TM₃ (TM = V, Fe, Mo, W) single-cluster catalyst confined on porous BN for electrocatalytic nitrogen reduction

Highlights

• Confined metal clusters in the triplet form as sub-nanometer reactors for NRR.

• Mo₃@p-BN exhibits excellent catalytic activity and selectivity.

• A rather low limiting potential (−0.34 V) was obtained.

• Mo atoms act as “cache” to accelerate electron transfer.

• Synergistic effect makes the “donation–backdonation” mechanism more efficient.

Abstract

Confined metal clusters as sub-nanometer reactors for electrocatalytic N₂ reduction reaction (eNRR) have received increasing attention due to the unique metal-metal interaction and higher activity than single-atom catalysts. Herein, the inspiration of the superior capacitance and unique microenvironment with regular surface cavities of the porous boron nitride (p-BN) nanosheets, we systematically studied the catalytic activity for NRR of transition-metal single-clusters in the triplet form (V₃, Fe₃, Mo₃ and W₃) confined in the surface cavities of the p-BN sheets by spin-polarized density functional theory (DFT) calculations. After a two-step screening strategy, Mo₃@p-BN was found to have high catalytic activity and selectivity with a rather low limiting potential (–0.34 V) for the NRR. The anchored Mo₃ single-cluster can be stably embedded on the surface cavities of the substrate preventing the diffusion of the active Mo atoms. More importantly, the Mo atoms in the Mo₃ single-cluster would act as “cache” to accelerate electron transfer between active metal centers and nitrogen-containing intermediates via the intimate Mo-Mo interactions. The cooperation of Mo atoms can also provide a large number of occupied and unoccupied d orbitals to make the "donation–backdonation" mechanism more effective. This work not only provides a quite promising electrocatalyst for NRR, but also brings new insights into the rational design of triple-atom NRR catalysts.

Introduction

Ammonia (NH₃) is a significant material for synthetic chemicals, which is widely used in agricultural fertilizer, industrial power generation combustion, energy storage and other fields due to its merits such as high energy density and easy transmission [1]. The N₂ can be converted into NH₃ by some prokaryotes through biological nitrogen fixation in the atmosphere, but the ammonia transferred by this method is far from satisfying the requirements of growing human activities. Generally, industrial NH₃ production mainly relies on the traditional Haber-Bosch process, which converts high-purity N₂ and H₂ into NH₃ on Fe-based catalysts [2]. However, this process operates at elevated temperature and pressure, which would consume a huge amount of energy and emit lots of greenhouse gasses (CO₂), aggravating environmental and energy problems [3, 4]. Therefore, it is urgent to find an environmentally friendly technology for synthesizing NH₃ [5], [6], [7]. Among many NH₃ production methods, the electrochemical/photochemical nitrogen fixation stands out because of mild conditions, simple process, safety, high activity, etc. [8], [9], [10]. However, most photo/electrocatalysts suffer from low activity, low selectivity and poor durability. It is of great importance to find an efficient catalyst to reduce the activation energy barrier of N₂ reduction and accelerate the NRR process.

Among the various eNRR catalysts, transition-metal (TM) atomic catalysts are extensively researched, such as single-atom catalyst (SAC) [11], [12], [13], [14]. For example, our previous work [15] found that V@BN has excellent NRR catalytic performance with a rather low overpotential of 0.25 V by high-throughput screening of 18 candidates with a single TM (3d, 4d and 5d) atom supported on a defective hexagonal boron nitride (h-BN) nanosheet. Besides, Chen et al. [16] studied a series of single metal atoms supported on g-C₃N₄ through DFT calculations, and among all the candidates, the onset potentials for NRR of the g-C₃N₄ catalysts supported by Ti, Co, Mo, W and Pt are lower than that of the Ru (0001) surface. Likewise, Zhao and Chen [11] found that Mo@h-BN catalyst shows high catalytic activity for the activation and reduction of N₂ through an enzymatic mechanism with a low overpotential of only 0.19 V. Some other SACs are also supposed to have high catalytic activity towards NRR through theoretical calculations, such as Fe@MoS₂ [17], Re@MoS₂ [18] and Re@WS₂ [19].

Although the SACs have achieved remarkable success in a wide spectrum of catalytic reactions, there still exist several obstacles for further improving their activities, such as the low loading amount of single-atom catalyst, the lack of cooperation between catalysts due to the large distance between them, and the unfavorable electron transfer rate between active metal center and substrate [20], [21], [22], [23]. In comparison to the SACs, the double-atom catalysts (DACs) and triple-atom catalysts (TACs) are suggested to possess higher catalytic activity owing to the synergistic effect of the two/three metal atoms [24, 25]. For example, Li et al. [26] found that the Fe-Rh double sites can synergistically optimize the adsorption and desorption of nitrogen-containing intermediates, and obtain significant NRR activity with the onset potential of –0.22 V. In addition, He and co-workers [27] verified that an asymmetrical dual-metal dimer catalytic center through FeMo cofactors can efficiently catalyze the reduction of N₂ to NH₃. For the TACs, recently, Ji and co-workers [28] found that Ru₃ trimer cluster anchored on N-doped carbon (Ru₃/CN) has a better catalytic conversion of 2-aminobenzyl alcohol than the SAC counterpart (Ru₁/CN). And Zheng et al. systematically studied the catalytic performance of various triple-TM-atom clusters immobilized on nitrogen-doped graphene for NRR and found that Co₃N4 possesses the highest activity with a limiting potential of –0.41 V [29]. There are relatively few reported cases of TACs for NRR in experiments, but still a lot of polyatomic records, such as Ag₃/Al₂O₃ catalyst for propylene epoxidation [30], and experimentally obtained Ru₃/CN for efficient oxidation of alcohols [28]. Interestingly, Wang et al. [20] recently experimentally constructed a “subnano reactor” by confining multiple Fe and Cu atoms in the surface cavities of g-C₃N₄ for high-efficiency NRR electrocatalysis. They found that a large portion of Fe and Cu atoms are in a triplet form with an average distance of 1.8–2.5 Å. The intimate contact between Fe and Cu atoms simultaneously accelerates the adsorption of N₂ and improves electrons transfer for NRR. Therefore, the multiple metal single-cluster catalysts are desirable for NRR due to their potential synergetic effect on modifying the electronic structure and lowing the NRR barrier.

Besides suitable atomic catalysts, an appropriate substrate is also necessary for the high activity of NRR. 2D materials usually have better performances in electrocatalysis than their 3D counterparts due to the high specific surface areas, abundant active sites, and allowing electronic modulation by doping and strain [31], [32], [33], [34], [35], [36]. Therefore, 2D materials are the most widely investigated substrates for electro/photocatalysis both in experiments and simulations [37, 38]. In this work, the porous BN (p-BN) nanosheet is chosen as the substrate, because of high specific surface areas and superior capacitance performance [39]. More importantly, the abundant and evenly distributed cavities in the p-BN can provide large space and numerous anchored sites for capturing active metal atoms, which can greatly restrict the diffusion of the metal atoms on the surface. Herein, we designed a series of TACs anchored on p-BN as electrocatalysts for NRR, including V₃, Fe₃, Mo₃ and W₃. Many previous reports showed that V, Fe, Mo, and W SACs possess excellent nitrogen fixation performance. For example, as mentioned previously, Mo@h-BN [11], V@BN [15], as well as W@g-C₃N₄ [16] exhibit high catalytic activity toward NRR. Moreover, Ling and co-workers proved that a single W atom embedded in graphene with three C atoms coordination (W₁C₃) exhibits the best performance from 540 systems [40]. And then, Li et al. found that the synergy between the graphene and FeN₃ equips the system with novel features for the NRR process [41]. In addition, Fe, Mo, and V are the elements existing in known nitrogenases for biological N₂ fixation, such as MoFe-, VFe-, FeFe-N₂ase [42]. Besides, Fe, Mo, V, and W metals are not as expensive as Pt, Pd and Au metals. Inspired by the above reasons, we selected these four metal TACs: V₃, Fe₃, Mo₃ and W₃, as the research objects. Subsequently, using the two-step screening strategy proposed by Ling et al. [40], the promising candidates were selected from the four catalysts. The final screened Mo₃@p-BN shows high catalytic activity, and its onset potential is only 0.18 V, which is much smaller than those of previously reported SACs and TACs [11, 16, 20, 29, 43, 44]. Our results reveal that the synergistic effect of Mo atoms not only accelerates electron transfer between active metal centers and nitrogen-containing intermediates through the intimate Mo-Mo interactions, but also provides enough occupied and unoccupied d orbitals to make the "donation–backdonation" mechanism more effective.