of their high abundance.

    KNi[Fe(CN)6] is a Prussian blue compound possessing a rigid and open framework structure and large-sized interstitial channels, which is commonly used as a sacrificial template in the synthesis of catalysts. Recently, via template sacrificed strategy, abundant Ni-Fe vacancies with rich-defect cubic nanoflowers (Ni-Fe-K0.23MnO2 CNFs-300) were successfully synthesized by annealing Ni-Fe-doped K0.23MnO2 cubic nanoflowers (Ni-Fe-K0.23MnO2 CNFs) under Ar atmosphere at 180 ℃ for 2 h. The Ni-Fe-K0.23MnO2 CNFs were obtained by hydrothermal treatment of KNi[Fe(CN)6], KMnO4 and KCl [98]. Theoretical calculation and experimental data manifested that the catalyst had excellent electrocatalytic performance over water decomposition, possessing much lower overpotential and better durability than other reported manganese oxide based electrocatalysts.

    The vacancy regulation concept emerges from the design philosophy of "less is more" and offers an opportunity to discover the anonymous quantum phenomenon. The defect fabrication in electrocatalysts with a suitable and well-defined vacancy is a key strategy for further optimizing the catalytic efficiency [99,100]. In addition, the incorporation of vacancies may favorably influence the electronic structure of catalysts, leading to fast charge transfer and optimal effective adsorption of intermediates for electrocatalytic reactions [101]. There are numerous types of vacancies including oxygen [102], sulfur [103], selenium [104] and metal-ion vacancies [63], which can ingeniously adjust the property of the surface adsorption for reaction intermediates thus enhancing the electrocatalytic activity [105].

    3.1.1   Oxygen vacancies

    OVs, the most prevailing anion vacancy in transition-metal oxides (TMOs) because of their low energy, are extensively accepted as a vital approach to perturb the surface electronic configuration and tailor the intrinsic physicochemical property of oxides [106]. The introduced OVs in TMOs may serve as active sites for the adsorption of reactants and intermediates, thereby decreasing the activation energy barrier and playing a significant role in promoting electrocatalytic reactions [107-111]. Moreover, the metastable electrons can be easily trapped in OVs to relocate to the anti-bonding orbital of N2 molecule, thus enhancing the cleavage energy of N≡N bond for corresponding catalytic reactions and fostering NRR activity. These attractive properties of OVs ensure their positive role in regulating the activities of various catalytic processes in the electrocatalysis field.

    Recently, numerous works have verified that TMOs can be utilized as NRR catalysts, especially with the assistance of OVs [112-119]. As we all know, Fe element plays an indispensable role in the Haber-Bosch process, and therefore Fe-based nanomaterials as well as their derivatives have been extensively investigated as NRR electrocatalysts. Zhang et al. proposed a task-specific ionic liquid approach to prepare α-Fe2O3 nanocubes with rich OVs for enhanced NRR performance, offering an ammonia yield of 32.13 µg h−1 mgcat−1 with FE of 6.63% at −0.3 V vs. RHE in 0.1 mol/L KOH [118]. Upon the introduction of the ionic liquid n-octylammonium formate in the reaction system, the reductant formic anion of ionic liquid in situ triggered rich OVs into α-Fe2O3 nanocubes. The formation of OVs left lone pair electrons on the surface of the catalyst, prompting the N2 polarity and enhancing the N2 activation and adsorption. Therefore, Fe2O3 coupled with OVs contributed to the exceptional activity of electrochemical NRR. Inspired by the Fe oxide-based materials, researchers have also studied other TMO-based materials as electrocatalysts for NRR such as MoO3 nanosheet, Nb2O5 nanofiber, multishelled hollow Cr2O3 microsphere, and MnO2 nanoarray, which all manifested potential in electrocatalytic NRR. However, many of these works only focus on selectivity or activity and have not taken the metal oxidation state into consideration. Therefore, it still remains unknown whether the OVs contribute to the activity improvement of NH3 synthesis or not.

    Some other works reported the importance of OVs concentration [103,120]. Han et al. tuned the OVs of TiO2 by controlling the synthetic temperature to study the influence of OVs concentration on NRR activity (Fig. 3a) [120]. At −0.12 V, rich OVs were obtained in anatase TiO2 with NH3 yield of about 3.0 µg h−1 mgcat−1 and FE of 6.5% (Fig. 3b). The pristine and rich OVs structures of TiO2 were simulated (Fig. 3c), and theoretical calculations further demonstrated the active role of OVs played in promoting the selectivity of nitrogen. The lowest energy reaction pathway was *N2 → *NNH → *NNH2 → *NHNH2 → *NH2NH2 → *NH2 + NH3 → *NH3 + NH3 → 2NH3. The first step *N2 → *NNH involved the cleavage of the nitrogen triple bond and the adsorption of hydrogen, and was considered to be a key step in electrolysis. DFT calculations showed that the energy barrier required for the anatase-OVs in the first hydrogenation step was 1.5 eV lower than that of the pristine precursor, indicating that the anatase-OVs can enhance the adsorption capacity of hydrogen and effectively inhibit HER. Thereby, the ammonium yield and Faraday efficiency were further improved (Fig. 3d).

    Figure 3

    Figure 3.  (a) O 1s XPS spectra of TiO2 under different temperatures. (b) NH3 yield rates and FEs of different catalysts at −0.12 V. (c) Optimized geometries (top-view) of pristine and the most stable surface oxygen vacancy containing anatase (101). The sky blue and red balls indicate Ti and O atoms, respectively and the free energy ΔG for NRR is shown in (d). Reproduced with permission [120]. Copyright 2019, Elsevier.

    In summary, the OVs existed in electrocatalysts exert a certain effect on bringing high NRR performance. However, it still needs more explanations to determine the OVs position, the exact electrocatalysts with OVs for excellent NRR performance.

    3.1.2   Nitrogen vacancies

    Oxygen vacancies as described above can activate N2 and significantly enhance interfacial electron transfer to boost NRR performance. That tempts us to turn our attention to study other defects. The DFT results of the transition metal nitride (VN, ZrN, NbN and CrN) show that these materials as electrocatalysts are more active towards NRR than towards the competing HER [121-123]. Therefore, the nitrogen vacancies (NVs) are expected to play a critical role in conversing N2 and adsorbing N2 molecules, which will in turn enhance the NRR performance.

    Lv et al. reported experimentally and theoretically that NRR performance could be improved through inducing NVs in the metal-free polymeric carbon nitride (PCN) [124]. Experimentally, NVs were generated inside PCN by pyrolysis under the argon atmosphere and the NVs concentration was regulated by adjusting the calcination condition. XRD patterns revealed a wide and weak peak, indicative of the formation of NVs (Fig. 4a). Observing from the EPR spectra, a strong signal at g = 2.0034 was detected, which was related to an unpaired electron owing to the existence of NVs (Fig. 4b). Fourier transform infrared spectroscopy and XPS (Fig. 4c) have been used to identify the exact location of the NVs in the PCN. The resulting catalyst, PCN-NV, exhibited superior electrocatalytic NRR performance with reasonable NH3 rate of 8.09 µg h−1 mgcat−1 and FE of 11.59%, nearly over 10-fold enhancement in comparison to pristine PCN (Fig. 4d). This exceptional performance revealed that the enrichment of NVs boosted the NRR electrocatalysis. Cycling tests proved that PCN-NV was stable during the electrochemical NRR process (Fig. 4e). DFT revealed that dinitrogen molecules might be chemisorbed to PCN by building a binuclear end-on bound structure for the transformation of spatial electrons, and NRR adopted an alternative associative pathway (Fig. 4f).

    Figure 4

    Figure 4.  Characterizations of PCN with and without NVs: (a) XRD patterns and (b) EPR spectra. (c) N 1s XPS spectra. (d) Ammonia yield. (e) Cycling test of PCN-NV. (f) Free energy diagram for NRR on NVs engineered PCN at equilibrium potential. Reproduced with permission [124]. Copyright 2018, John Wiley and Sons. (g) NH3 yields and FE of NV-W2N3 and W2N3 at −0.2 V vs. RHE. Reproduced with permission [125]. Copyright 2019, John Wiley and Sons.

    Besides the metal-free catalysts, Jin et al. successfully synthesized a new ultrathin 2D layered W2N3 with surface NVs for electrochemical NRR [125]. Computational studies proposed that NVs on W2N3 could produce an electron-deficient environment that would enable nitrogen adsorption and diminish the thermodynamic potential of NRR. The activity and stability of NVs were estimated by electrochemical measurements and various ex-situ characterizations. The NVs-engineered 2D W2N3 achieved a higher NRR performance with an average NH3 formation rate of 11.66 ± 0.98 µg h−1 mgcat−1 (3.80 ± 0.32 × 10−11 mol cm−2 s−1) and FE of 11.67% ± 0.93% at −0.2 V vs. RHE than that of pristine W2N3 (Fig. 4g). Nuclear magnetic resonance (NMR) test utilizing 15N2 as feeding gas indicated that NH3 was produced by NRR rather than catalyst decomposition or contamination. In addition, the synergistic effect between NVs and doped ions contributes to enhancing NRR performance as well. Li et al. proposed that Fe-doped MoN2 could act as a promising NRR catalyst through DFT calculations [123]. This is because that the NVs played a dominant role in adsorbing and activating nitrogen, while Fe-dopants mitigated the interaction between Mo-N2 and promoted the second ammonia formation.

    Based on the computational studies, transition metals (TMs) nitrides electrocatalysts have been confirmed to be more active towards NRR than HER. Therefore, we believe forming NVs in TMs nitrides as electrocatalysts can augment the number of active sites, which would in turn, enhance the NRR electrocatalytic performance. Additionally, isotopically labelled 15N2 and 14N2 are highly required to determine the detected ammonia origin, particularly for the NRR electrocatalysts which contain nitrogen species.

    3.1.3   Sulfur vacancies

    In the biological process, nitrogen fixation naturally occurs through nitrogenase enzymes, especially those containing a cofactor of iron-molybdenum-sulfur (FeMoS), which encourages the researchers to design catalysts based on the active components of nitrogenase enzymes to boost the NRR performance [126-129]. During the design of catalysts, the sulfur defects are indispensable parts in improving the NRR [126]. For example, defect-rich MoS2 nanoflowers attained high FE of 8.34% and NH3 yield of 29.28 µg h−1 mg−1 at −0.40 V vs. RHE, outperforming its defect-free counterparts. DFT calculations found that NRR on defected MoS2 would proceed at a lower energy barrier of 0.60 eV in the distal associative pathway for the reaction of *NH → *NH2, relative to 0.68 eV in defect-free MoS2. The positively charged plain Mo atoms at the edge of the defect could trigger the N2 molecules, contributing to the decline of electrons between N atoms (Fig. 5a) [128-131].

    Figure 5

    Figure 5.  (a) Free-energy profile for NRR at MoS2 edge site. Reproduced with permission [126] (b) Scheme of Ru/2H-MoS2 electronic properties and DFT for the mechanism of NRR. Reproduced with permission [129]. Copyright 2019, American Chemical Society. (c) Co-doped MoS2−x NRR performance at different potentials. (d) NRR performance of MoS2−x, Co-doped MoS2−x and CoS2 at −0.3 V. (e) Free-energy diagram of ammonia synthesis. Reproduced with permission [128]. Copyright 2019, American Chemical Society.

    The synergistic effect between sulfur vacancies (SVs) and doping of TMs brings about a new strategy for designing NRR catalysts. Suryanto et al. synthesized Ru nanoparticles on SVs-rich MoS2 [129]. Ru/MoS2 attained a high FE of 17.6% and an NH3 yield of 1.14 × 10−10 mol s−1 cm−2 at 50 ℃. Both theoretical and experimental studies suggested that SVs in the MoS2 matrix could act as the active H+ binding sites and Ru nanoparticles as the active N2 binding site (Fig. 5b). Recently, Zhang et al. synthesized a novel NRR electrocatalyst by doping Co into MoS2− nanosheets with rich SVs [128]. Co-doped MoS2− attained excellent performance, with an NH3 yield of 0.63 mmol h−1 g−1 and FE of more than 10% at −0.3 V vs. RHE (Figs. 5c and d). From both experimental and theoretical perspectives, Co doping into defective MoS2− was eligible to efficiently induce NRR activity in its basal plane (Fig. 5e). DFT calculations indicated that in the first nitrogen adsorption step N2 → *N2 of NRR reaction, Co-doped MoS2− would proceed at a lower energy barrier of 0.94 eV than that of defective MoS2− (1.19 eV). The energy barrier for the rate-limiting step *NH → *NH2 was 1.62 eV in the undoped material, while reduced to 0.52 eV in the Co-doped case. Therefore, Co-doping led to a decrease in the overpotential required for electrolytic reaction, thus greatly improving the yield of ammonia and Faraday efficiency. The above endeavors not only show the NRR electrocatalyst with good activity under environmental conditions, but also pave the way for exploring TMs with SVs as NRR catalysts.

    Heteroatom doping can be advantageous to augmenting the active sites and reducing the energy barrier of NRR by building a unique electronic structure and coordination environment. It has also been thoroughly investigated as an effective way of improving the electrocatalytic activity of catalysts [132]. The latest advances in NRR electrocatalysts based on heteroatom doping will be recapitulated in this section.

    3.2.1   Metal-atom doping

    The heteroatom doping in carbon-based materials usually causes defects, which can not only tailor the electron structure of the neighboring carbon atoms for modulating polarity but also provide active sites to allow better adsorption of reactants. It has been established as supports for anchoring single metal atoms or double metal atoms in carbon-based materials to effectively tune the electronic properties of carbon. In this way, the performance of the materials will be boosted in all aspects, such as improved catalytic efficiency, increased utilization of atoms and lower cost [133,134].

    Many research groups have constructed single atom catalysts (SACs) to reveal NRR efficiency through computational study. For example, Wei et al. performed spin-polarized DFT which proposed that the Fe-doped monolayer phosphorene (Fe@P) single metal atom catalyst (SMAC) could show superior catalytic efficiency than the bare matrix [135]. Interestingly, the doped Fe atom not only reduced the bandgap to 0.65 eV and sped up the charging-transfer process but also functioned as a catalytic active site for NRR, as clearly demonstrated by the spin-density and asymmetric charge distribution (Fig. 6a).

    Figure 6

    Figure 6.  (a) Top view of the distribution of spin density. The positive and negative densities of the spin are blue and yellow isosurfaces. Reproduced with permission [135]. Copyright 2018, Royal Society of Chemistry. (b) Differences between H adsorption free energy on the SAC (ΔGSAC(*H)) and adsorption free energy on the surface (ΔGsurface(*H)) of the same metal atoms (eV). Horizontal dashed line denotes ΔGSAC(*H) = ΔGsurface(*H). Reproduced with permission [136]. Copyright 2018, American Chemical Society. (c) ΔGmax as a function of adsorption energy of N2. Reproduced with permission [138]. Copyright 2019, Elsevier. (d) NH3 yields and FEs of the pure Ni2P catalyst as well as the Fe-doped Ni2P catalysts with various Fe/Ni ratios (including Fe0.2Ni1.8P, Fe0.25Ni1.75P, Fe0.4Ni16P and Fe1Ni1P). (e) The structures of intermediates for N2 reduction on Fe-doped Ni2P (001). The adsorption site is denoted by an asterisk (*) and the H, N, Ni, Fe and P atoms are shown in light pink, blue, gray-blue, brown and pink colors, respectively. Reproduced with permission [139]. Copyright 2020, Elsevier. (f) Band structure diagram for Fe-ReS2 and pure ReS2. Reproduced with permission [140]. Copyright 2020, John Wiley and Sons.

    Choi et al. [136] have constructed a sequence of SACs on defective graphene comprising N defects and C vacancies towards NRR, and many of them (e.g., V@N4 and Ti@N4) have served as effective catalysts with good selectivity and low overpotential. The calculation results indicated that H adsorption on SACs can be effectively suppressed by because the H adsorption energy on the SAC (ΔG(*H)) is greater than the H adsorption energy on the metal surface consisting of the corresponding metal atoms (ΔGsurface(*H)). For most metals, the values of ΔGSAC(*H) − ΔGsurface(*H) always remain in the positive region with a highly discrete trend (Fig. 6b). Similarly, Chen et al. prepared multiple single transition metal atoms (such as Ti, Pt and V) which were anchored on graphite carbon nitride (g-C3N4) containing NVs (TM@NVs-g-C3N4) [137]. Since the electrons in the TM orbits can be transferred to the 2p* antibonding orbit of N2, the hydrogenation efficiency of N2 would be increased.

    Apart from the SACs, Yang et al. investigated double-atom catalysts Fe-TM supported on graphene-based substrates (Fe-TMDA/GS) for electroreduction of nitrogen to ammonia [138]. The introduction of a second metal atom facilitied the electron transfer between N2 and graphene surface, thereby further enhancing the catalytic activity of NRR. As shown in Fig. 6c, the sequence of catalytic activity of Fe-TMDA/GS in NRR can be explained based on the inverted volcano diagram of ΔGmax and the adsorption energy N2. According to the Sabatier principle, the catalytic activity of Fe TMDA/GS (TM = Cu) was obviously higher than that of single atom Fe catalyst, indicating that the regulation of TM improved the catalytic activity of NRR.

    For practical concern, Guo et al. synthesized Fe-doped Ni2P nanosheets by hydrothermal method as electrocatalyst for NRR [139]. Among the prepared Fe-doped Ni2P nanosheets, Fe0.4Ni1.6P exhibited superior NRR activity, with FE and ammonia yield of 7.92% and 88.51 µg h−1 mgcat−1, respectively, at −0.3 V vs. RHE than pure Ni2P without Fe-doped (Fig. 6d). Theoretical calculations revealed the changes in the related intermediates in NRR and the energy barrier difference between Ni2P (001) and Fe-doped Ni2P (001), involving the transfer of six electrons and H+ ions. The energy barrier of Fe-doped Ni2P in the first potential determination step *N–N* → *NH–N* was 0.65 eV lower than that of Ni2P, implying that Fe doping promoted N2 activation and hydrogen selectivity in NRR. Due to the lower energy barrier, *NNH reacted in an associative distal pathway and thus was hydrogenated to form the reaction intermediate *NNH2 instead of *NHNH (Fig. 6e). More recently, Lai et al. reported that low-valent metal doping to tune the energy level of electrocatalysts was effective in addressing the bottlenecks for HER and NRR [140]. They doped ReS2 nanosheet with low-valent TMs (such as Fe, Co, Ni, Cu, and Zn) to form more active sites for N2/H2O chemisorption and electron transfer. Nitrogen-coated carbon nanofiber of Fe-doped ReS2 nanosheets (Fe-ReS2@N-CNF) obtained the highest electrochemical activity with NH3 yield of 80.4 µg h−1 mgcat−1, which was 8-fold higher than bare ReS2. Compared with the energy levels of undoped ReS2 nanosheets, the energy levels of Fe-ReS2 nanosheets were appropriately shifted to near the reduction potentials of H2O and N2 (Fig. 6f), simplifying the electron exchange process between the electrocatalysts and the reaction species [141]. DFT studies have shown that the introduction of low-valent transition metal into ReS2 nanosheet structure could create abundant anchoring sites for direct adsorption and subsequent activation of N2 molecules responsible for enhanced NRR performance.

    3.2.2   Non-metal doping

    Non-noble transition-metal-based catalysts have been studied in the production of lower-cost and high-efficient electrocatalysts for NRR. Due to facile synthesis, N-doping becomes the most popular in carbon-based electrocatalysts that are typically accomplished through the carbonization of N-rich substrates. As the nitrogen atom has a higher electronegativity than carbon atom, the electron cloud density of the neighboring carbon atom connected to nitrogen is lower, which is beneficial to the N2 adsorption [142,143]. Moreover, the nitrogen atom has one more extranuclear electron and less electron affinity than carbon atom, yielding higher positive charge density of the carbon atom adjacent to the nitrogen. Furthermore, there exists conjugation interaction between the lone pair of electrons in the nitrogen atom and the π bond of the carbon atom; therefore, the N-doped carbons show excellent electrochemical performance. Yang et al. prepared N-containing hierarchical carbon foams (NCFs) by combining ZnCl2 with cicada sloughs [144]. The NCFs attained a high NH3 yield rate of 15.7 µg h−1 mgcat−1 and FE of 1.45% at −0.2 V vs. RHE. N-containing polymers or compounds can also be used as precursors in order to prepare N-doped carbon to promote NRR performance. Wang et al. introduced a technique for developing a nitrogen-doped carbon membrane (NCM) with a hierarchical structure for NRR [145]. The pristine NCMs obtained an NH3 yield of 0.08 gm−2h−1 at −0.3 V vs. RHE and FE of 5.2% at −0.2 V vs. RHE. By using semiconductive NCMs with high porosity as a loading substrate for Au nanoparticles, the efficiency improved to 0.361 gm−2h−1 and 22%, respectively.

    Inspired by the improved electrocatalytic performance generated by the N-doping materials, some groups have constructed the synergistic doping species models into the carbon matrix. The incorporation of nonmetals such as boron into the carbon matrix can redistribute electron density, where the electron-deficient boron sites offer superior binding capabilities to the N2 molecules [146]. Yu et al. synthesized boron atoms incorporated into graphene sheets (BC3) by thermal reduction of graphene oxide [146]. BC3 attained ammonia yield rate and FE of 9.8 µg h−1 cm−2 and 10.8%, respectively, at −0.5 V vs. RHE. The NH3 yield and FE results were 5 times and 10 times fold higher than the undoped graphene, respectively. In fact, the pristine boron nitride (BN) replaced the C atom at the edge or defect sites of the G framework caused by the sluggish ammonia desorption. To overcome this effect, Mao et al. developed a single B-atom incorporated on BN edge (B@BN) with an incredibly low overpotential of only 0.13 V using a distal pathway. In addition, with a low energy barrier of 0.35 eV, the created NH3 could be quickly removed [147].

    Besides being a single non-metal element for doping, Bi-non-metal can be developed as an NRR catalyst because the more flexible configurations could activate the adjacent carbon atoms and put them into carbon-based matrixes. Tian et al. studied N/S co-doped graphene (NSG) as an effective metal-free catalyst for NRR [148]. The as-prepared NSG performed an NH3 yield of 7.7 µg h−1 mgcat−1 and FE of 5.8% at −0.6 V vs. RHE higher than undoped and single-doped counterparts. The existence of pyrrole-N, pyridine-N, and thiophene-S was expected to trigger positively charged on adjacent C atoms, which could facilitate N2 adsorption, weaken N≡N bond, and boost electron transfer.

    Amorphous materials possessing dangling bonds raise the number of unsaturated coordination sites as well as increase the number of catalytic defects, thereby improving their catalytic activity and promoting NRR efficiency. Amorphous catalysts have been thoroughly researched for their high performance in different electrocatalytic reactions (such as oxygen evolution reaction, HER and NRR) and most of them have shown better performances over their crystalline forms [149-152]. Throughout this section, we will outline the recent progress of the amorphous materials (noble and non-noble metals) in the NRR.

    3.3.1   Amorphous noble metal-based materials

    Noble metals are applied broadly as electrocatalysts for a variety of reactions as they can provide strong binding for a wide range of reactants and have good electric conductivity. Lately, noble metal-based catalysts, including Au [153-155], Rh [156], Pt [157] and Ru [158] have been widely utilized as NRR electrocatalysts.

    Through crystallinity modulation, precious metal-based materials can be transferred to amorphous materials to further create defects offering active sites. Li et al. [153] transformed crystallized Au nanoparticles (~5 nm) to the amorphous state and used reduced graphite oxide (rGO) as a platform for anchoring Au nanoparticles (a-Au/CeOx-rGO) by an easy co-reduction process under ambient conditions. Low-crystalline cerium oxides (CeOx) played a critical role in enhancing NRR performance because they had a high number of unsaturated coordination sites. As a consequence, a-Au/CeOx-rGO with low Au loading of 1.31 wt% achieved an ammonia yield of 8.3 µg h−1 mg−1 and FE of 10.10% at −0.2 V vs. RHE at ambient temperature, better than its bare equivalent without CeOx (c-Au/rGO, with an NH3 rate of 3.5 µg h−1 mg−1 and FE of 3.67%) (Figs. 7ac). Moreover, there was no hydrazine formed in this work, proving that the catalyst was effective in boosting both the activity and the selectivity.

    Figure 7

    Figure 7.  (a) Schematic representation for the electrochemical NRR by catalysts of c-Au/rGO and a-Au/CeOx-rGO. (b, c) NRR performance of a-Au/CeOx-rGO and other catalysts at −0.2 V vs. RHE, Reproduced with permission [153]. Copyright 2017, John Wiley and Sons. (d) Representative XRD patterns of Cu/rGO, Pd/rGO and Pd0.2Cu0.8/rGO. (e, f) NRR performance of Pd0.2Cu0.8/rGO and different catalysts at −0.2 V vs. RHE. Reproduced with permission [159]. Copyright 2018, John Wiley and Sons. (g) NRR performance of BVC-A at −0.2 V vs. RHE. (h) Illustration of fabricating BVC-A and BVC-crystalline (BVC-C) NRR electrocatalysts. Reproduced with permission [112]. Copyright 2018, John Wiley and Sons.

    As HER occupies a dominant position on the noble metal catalysts, the NRR activity cannot be fully substantiated. Generally, the expensiveness and paucity of noble metals are considered as obstacles for practical large-scale applications. To address these thorny issues, alloying the noble metals and TMs is proposed as a solution to generate various active sites and enhance the catalytic performance. As shown by Shi et al., the ultrafine PdxCu1− amorphous nanoclusters anchored on rGO (Pd0.2Cu0.8/rGO) were prepared using a co-reduction method [159]. The X-ray diffraction (XRD) pattern of Pd0.2Cu0.8/rGO has no obvious bimetallic peaks compare with Pd/rGO and Cu/rGO except for the diffraction peaks of reduced graphene oxide, which indicates its amorphous state (Fig. 7d). By tuning the bimetallic ratio of Cu and Pd, the optimal performance was attained for Pd0.2Cu0.8/rGO composite with an NH3 yield of 2.80 µg h−1 mg−1 at −0.2 V vs. RHE (Fig. 7e). As the electronic interaction for both Cu and Pd considerably altered the electronic states of the monometal compounds, the NH3 yield provided by the bimetallic Pd0.2Cu0.8/rGO was 2.1 times higher than that of the Pd/rGO and 2.4 times higher than that of the Cu/rGO equivalents (Fig. 7f). The outstanding catalytic performance of alloy catalysts with low noble metal proportions would highly inspire the researchers to deepen study about amorphous bimetallic or even trimetallic catalysts on NRR process in the future.

    3.3.2   Amorphous non-noble metal-based materials

    Lately, non-noble metals (particularly TMs) have emerged as potential candidate electrocatalysts for NRR owing to their abundant resources, low cost, and favorable catalytic activity [160]. Considering the fact that works in this field are few, researchers should devote more effort to unraveling the effect of different amorphous non-noble metals on catalytic efficiency of NRR. As we have already concluded, existing OVs in TMOs could play an active role in the adsorption of reactants as well as reaction intermediates, thus decreasing the energy activation barrier and boosting electrocatalytic efficiency. Lv et al. prepared Bi4V2O11/CeO2 composite with amorphous phase (BVC-A) consisting of abundant OVs, which gave superior electrocatalytic NRR efficiency with an ammonia production of 23.21 µg h−1 mg−1 and FE of 10.16% at −0.2 V vs. RHE under ambient conditions (Fig. 7g) [112]. It was also demonstrated that the mole ratios of Ce/Bi were crucial for forming the amorphous Bi4V2O11 which possessed rich defects, as illustrated by XPS and high resolution transmission electron microscope (HRTEM) results (Fig. 7h). The disturbed structure was abundant with dangling bonds, which provided a high concentration of defect sites and decreased the energy barrier, serving as a good catalyst for enhancing NRR activity.

    In addition to BVC-A, a sequence of amorphous Sn/crystalline SnS2 (Sn/SnS2) nanosheets were synthesized by Li et al. through the hydrothermal method, followed by in situ electrochemical reduction [161]. The different percentages of amorphous Sn could be tuned through selecting electrolytes with various pH values. In N2-saturated 0.1 mol/L phosphate buffer solution (PBS), Sn/SnS2 electrocatalyst attained a high ammonia yield of 23.8 µg h−1 mg−1 at −0.8 V vs. RHE and FE of 6.5% at −0.7 V vs. RHE. No observable hydrazine was produced, implying its favorable selectivity for NH3. Moreover, NMR proved that the cathode reduction products were derived from the reduced N2. These findings will provide guidance for us to investigate other NRR catalysts based on TMs via using the amorphous approach in the future.

    The surface area plays a key role in various reactions, as the reactants are adsorbed and bound to the catalyst surface. The electrical fields can be concentrated by the high-curvature structure which will influence ion concentrations [162]. Therefore, creation of edges or corners may be an efficient way to improve the performance of NRR. Song et al. [163] have created a specific carbon nanospikes (CNS) with scratchy surfaces made up of sharp spikes. The nitrogen-doped CNS contained multiple-oriented spikes of approximately 50–80 nm in length, ending at approximately ~1 nm in the tip, serving as a good physical catalyst to stimulate the electrical field. The pure CNS demonstrated higher NRR efficiency with an outstanding NH3 yield of 97.18 µg h−1 cm−2 and a good FE of 11.5% at −1.19 V vs. RHE than oxygen-plasma etched (O-Etched) CNS without a sharp tip.

    The electrolyte cations with the sequence of Li+ > Na+ > K+ have been determined to be critical during NRR process, implying that the counterions with the smallest ionic radius could enhance the electric field and serve as promoters to enhance the N2 concentration close to the tips (Figs. 8a and b), in addition to creating a dehydrated cation shield around the tip to block the water and the proton to pass through. After Ar plasma treatment, the efficiency of NRR became very low, showing the beneficial impact of the spike structure (Figs. 8c and d). These outcomes showed that the NRR was driven by the powerful electric fields from the catalyst spike.

    Figure 8

    Figure 8.  (a, b) Ammonia output and FE of CNS in different countercations with the order of Li+ > Na+ > K+. (c, d) FEs and ammonia formation rates of various catalysts. Reproduced with permission [163]. Copyright 2018, American Association for the Advancement of Science. (e) Schematic diagram of NRR on triangular AgTPs and rounded AgNPs. (f) NRR performance of AgTPs and AgNPs. Reproduced with permission [164]. Copyright 2019, The Royal Society of Chemistry.

    More recently, Gao et al. reported that NRR efficiency could be improved by using Ag triangular nanoplates (AgTPs) with sharp edges as catalysts [164]. The rounded Ag nanoparticles (AgNPs) were mainly wrapped by {111} and {100} facets, which served as active sites for H+/e transfer and nitrogen fixation. While for AgTPs, a large number of atoms were located at the sharp edges and corners with lower coordination numbers and stronger ability to adsorb nitrogen species (Fig. 8e). Therefore, the energy barrier of the reaction determination step *NN → *NNH was highly reduced, promising a good yield and high efficiency at a lower applied potential for NRR. AgTPs with sharp edges have attained an ammonia yield of 58.5 mg gAg−1 h−1 and a high FE of 25% at an overpotential of −0.25 V vs. RHE under environmental conditions. In contrast, by using rounded Ag nanoparticles, the FE and NH3 yield have been decreased significantly to 16% and 38 mg gAg−1 h−1 respectively at a larger overpotential −0.35 V vs. RHE (Fig. 8f). Such enhanced performance verified that the AgTPs was one of the most efficient transition metal catalysts for NRR.

    Porousness existed in the structures usually possess rich channels, which could facilitate the N2 enrichment, prolong N2 residue time, slow down the escape of reaction intermediates, shorten the transport path of electrons and protons, and further accelerate NRR process. For example, Ying et al. transformed commercial Fe3O4 into porous structure based on deep eutectic solvent regeneration strategy [165]. The flaky and porous structure endowed the treated Fe3O4 with large active surface area and high porosity. Compared with commercial Fe3O4, the treated material showed better electrocatalytic performance, with a high yield of 12.09 µg h−1 mgcat−1 and a FE of 34.38% at 0.1 mol/L Na2SO4. Besides, alloying two metals can change the electronic structure of each metal, and then synergistically change the adsorption energy of reactants on the catalyst [166]. For example, Yu et al. proposed a one-pot reduction method for direct synthesis of bimetallic Ag3Cu porous network (Ag3Cu BPNs) [167]. Due to the porous network and bimetallic composition, Ag3Cu BPNs exhibited a high ammonia yield of 24.59 µg h−1 mgcat−1 and a superior FE of 13.28% in 0.1 mol/L NaSO4. Similarly, Pang et al. designed a 3D layered interconnected porous network Pd/Au alloy with excellent NRR performance (Fig. 9a) [168]. The large specific surface area and high active site density effectively inhibited HER and contributed to an ammonia yield up to 39.9 µg h−1 mgcat−1. In terms of stability, the alloy maintained 100% of the original activity after 18 h of operation.

    Figure 9

    Figure 9.  (a) SEM image of Pd3Cu1 alloy. Reproduced with permission [168]. Copyright 2019, Elsevier. (b) NH3 yield (bar graph) and Faradaic efficiency (blue point) by using Vr-ReSe2@CBC NFs and Vr-ReSe2 bulk as contrast samples. (c) DOS for V-ReSe2 and ReSe2 models, as well as the charge density difference of V-ReSe2 (The pink and yellow areas present electrons accumulation and depletion, respectively) (left figure), and free energy diagrams for nitrogen reduction reaction on the surface of V-ReSe2 and ReSe2 catalysts (right figure). (d) Schematic diagram of synthesis process of CBC@Vr-ReSe2@CBC. (e) NRR performance (NH3 yield and Faradaic efficiency) of CBC@Vr-ReSe2@CBC NFs in 0.1 mol/L Na2SO4 electrolyte under ambient conditions. (f) Distributions of concentration of H+ ions in the electrode in the COMSOL multiphysics simulations. Reproduced with permission [169]. Copyright 2020, John Wiley and Sons.

    However, the enhanced N2 activation due to porous active site modifications such as vacancy formation and doping, inevitably leads to a reduction in the proton coupling potential and thus enhances the occurrence of the competitive hydrogen precipitation reaction. The trap engineering strategies discussed in the previous sections do not effectively address this issue. Porous materials provide an alternative solution, as they may involve processes such as interfacial wettability and pore-induced diffusion that can facilitate enhanced nitrogen fixation while reducing the hydrogen precipitation performance of the catalyst. Recently, some progress has been made in enhancing electrocatalytic nitrogen fixation through modification of interfacial wettability. Lai et al. prepared vacancy-rich ReSe2@carbonized bacterial cellulose nanofibers (Vr-ReSe2@CBC NFs) by heat treatment methods [169]. Monolayer porous CBC as a porous carrier enabled uniform distribution of Vr-ReSe2 nanosheets compared to unloaded selenium-rich vacant ReSe2, thus enhancing their charge transfer capacity and producing higher ammonia yields (Fig. 9b). The total density of states (DOS) plots and free energy compared to ReSe2 bulk are shown in Fig. 9c, with d-band centering at 4.19 eV and 4.42 eV for V-ReSe2 and ReSe2, respectively. This indicated that more electrons were distributed around the Fermi energy level, causing stronger interaction between N2/water and the V-ReSe2 surface, and allowing the electrons to be easily transferred to N2/water. In such way, the adsorption and activation of V-ReSe2 to both N2 and water were enhanced. To confine HER, the same group buried Vr-ReSe2@CBC nanofibers between two CBC layers (CBC/Vr-ReSe2@CBC/CBC) (Fig. 9d). The CBC layers can achieve a hydrophobic interior of Vr-ReSe2@CBC using trace evaporated water, which hindered the HER process and reduced the FE. Therefore, a large amount of moist N2 gas was injected into the Vr-ReSe2@CBC NFs and aggregated on its surface under thermodynamically favorable adsorption to facilitate the NRR process. In terms of electrochemical properties, the CBC/Vr-ReSe2@CBC/CBC exhibited a FE four times higher than that of Vr-ReSe2@CBC (Fig. 9e). The H+ concentration near the Vr-ReSe2@CBC interface was significantly reduced using COMSOL multiphysics simulations (Fig. 9f), indicating that the CBC layer covering the top of Vr-ReSe2@CBC was hydrophobic and porous to water molecules.

    Generally, defect engineering, including generating cations/anions vacancies, doping defects, creating amorphous structures, edge defects, and porous defects has been demonstrated to be effective in enhancing both activity and selectivity of the NRR (summarized in Table 1) [26,84,94,112115,117120,125128,139,140,146,155,159,161,163,164,167,170195].

    Table 1

    Table 1.  A brief summary of typical NRR electrocatalysts with defects identified in the latest years and their catalytic efficiency.
    DownLoad: CSV

    In this review, recent research progress on the effect of several different defect structures in defect engineering technologies on electrochemical NRR is comprehensively investigated. For vacancy (e.g., OVs, SVs and NVs) defect structures, the alteration of the local electrons allows the vacancy to act as the active center of NRR, which can effectively adsorb and activate N2 molecules. DFT theoretical calculations show that the presence of vacancies largely reduces the energy barrier required for the N2 hydrogenation reaction compared to catalysts without vacancy structures. For doped defect structures, theoretical calculations are discussed to summarize the feasibility of atom doping of different metal species as active sites to tune the electron configuration and thus improve the nitrogen fixation capacity. Meanwhile, non-metals are also an important means of doping, such as nitrogen doping, where the smaller affinity of nitrogen progeny and the existence of conjugation between its lone pair of electrons and π-bonds of carbon atoms promise the excellent properties. Compared to crystalline structures, amorphous catalysts can help expose more defective structures with active sites, and both volume- and surface-constrained electrocatalytic processes may exist in amorphous catalysts, thus ensuring a larger electrolyte-catalyst interface. The presence of edge defect structures enriches the dangling bonds on the catalyst surface, which provides a high concentration of defect sites, lowers the energy barrier and serves as a good catalyst for enhanced NRR activity. Some recent studies have shown that the selectivity of ammonia and hydrogen corresponding to different kinds of exposed high-energy atoms on the edge differs, so subsequent studies on the cross section may involve the separation of active sites to achieve simultaneous enhancement of selectivity and activity for ammonia. Porousness as another popular defective structure has abundant specific surface area and provides a large number of active site structures, which facilitates catalyst conductivity enhancement and nitrogen activation. Hydrogen evolution reaction occurs on proton-rich hydrophilic surfaces, and modification of interfacial properties by porous hydrophobic materials can selectively attenuate proton coupling and thus enhance the selectivity for ammonia. Besides, the types of defects are often not present in a single structural form; for example, there may be simultaneous abundant edges/corners around the porous structures. Amorphous and doped structures may be accompanied by the formation of vacancies (e.g., OVs and SVs). These defect types interact and interconnect with each other while exposing abundant active sites, which play an indispensable role in improving NRR performance. Despite of some pioneering and creative research on defect engineering of electrocatalytic materials, this area of research is still in its preliminary stage and there are substantial challenges ahead that need to be properly addressed.

    (1) The controlled generation of defects. The current methods of preparing defects either in situ or post-processing are not effective in achieving a single presence of defects, which impedes the study of their mechanism during electrocatalytic reactions and the understanding of the important role they play. This calls for precise synthetic methods to control the generation of defects, which will help to isolate the complex factors in the electrochemical process and pave the way for a comprehensive and profound understanding of defect effects.

    (2) Large-scale synthesis of defective electrocatalysts. From the perspective of materials synthesis, the key properties such as morphology, electronic and energy band structure, and charge transfer capability of defective catalytic materials can be optimally designed by controlled synthetic pathways to achieve efficient nitrogen fixation. However, since defect formation in catalysts is a high-energy process and is currently prepared only in the laboratory, the cost issue of defect fabrication deserves researchers' attention. Therefore, an effective strategy for large-scale synthesis of defective electrocatalysts is urgently needed to meet the economic demand for practical applications in electrocatalytic nitrogen fixation.

    (3) Defect characterization. In general, conversion and reorganization of active sites may occur during reduction reactions, and traditional in situ characterization methods have difficulty in reproducing such structural changes. Therefore, in situ/operational measurements with high spatial, spectral and temporal resolution have become an effective technique for gaining insight into the role of defects in electrocatalytic reactions. This study responds to a real catalytic mechanism that can guide researchers to effectively design and protect the active sites of catalysts and extend the catalyst lifetime.

    In summary, the progress of defect engineering for guiding the design of NRR electrocatalysts were discussed. Although future research work in this area will be more challenging, we believe that the combination of accurate characterization techniques, controlled synthetic methods and scientifically rigorous studies, together with reasonable model simulations, will further accelerate the development of defect structure systems driven NRR electrocatalysts. We are convinced that the field of electrocatalytic nitrogen fixation will eventually achieve greater breakthroughs.

    The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

    This work was supported by the National Natural Science Foundation of China (Nos. 22075133 and 21701086).


    1. [1]

      B.H.R. Suryanto, K. Matuszek, J. Choi, et al., Science 372 (2021) 1187–1191. doi: 10.1126/science.abg2371

    2. [2]

      K. Li, S.Z. Andersen, M.J. Statt, et al., Science 374 (2021) 1593–1597. doi: 10.1126/science.abl4300

    3. [3]

      X. Kong, H.Q. Peng, S. Bu, et al., J. Mater. Chem. A 8 (2020) 7457–7473. doi: 10.1039/D0TA01453B

    4. [4]

      X. Zhao, G. Hu, G.F. Chen, et al., Adv. Mater. 33 (2021) 2007650. doi: 10.1002/adma.202007650

    5. [5]

      M. Wang, M.A. Khan, I. Mohsin, et al., Energy Environ. Sci. 14 (2021) 2535–2548. doi: 10.1039/D0EE03808C

    6. [6]

      Y. Wan, H. Zhou, M. Zheng, et al., Adv. Funct. Mater. 31 (2021) 2100300. doi: 10.1002/adfm.202100300

    7. [7]

      Y. Tan, L. Yan, C. Huang, et al., Small 17 (2021) 2100372. doi: 10.1002/smll.202100372

    8. [8]

      K.A. Brown, D.F. Harris, M.B. Wilker, et al., Science 352 (2016) 448. doi: 10.1126/science.aaf2091

    9. [9]

      S. Shang, W. Xiong, C. Yang, et al., ACS Nano 15 (2021) 9670–9678. doi: 10.1021/acsnano.0c10947

    10. [10]

      L. Li, C. Tang, X. Cui, et al., Angew. Chem. Int. Ed. 60 (2021) 14131–14137. doi: 10.1002/anie.202104394

    11. [11]

      Y. Feng, Z. Zhang, K. Zhao, et al., J. Colloid Interface Sci. 583 (2021) 499–509. doi: 10.1016/j.jcis.2020.09.089

    12. [12]

      F. Gorky, J.M. Lucero, J.M. Crawford, et al., ACS Appl. Mater. Interfaces 13 (2021) 21338–21348. doi: 10.1021/acsami.1c03115

    13. [13]

      A. Anastasopoulou, R. Keijzer, S. Butala, et al., J. Phys. D: Appl. Phys. 53 (2020) 234001. doi: 10.1088/1361-6463/ab71a8

    14. [14]

      K. Han, J. Luo, Y. Feng, et al., Energy Environ. Sci. 13 (2020) 2450–2458. doi: 10.1039/D0EE01102A

    15. [15]

      C.J.M. van der Ham, M.T.M. Koper, D.G.H. Hetterscheid, Chem. Soc. Rev. 43 (2014) 5183–5191. doi: 10.1039/C4CS00085D

    16. [16]

      X. Cui, C. Tang, Q. Zhang, Adv. Energy Mater. 8 (2018) 1800369. doi: 10.1002/aenm.201800369

    17. [17]

      G. Qing, R. Ghazfar, S.T. Jackowski, et al., Chem. Rev. 120 (2020) 5437–5516. doi: 10.1021/acs.chemrev.9b00659

    18. [18]

      F. Wang, X. Xu, J. Mao, Diam. Relat. Mater. 109 (2020) 108037. doi: 10.1016/j.diamond.2020.108037

    19. [19]

      S. Giddey, S.P.S. Badwal, A. Kulkarni, Int. J. Hydrog. Energy 38 (2013) 14576–14594. doi: 10.1016/j.ijhydene.2013.09.054

    20. [20]

      M. Jin, X. Zhang, M. Han, et al., J. Mater. Chem. A 8 (2020) 5936–5942. doi: 10.1039/C9TA13135C

    21. [21]

      Q. Wang, Y. Lei, D. Wang, et al., Energy Environ. Sci. 12 (2019) 1730–1750. doi: 10.1039/C8EE03781G

    22. [22]

      W. Tong, B. Huang, P. Wang, et al., Natl. Sci. Rev. 8 (2021) nwaa088. doi: 10.1093/nsr/nwaa088

    23. [23]

      L. Wang, M. Wu, X. Lang, et al., ChemCatChem 12 (2020) 3937–3945. doi: 10.1002/cctc.202000185

    24. [24]

      M. Li, H. Huang, J. Low, et al., Small Methods 3 (2019) 1800388. doi: 10.1002/smtd.201800388

    25. [25]

      B. Liu, B. He, H.Q. Peng, et al., Adv. Sci. 5 (2018) 1800406. doi: 10.1002/advs.201800406

    26. [26]

      H. Wang, H. Yu, Z. Wang, et al., Small 15 (2019) 1804769. doi: 10.1002/smll.201804769

    27. [27]

      A. Liu, Y. Yang, D. Kong, et al., Appl. Surf. Sci. 536 (2021) 147851. doi: 10.1016/j.apsusc.2020.147851

    28. [28]

      B. Fan, H. Wang, H. Zhang, et al., Adv. Funct. Mater. 32 (2022) 2110783. doi: 10.1002/adfm.202110783

    29. [29]

      C. Xie, D. Yan, W. Chen, et al., Mater. Today 31 (2019) 47–68. doi: 10.1016/j.mattod.2019.05.021

    30. [30]

      Y. Wang, M.M. Shi, D. Bao, et al., Angew. Chem. Int. Ed. 58 (2019) 9464–9469. doi: 10.1002/anie.201903969

    31. [31]

      V. Sorkin, Y. Cai, Z. Ong, et al., Crit. Rev. Solid State Mater. Sci. 42 (2017) 1–82. doi: 10.1080/10408436.2016.1182469

    32. [32]

      S. Qiao, Q. Wang, D. Lei, et al., J. Mater. Chem. A 10 (2022) 11702–11711. doi: 10.1039/D2TA01107G

    33. [33]

      Z. Zhang, X. Liu, D. Wang, et al., Chem. Eng. J. 446 (2022) 137037. doi: 10.1016/j.cej.2022.137037

    34. [34]

      Y. Zhang, J. Lu, L. Zhang, et al., Appl. Catal. B 309 (2022) 121249. doi: 10.1016/j.apcatb.2022.121249

    35. [35]

      X. Zhang, C. Wang, Y. Guo, et al., J. Mater. Chem. A 10 (2022) 6448–6453. doi: 10.1039/D2TA00661H

    36. [36]

      M. Li, X. Wu, K. Liu, et al., J. Energy Chem. 69 (2022) 506–515. doi: 10.1016/j.jechem.2022.01.031

    37. [37]

      Y. Zhu, Y. Sun, J. Khan, et al., Chem. Eng. J. 443 (2022) 136501. doi: 10.1016/j.cej.2022.136501

    38. [38]

      L. Wu, P. Guo, X. Wang, et al., Chem. Eng. J. 446 (2022) 136759. doi: 10.1016/j.cej.2022.136759

    39. [39]

      T. Tian, X. Jin, N. Guo, et al., Appl. Catal. B 308 (2022) 121227. doi: 10.1016/j.apcatb.2022.121227

    40. [40]

      Y. Fang, Y. Xue, L. Hui, et al., Adv. Sci. 9 (2022) 2102721. doi: 10.1002/advs.202102721

    41. [41]

      M. Yarali, H. Brahmi, Z. Yan, et al., ACS Appl. Mater. Interfaces 10 (2018) 4921–4928. doi: 10.1021/acsami.7b14310

    42. [42]

      Y. Kou, K. Wang, M. Wumaer, et al., Appl. Surf. Sci. 589 (2022) 153031. doi: 10.1016/j.apsusc.2022.153031

    43. [43]

      I.E. Khalil, C. Xue, W. Liu, et al., Adv. Funct. Mater. 31 (2021) 2010052. doi: 10.1002/adfm.202010052

    44. [44]

      X. Guo, W. Ye, Z.A. Chen, et al., Appl. Catal. B 310 (2022) 121334. doi: 10.1016/j.apcatb.2022.121334

    45. [45]

      Z. Chen, J. Cao, X. Wu, et al., ACS Appl. Mater. Interfaces 14 (2022) 12223–12233. doi: 10.1021/acsami.1c23718

    46. [46]

      H.J. Liu, C.Y. Chiang, Y.S. Wu, et al., ACS Catal. 12 (2022) 6132–6142. doi: 10.1021/acscatal.1c05539

    47. [47]

      C. Yan, W. Luo, H. Yuan, et al., Appl. Catal. B 308 (2022) 121191. doi: 10.1016/j.apcatb.2022.121191

    48. [48]

      Y. Zou, J. Hu, B. Li, et al., Appl. Catal. B 312 (2022) 121408. doi: 10.1016/j.apcatb.2022.121408

    49. [49]

      K. Li, W. Cai, Z. Zhang, et al., Chem. Eng. J. 435 (2022) 135017. doi: 10.1016/j.cej.2022.135017

    50. [50]

      T. Ma, H. Cao, S. Li, et al., Adv. Mater. 33 (2022) 2206368.

    51. [51]

      X. Wang, L. Yu, B.Y. Guan, et al., Adv. Mater. 30 (2018) 1801211. doi: 10.1002/adma.201801211

    52. [52]

      L. Xu, Y. Zou, Z. Xiao, et al., J. Energy Chem. 35 (2019) 24–29. doi: 10.1016/j.jechem.2018.09.013

    53. [53]

      R. Grantab, V.B. Shenoy, R.S. Ruoff, Science 330 (2010) 946–948. doi: 10.1126/science.1196893

    54. [54]

      H. Liu, X.M. Wang, H. Liang, et al., Int. J. Solids Struct. 191-192 (2020) 464–472.

    55. [55]

      Y. Kang, H. Du, B. Jiang, et al., J. Mater. Chem. A 10 (2022) 6560–6568. doi: 10.1039/D1TA10742A

    56. [56]

      Z. Wu, T. Wang, J.J. Zou, et al., ACS Catal. 12 (2022) 5911–5920. doi: 10.1021/acscatal.2c01829

    57. [57]

      R. Shi, Y. Zhao, G.I.N. Waterhouse, et al., ACS Catal. 9 (2019) 9739–9750. doi: 10.1021/acscatal.9b03246

    58. [58]

      B. Garlyyev, J. Fichtner, O. Piqué, et al., Chem. Sci. 10 (2019) 8060–8075. doi: 10.1039/C9SC02654A

    59. [59]

      H. Huang, L. Chen, C. Liu, et al., J. Mater. Chem. A 4 (2016) 14577–14585. doi: 10.1039/C6TA06174E

    60. [60]

      B. Jiang, T. Yang, T. Wang, et al., Chem. Eng. J. 442 (2022) 136119. doi: 10.1016/j.cej.2022.136119

    61. [61]

      C. Xue, X. Zhou, X. Li, et al., Adv. Sci. 9 (2022) 2104183. doi: 10.1002/advs.202104183

    62. [62]

      G. Chen, P. Liu, Z. Liao, et al., Adv. Mater. 32 (2020) 1907399. doi: 10.1002/adma.201907399

    63. [63]

      X. Wang, Y. Zhang, H. Si, et al., J. Am. Chem. Soc. 142 (2020) 4298–4308. doi: 10.1021/jacs.9b12113

    64. [64]

      X. She, X. Zhu, J. Yang, et al., Nano Energy 84 (2021) 105869. doi: 10.1016/j.nanoen.2021.105869

    65. [65]

      J. Ge, P. Yin, Y. Chen, et al., Adv. Mater. 33 (2021) 2006711. doi: 10.1002/adma.202006711

    66. [66]

      G. Li, K. He, F. Zhang, et al., Appl. Catal. B 309 (2022) 121231. doi: 10.1016/j.apcatb.2022.121231

    67. [67]

      X. Wan, W. Guo, X. Dong, et al., Green Chem. 24 (2022) 1090–1095. doi: 10.1039/D1GC04483D

    68. [68]

      M. Jiang, A. Tao, Y. Hu, et al., ACS Appl. Mater. Interfaces 14 (2022) 17470–17478. doi: 10.1021/acsami.2c02048

    69. [69]

      Y. Liu, H. Cheng, M. Lyu, et al., J. Am. Chem. Soc. 136 (2014) 15670–15675. doi: 10.1021/ja5085157

    70. [70]

      P. Li, Z. Jin, Z. Fang, et al., Energy Environ. Sci. 14 (2021) 3522–3531. doi: 10.1039/D1EE00545F

    71. [71]

      W. Li, D. Wang, Y. Zhang, et al., Adv. Mater. 32 (2020) 1907879. doi: 10.1002/adma.201907879

    72. [72]

      D. Wang, Y. Zou, L. Tao, et al., Chin. Chem. Lett. 30 (2019) 826–838. doi: 10.1016/j.cclet.2019.03.051

    73. [73]

      L. Ma, C. Wang, B.Y. Xia, et al., Angew. Chem. Int. Ed. 54 (2015) 5666–5671. doi: 10.1002/anie.201500947

    74. [74]

      H. Rong, J. Mao, P. Xin, et al., Adv. Mater. 28 (2016) 2540–2546. doi: 10.1002/adma.201504831

    75. [75]

      X. Wang, L. Figueroa-Cosme, X. Yang, et al., Nano Lett. 16 (2016) 1467–1471. doi: 10.1021/acs.nanolett.5b05140

    76. [76]

      J. Zhu, Y. Huang, W. Mei, et al., Angew. Chem. Int. Ed. 58 (2019) 3859–3864. doi: 10.1002/anie.201813805

    77. [77]

      Á. Morales-García, F. Calle-Vallejo, F. Illas, ACS Catal. 10 (2020) 13487–13503. doi: 10.1021/acscatal.0c03106

    78. [78]

      X. Wu, S. Xiao, Y. Long, et al., Small 18 (2022) 2105831. doi: 10.1002/smll.202105831

    79. [79]

      R. Meshkian, M. Dahlqvist, J. Lu, et al., Adv. Mater. 30 (2018) 1706409. doi: 10.1002/adma.201706409

    80. [80]

      M. Ma, X. Han, H. Li, et al., Appl. Catal. B 265 (2020) 118568. doi: 10.1016/j.apcatb.2019.118568

    81. [81]

      S. Dou, L. Tao, J. Huo, et al., Energy Environ. Sci. 9 (2016) 1320–1326. doi: 10.1039/C6EE00054A

    82. [82]

      B. Ma, S. Ren, P. Wang, et al., Nano Res. 12 (2019) 137–142. doi: 10.1007/s12274-018-2192-8

    83. [83]

      Y. Tian, Z. Wei, X. Wang, et al., Int. J. Hydrog. Energy 42 (2017) 4184–4192. doi: 10.1016/j.ijhydene.2016.09.142

    84. [84]

      L. Zhang, X.Y. Xie, H. Wang, et al., Chem. Commun. 55 (2019) 4627–4630. doi: 10.1039/C9CC00936A

    85. [85]

      F. Guo, P. Yang, Z. Pan, et al., Angew. Chem. Int. Ed. 56 (2017) 8231–8235. doi: 10.1002/anie.201703789

    86. [86]

      B. Zhang, Y. Zheng, T. Ma, et al., Adv. Mater. 33 (2021) 2006042. doi: 10.1002/adma.202006042

    87. [87]

      Y.N. Gong, L. Jiao, Y. Qian, et al., Angew. Chem. Int. Ed. 59 (2020) 2705–2709. doi: 10.1002/anie.201914977

    88. [88]

      S. Chen, H. Jang, J. Wang, et al., J. Mater. Chem. A 8 (2020) 2099–2104. doi: 10.1039/C9TA10524G

    89. [89]

      R. Anwar, N. Iqbal, S. Hanif, et al., Catalysts 10 (2020) 799. doi: 10.3390/catal10070799

    90. [90]

      C. Young, J. Wang, J. Kim, et al., Chem. Mater. 30 (2018) 3379–3386. doi: 10.1021/acs.chemmater.8b00836

    91. [91]

      N. Bhuvanendran, S. Ravichandran, S. Kandasamy, et al., Int. J. Hydrog. Energy 46 (2021) 2128–2142. doi: 10.1016/j.ijhydene.2020.10.115

    92. [92]

      J. Wang, X. Xiao, Y. Liu, et al., J. Mater. Chem. A 7 (2019) 17675–17702. doi: 10.1039/C9TA04804A

    93. [93]

      C. Chen, D. Yan, Y. Wang, et al., Small 15 (2019) 1805029. doi: 10.1002/smll.201805029

    94. [94]

      Y. Wang, Y. Tian, J. Zhang, et al., ACS Appl. Energy Mater. 3 (2020) 9516–9522. doi: 10.1021/acsaem.0c01834

    95. [95]

      J. Mao, Y. Chen, J. Pei, et al., Chem. Commun. 52 (2016) 5985–5988. doi: 10.1039/C6CC02264B

    96. [96]

      X. Yang, Y. Ma, Y. Liu, et al., ACS Appl. Mater. Interfaces 13 (2021) 19864–19872. doi: 10.1021/acsami.0c22623

    97. [97]

      X. Wang, Z. Feng, B. Xiao, et al., Green Chem. 22 (2020) 6157–6169. doi: 10.1039/D0GC01149E

    98. [98]

      H. Liao, X. Guo, Y. Hou, et al., Small 16 (2020) 1905223. doi: 10.1002/smll.201905223

    99. [99]

      J. Han, Z. Liu, Y. Ma, et al., Nano Energy 52 (2018) 264–270. doi: 10.1016/j.nanoen.2018.07.045

    100. [100]

      Y. Liu, C. Xiao, Z. Li, et al., Adv. Energy Mater. 6 (2016) 1600436. doi: 10.1002/aenm.201600436

    101. [101]

      X. Wang, X. Gan, T. Hu, et al., Adv. Mater. 29 (2017) 1603617. doi: 10.1002/adma.201603617

    102. [102]

      Z. Cai, Y. Bi, E. Hu, et al., Adv. Energy Mater. 8 (2018) 1701694. doi: 10.1002/aenm.201701694

    103. [103]

      Y. Zhao, Y. Zhao, R. Shi, et al., Adv. Mater. 31 (2019) 1806482. doi: 10.1002/adma.201806482

    104. [104]

      Y. Zhang, C. Zhang, Y. Guo, et al., J. Mater. Chem. A 7 (2019) 2536–2540. doi: 10.1039/C8TA11407B

    105. [105]

      S. Dou, X. Wang, S. Wang, Small Methods 3 (2019) 1800211. doi: 10.1002/smtd.201800211

    106. [106]

      B.K. Sharma, A. Stoesser, S.K. Mondal, et al., ACS Appl. Mater. Interfaces 10 (2018) 22408–22418. doi: 10.1021/acsami.8b04892

    107. [107]

      S. Gao, Z. Sun, W. Liu, et al., Nat. Commun. 8 (2017) 14503. doi: 10.1038/ncomms14503

    108. [108]

      X. Yao, J. Liu, W. Wang, et al., J. Chem. Phys. 146 (2017) 224703. doi: 10.1063/1.4985157

    109. [109]

      J. Kim, X. Yin, K.C. Tsao, et al., J. Am. Chem. Soc. 136 (2014) 14646–14649. doi: 10.1021/ja506254g

    110. [110]

      S. Choi, Y. Park, H. Yang, et al., CrystEngComm 22 (2020) 1500–1513. doi: 10.1039/C9CE01883B

    111. [111]

      T. Ling, D.Y. Yan, Y. Jiao, et al., Nat. Commun. 7 (2016) 12876. doi: 10.1038/ncomms12876

    112. [112]

      C. Lv, C. Yan, G. Chen, et al., Angew. Chem. Int. Ed. 57 (2018) 6073–6076. doi: 10.1002/anie.201801538

    113. [113]

      W. Fu, P. Zhuang, M.O.L. Chee, et al., ACS Sustain. Chem. Eng. 7 (2019) 9622–9628. doi: 10.1021/acssuschemeng.9b01178

    114. [114]

      Y. Fang, Z. Liu, J. Han, et al., Adv. Energy Mater. 9 (2019) 1803406. doi: 10.1002/aenm.201803406

    115. [115]

      J. Zhang, Y. Ji, P. Wang, et al., Adv. Funct. Mater. 30 (2020) 1906579. doi: 10.1002/adfm.201906579

    116. [116]

      Y. Liu, X. Kong, X. Guo, et al., ACS Catal. 10 (2020) 1077–1085. doi: 10.1021/acscatal.9b03864

    117. [117]

      C. Fang, T. Bi, X. Xu, et al., Adv. Mater. Interfaces 6 (2019) 1901034. doi: 10.1002/admi.201901034

    118. [118]

      C. Zhang, S. Liu, T. Chen, et al., Chem. Commun. 55 (2019) 7370–7373. doi: 10.1039/C9CC03221E

    119. [119]

      S. Cheng, Y.J. Gao, Y.L. Yan, et al., J. Energy Chem. 39 (2019) 144–151. doi: 10.1016/j.jechem.2019.01.020

    120. [120]

      Z. Han, C. Choi, S. Hong, et al., Appl. Catal. B 257 (2019) 117896. doi: 10.1016/j.apcatb.2019.117896

    121. [121]

      Y. Abghoui, A.L. Garden, V.F. Hlynsson, et al., Phys. Chem. Chem. Phys. 17 (2015) 4909–4918. doi: 10.1039/C4CP04838E

    122. [122]

      Y. Abghoui, A.L. Garden, J.G. Howalt, et al., ACS Catal. 6 (2016) 635–646. doi: 10.1021/acscatal.5b01918

    123. [123]

      Q. Li, L. He, C. Sun, et al., J. Phys. Chem. C 121 (2017) 27563–27568. doi: 10.1021/acs.jpcc.7b10522

    124. [124]

      C. Lv, Y. Qian, C. Yan, et al., Angew. Chem. Int. Ed. 57 (2018) 10246–10250. doi: 10.1002/anie.201806386

    125. [125]

      H. Jin, L. Li, X. Liu, et al., Adv. Mater. 31 (2019) 1902709. doi: 10.1002/adma.201902709

    126. [126]

      X. Li, T. Li, Y. Ma, et al., Adv. Energy Mater. 8 (2018) 1801357. doi: 10.1002/aenm.201801357

    127. [127]

      L. Zhang, X. Ji, X. Ren, et al., Adv. Mater. 30 (2018) 1800191. doi: 10.1002/adma.201800191

    128. [128]

      J. Zhang, X. Tian, M. Liu, et al., J. Am. Chem. Soc. 141 (2019) 19269–19275. doi: 10.1021/jacs.9b02501

    129. [129]

      B.H.R. Suryanto, D. Wang, L.M. Azofra, et al., ACS Energy Lett. 4 (2019) 430–435. doi: 10.1021/acsenergylett.8b02257

    130. [130]

      L.M. Azofra, C. Sun, L. Cavallo, et al., Chem. Eur. J. 23 (2017) 8275–8279. doi: 10.1002/chem.201701113

    131. [131]

      H. Zhang, C. Cui, Z. Luo, J. Phys. Chem. C 124 (2020) 6260–6266. doi: 10.1021/acs.jpcc.0c00486

    132. [132]

      W. Zhang, B.W. Zhang, Nanomicro Lett. 13 (2021) 106.

    133. [133]

      X. Yan, Y. Jia, X. Yao, Chem. Soc. Rev. 47 (2018) 7628–7658. doi: 10.1039/C7CS00690J

    134. [134]

      C. Zhu, S. Fu, Q. Shi, et al., Angew. Chem. Int. Ed. 56 (2017) 13944–13960. doi: 10.1002/anie.201703864

    135. [135]

      Z. Wei, Y. Zhang, S. Wang, et al., J. Mater. Chem. A 6 (2018) 13790–13796. doi: 10.1039/C8TA03989E

    136. [136]

      C. Choi, S. Back, N.Y. Kim, et al., ACS Catal. 8 (2018) 7517–7525. doi: 10.1021/acscatal.8b00905

    137. [137]

      X. Chen, X. Zhao, Z. Kong, et al., J. Mater. Chem. A 6 (2018) 21941–21948. doi: 10.1039/C8TA06497K

    138. [138]

      W. Yang, H. Huang, X. Ding, et al., Electrochim. Acta 335 (2020) 135667. doi: 10.1016/j.electacta.2020.135667

    139. [139]

      C. Guo, X. Liu, L. Gao, et al., Appl. Catal. B 263 (2020) 118296. doi: 10.1016/j.apcatb.2019.118296

    140. [140]

      F. Lai, N. Chen, X. Ye, et al., Adv. Funct. Mater. 30 (2020) 1907376. doi: 10.1002/adfm.201907376

    141. [141]

      Y. Shi, Y. Zhou, D.R. Yang, et al., J. Am. Chem. Soc. 139 (2017) 15479–15485. doi: 10.1021/jacs.7b08881

    142. [142]

      J. Su, R. Ge, Y. Dong, et al., J. Mater. Chem. A 6 (2018) 14025–14042. doi: 10.1039/C8TA04064H

    143. [143]

      Y. Zhang, L. Guo, L. Tao, et al., Small Methods 3 (2019) 1800406. doi: 10.1002/smtd.201800406

    144. [144]

      X. Yang, K. Li, D. Cheng, et al., J. Mater. Chem. A 6 (2018) 7762–7769. doi: 10.1039/C8TA01078A

    145. [145]

      H. Wang, L. Wang, Q. Wang, et al., Angew. Chem. Int. Ed. 57 (2018) 12360–12364. doi: 10.1002/anie.201805514

    146. [146]

      X. Yu, P. Han, Z. Wei, et al., Joule 2 (2018) 1610–1622. doi: 10.1016/j.joule.2018.06.007

    147. [147]

      X. Mao, S. Zhou, C. Yan, et al., Phys. Chem. Chem. Phys. 21 (2019) 1110–1116. doi: 10.1039/C8CP07064D

    148. [148]

      Y. Tian, D. Xu, K. Chu, et al., J. Mater. Sci. 54 (2019) 9088–9097. doi: 10.1007/s10853-019-03538-0

    149. [149]

      J.S. Kim, B. Kim, H. Kim, et al., Adv. Energy Mater. 8 (2018) 1702774. doi: 10.1002/aenm.201702774

    150. [150]

      J. Zhang, R. Yin, Q. Shao, et al., Angew. Chem. Int. Ed. 58 (2019) 5609–5613. doi: 10.1002/anie.201900167

    151. [151]

      W. Xu, S. Zhu, Y. Liang, et al., J. Mater. Chem. A 5 (2017) 18793–18800. doi: 10.1039/C7TA05314B

    152. [152]

      P. Cai, J. Huang, J. Chen, et al., Angew. Chem. Int. Ed. 56 (2017) 4858–4861. doi: 10.1002/anie.201701280

    153. [153]

      S.J. Li, D. Bao, M.M. Shi, et al., Adv. Mater. 29 (2017) 1700001. doi: 10.1002/adma.201700001

    154. [154]

      Z. Wang, Y. Li, H. Yu, et al., ChemSusChem 11 (2018) 3480–3485. doi: 10.1002/cssc.201801444

    155. [155]

      M.M. Shi, D. Bao, B.R. Wulan, et al., Adv. Mater. 29 (2017) 1606550. doi: 10.1002/adma.201606550

    156. [156]

      H. Yu, Z. Wang, W. Tian, et al., Mater. Today Energy 20 (2021) 100681. doi: 10.1016/j.mtener.2021.100681

    157. [157]

      R. Lan, J.T.S. Irvine, S. Tao, Sci. Rep. 3 (2013) 1145. doi: 10.1038/srep01145

    158. [158]

      M. Kitano, S. Kanbara, Y. Inoue, et al., Nat. Commun. 6 (2015) 6731. doi: 10.1038/ncomms7731

    159. [159]

      M.M. Shi, D. Bao, S.J. Li, et al., Adv. Energy Mater. 8 (2018) 1800124. doi: 10.1002/aenm.201800124

    160. [160]

      C. Guo, J. Ran, A. Vasileff, et al., Energy Environ. Sci. 11 (2018) 45–56. doi: 10.1039/C7EE02220D

    161. [161]

      P. Li, W. Fu, P. Zhuang, et al., Small 15 (2019) 1902535. doi: 10.1002/smll.201902535

    162. [162]

      P. De Luna, W. Liang, A. Mallick, et al., ACS Appl. Mater. Interfaces 10 (2018) 31225–31232. doi: 10.1021/acsami.8b04848

    163. [163]

      Y. Song, D. Johnson, R. Peng, et al., Sci. Adv. 4 (2018) e1700336.

    164. [164]

      W.Y. Gao, Y.C. Hao, X. Su, et al., Chem. Commun. 55 (2019) 10705–10708. doi: 10.1039/C9CC04691G

    165. [165]

      H. Ying, T. Chen, C. Zhang, et al., J. Colloid Interface Sci. 602 (2021) 64–72. doi: 10.1016/j.jcis.2021.05.185

    166. [166]

      H. Wang, H. Yu, Y. Li, et al., Nanotechnology 29 (2018) 175403. doi: 10.1088/1361-6528/aaaf3f

    167. [167]

      H. Yu, Z. Wang, D. Yang, et al., J. Mater. Chem. A 7 (2019) 12526–12531. doi: 10.1039/C9TA03297E

    168. [168]

      F. Pang, Z. Wang, K. Zhang, et al., Nano Energy 58 (2019) 834–841. doi: 10.1016/j.nanoen.2019.02.019

    169. [169]

      F. Lai, W. Zong, G. He, et al., Angew. Chem. Int. Ed. 59 (2020) 13320–13327. doi: 10.1002/anie.202003129

    170. [170]

      B. Xu, L. Xia, F. Zhou, et al., ACS Sustain. Chem. Eng. 7 (2019) 2889–2893. doi: 10.1021/acssuschemeng.8b05007

    171. [171]

      C. Ma, N. Zhai, B. Liu, et al., Electrochim. Acta 370 (2021) 137695. doi: 10.1016/j.electacta.2020.137695

    172. [172]

      L. Wen, X. Li, R. Zhang, et al., ACS Appl. Mater. Interfaces 13 (2021) 14181–14188. doi: 10.1021/acsami.0c22767

    173. [173]

      G. Zhang, Q. Ji, K. Zhang, et al., Nano Energy 59 (2019) 10–16. doi: 10.1016/j.nanoen.2019.02.028

    174. [174]

      L. Yang, T. Wu, R. Zhang, et al., Nanoscale 11 (2019) 1555–1562. doi: 10.1039/C8NR09564G

    175. [175]

      K. Chu, F. Liu, J. Zhu, et al., Adv. Energy Mater. 11 (2021) 2003799. doi: 10.1002/aenm.202003799

    176. [176]

      Z. Sun, R. Huo, C. Choi, et al., Nano Energy 62 (2019) 869–875. doi: 10.1016/j.nanoen.2019.06.019

    177. [177]

      J. Xia, H. Guo, M. Cheng, et al., J. Mater. Chem. A 9 (2021) 2145–2151. doi: 10.1039/D0TA08089F

    178. [178]

      X. Yang, F. Ling, J. Su, et al., Appl. Catal. B 264 (2020) 118477. doi: 10.1016/j.apcatb.2019.118477

    179. [179]

      W. Zang, T. Yang, H. Zou, et al., ACS Catal. 9 (2019) 10166–10173. doi: 10.1021/acscatal.9b02944

    180. [180]

      X.W. Lv, X.L. Liu, L.J. Gao, et al., J. Mater. Chem. A 9 (2021) 4026–4035. doi: 10.1039/D0TA11244E

    181. [181]

      S. Xiao, F. Luo, H. Hu, et al., Chem. Commun. 56 (2020) 446–449. doi: 10.1039/C9CC07708A

    182. [182]

      T. Wu, H. Zhao, X. Zhu, et al., Adv. Mater. 32 (2020) 2000299. doi: 10.1002/adma.202000299

    183. [183]

      T. Wu, X. Li, X. Zhu, et al., Chem. Commun. 56 (2020) 1831–1834. doi: 10.1039/C9CC09179C

    184. [184]

      W. Peng, M. Luo, X. Xu, et al., Adv. Energy Mater. 10 (2020) 2001364. doi: 10.1002/aenm.202001364

    185. [185]

      K. Chu, Y.P. Liu, Y.B. Li, et al., Appl. Catal. B 264 (2020) 118525. doi: 10.1016/j.apcatb.2019.118525

    186. [186]

      H. Chen, Y. Xiao, C. Chen, et al., ACS Appl. Mater. Interfaces 11 (2019) 11459–11465. doi: 10.1021/acsami.8b22564

    187. [187]

      S. Zhao, H.X. Liu, Y. Qiu, et al., J. Mater. Chem. A 8 (2020) 6586–6596. doi: 10.1039/D0TA00658K

    188. [188]

      X. Wang, W. Wang, M. Qiao, et al., Sci. Bull. 63 (2018) 1246–1253. doi: 10.1016/j.scib.2018.07.005

    189. [189]

      K. Chu, H. Nan, Q. Li, et al., J. Energy Chem. 53 (2021) 132–138. doi: 10.1016/j.jechem.2020.04.074

    190. [190]

      T. Pan, L. Wang, Y. Shen, et al., ACS Appl. Mater. Interfaces 14 (2022) 14474–14481. doi: 10.1021/acsami.2c00018

    191. [191]

      Y. Liu, W. Wang, S. Zhang, et al., ACS Sustain. Chem. Eng. 8 (2020) 2320–2326. doi: 10.1021/acssuschemeng.9b07679

    192. [192]

      Y. Liu, M. Han, Q. Xiong, et al., Adv. Energy Mater. 9 (2019) 1803935. doi: 10.1002/aenm.201803935

    193. [193]

      H. Cheng, L.X. Ding, G.F. Chen, et al., Adv. Mater. 30 (2018) 1803694. doi: 10.1002/adma.201803694

    194. [194]

      H. Sun, H.Q. Yin, W. Shi, et al., Nano Res. 15 (2022) 3026–3033. doi: 10.1007/s12274-021-3937-3

    195. [195]

      Z. Wang, J. Niu, Y. Xu, et al., ACS Sustain. Chem. Eng. 8 (2020) 12588–12594. doi: 10.1021/acssuschemeng.0c03970

  • Scheme 1  Schematic diagram of defects engineering for NRR (vacancy, dopant, amorphism, edge/corner, and porousness). The advantages of defects towards NRR can be summarized into the following four points: (1) adjusting electronic structure; (2) modifying physical structure; (3) increasing reaction active site; and (4) enhancing nitrogen adsorption.

    Figure 1  Characterizations of defects in nanomaterials. (a) HAADF-STEM images of SA-Fe-NHP highlighting isolated Fe atoms. Reproduced with permission [62]. Copyright 2020, John Wiley and Sons. (b) HAADF-STEM images of sulfur vacancies on MoS2. Reproduced with permission [63]. Copyright 2020, American Chemical Society. (c) HRTEM image of defect rich SnO2, and the arrows indicate the grain-boundary surface terminations. Reproduced with permission [64]. Copyright 2021, Elsevier. (d) XPS spectra of a/c-Rh nanosheets (NSs) and c-Rh NSs. Reproduced with permission [65]. Copyright 2021, John Wiley and Sons. (e) EPR spectra of CoMnNiOx with different contents of oxygen vacancies. Reproduced with permission [66]. Copyright 2022, Elsevier. (f) PL spectra of NbOx and Nb2O5. Reproduced with permission [67]. Copyright 2022, Royal Society of Chemistry. (g) XRD patterns of amorphous and crystalline Ru/CNTs catalyst. Reproduced with permission [68]. Copyright 2022, American Chemical Society. (h) Positron lifetime spectra of ultrathin CoSe2 nanosheets and bulk CoSe2. Reproduced with permission [69]. Copyright 2014, American Chemical Society. (i) EXAFS spectra of Fe foil, Fe-PPy and Fe2O3 with different doping sites. Reproduced with permission [70]. Copyright 2021, Royal Society of Chemistry.

    Figure 2  (a) Illustration of the KOH etching synthetic method of PD-C. Reproduced with permission [76]. Copyright 2019, John Wiley and Sons. (b) W1.33C MXene with ordered vacancies after selective etching and delamination. Reproduced with permission [79]. Copyright 2018, John Wiley and Sons. (c) Schematic illustration of the preparation process of SG-P and the plasma system. Reproduced with permission [83]. Copyright 2017, Elsevier. (d) Schematic illustration of the preparation procedure for Ni-implanted N-doped nanoporous carbons. Reproduced with permission [87]. Copyright 2020, John Wiley and Sons. (e) Synthetic procedure of Ce-Bi2WO6. Reproduced with permission [96]. Copyright 2021, American Chemical Society.

    Figure 3  (a) O 1s XPS spectra of TiO2 under different temperatures. (b) NH3 yield rates and FEs of different catalysts at −0.12 V. (c) Optimized geometries (top-view) of pristine and the most stable surface oxygen vacancy containing anatase (101). The sky blue and red balls indicate Ti and O atoms, respectively and the free energy ΔG for NRR is shown in (d). Reproduced with permission [120]. Copyright 2019, Elsevier.

    Figure 4  Characterizations of PCN with and without NVs: (a) XRD patterns and (b) EPR spectra. (c) N 1s XPS spectra. (d) Ammonia yield. (e) Cycling test of PCN-NV. (f) Free energy diagram for NRR on NVs engineered PCN at equilibrium potential. Reproduced with permission [124]. Copyright 2018, John Wiley and Sons. (g) NH3 yields and FE of NV-W2N3 and W2N3 at −0.2 V vs. RHE. Reproduced with permission [125]. Copyright 2019, John Wiley and Sons.

    Figure 5  (a) Free-energy profile for NRR at MoS2 edge site. Reproduced with permission [126] (b) Scheme of Ru/2H-MoS2 electronic properties and DFT for the mechanism of NRR. Reproduced with permission [129]. Copyright 2019, American Chemical Society. (c) Co-doped MoS2−x NRR performance at different potentials. (d) NRR performance of MoS2−x, Co-doped MoS2−x and CoS2 at −0.3 V. (e) Free-energy diagram of ammonia synthesis. Reproduced with permission [128]. Copyright 2019, American Chemical Society.

    Figure 6  (a) Top view of the distribution of spin density. The positive and negative densities of the spin are blue and yellow isosurfaces. Reproduced with permission [135]. Copyright 2018, Royal Society of Chemistry. (b) Differences between H adsorption free energy on the SAC (ΔGSAC(*H)) and adsorption free energy on the surface (ΔGsurface(*H)) of the same metal atoms (eV). Horizontal dashed line denotes ΔGSAC(*H) = ΔGsurface(*H). Reproduced with permission [136]. Copyright 2018, American Chemical Society. (c) ΔGmax as a function of adsorption energy of N2. Reproduced with permission [138]. Copyright 2019, Elsevier. (d) NH3 yields and FEs of the pure Ni2P catalyst as well as the Fe-doped Ni2P catalysts with various Fe/Ni ratios (including Fe0.2Ni1.8P, Fe0.25Ni1.75P, Fe0.4Ni16P and Fe1Ni1P). (e) The structures of intermediates for N2 reduction on Fe-doped Ni2P (001). The adsorption site is denoted by an asterisk (*) and the H, N, Ni, Fe and P atoms are shown in light pink, blue, gray-blue, brown and pink colors, respectively. Reproduced with permission [139]. Copyright 2020, Elsevier. (f) Band structure diagram for Fe-ReS2 and pure ReS2. Reproduced with permission [140]. Copyright 2020, John Wiley and Sons.

    Figure 7  (a) Schematic representation for the electrochemical NRR by catalysts of c-Au/rGO and a-Au/CeOx-rGO. (b, c) NRR performance of a-Au/CeOx-rGO and other catalysts at −0.2 V vs. RHE, Reproduced with permission [153]. Copyright 2017, John Wiley and Sons. (d) Representative XRD patterns of Cu/rGO, Pd/rGO and Pd0.2Cu0.8/rGO. (e, f) NRR performance of Pd0.2Cu0.8/rGO and different catalysts at −0.2 V vs. RHE. Reproduced with permission [159]. Copyright 2018, John Wiley and Sons. (g) NRR performance of BVC-A at −0.2 V vs. RHE. (h) Illustration of fabricating BVC-A and BVC-crystalline (BVC-C) NRR electrocatalysts. Reproduced with permission [112]. Copyright 2018, John Wiley and Sons.

    Figure 8  (a, b) Ammonia output and FE of CNS in different countercations with the order of Li+ > Na+ > K+. (c, d) FEs and ammonia formation rates of various catalysts. Reproduced with permission [163]. Copyright 2018, American Association for the Advancement of Science. (e) Schematic diagram of NRR on triangular AgTPs and rounded AgNPs. (f) NRR performance of AgTPs and AgNPs. Reproduced with permission [164]. Copyright 2019, The Royal Society of Chemistry.

    Figure 9  (a) SEM image of Pd3Cu1 alloy. Reproduced with permission [168]. Copyright 2019, Elsevier. (b) NH3 yield (bar graph) and Faradaic efficiency (blue point) by using Vr-ReSe2@CBC NFs and Vr-ReSe2 bulk as contrast samples. (c) DOS for V-ReSe2 and ReSe2 models, as well as the charge density difference of V-ReSe2 (The pink and yellow areas present electrons accumulation and depletion, respectively) (left figure), and free energy diagrams for nitrogen reduction reaction on the surface of V-ReSe2 and ReSe2 catalysts (right figure). (d) Schematic diagram of synthesis process of CBC@Vr-ReSe2@CBC. (e) NRR performance (NH3 yield and Faradaic efficiency) of CBC@Vr-ReSe2@CBC NFs in 0.1 mol/L Na2SO4 electrolyte under ambient conditions. (f) Distributions of concentration of H+ ions in the electrode in the COMSOL multiphysics simulations. Reproduced with permission [169]. Copyright 2020, John Wiley and Sons.

    Table 1.  A brief summary of typical NRR electrocatalysts with defects identified in the latest years and their catalytic efficiency.

    下载: 导出CSV
  • 加载中
计量
  • PDF下载量:  2
  • 文章访问数:  1456
  • HTML全文浏览量:  40
文章相关
  • 发布日期:  2023-07-15
  • 收稿日期:  2022-08-08
  • 接受日期:  2022-09-20
  • 修回日期:  2022-09-02
  • 网络出版日期:  2022-09-23
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

/

返回文章