NH
3 holds significant importance as a vital chemical compound within the industrial sector, playing a crucial role in modern economic development [
1,
2]. The electrochemical reduction of N
2 (NRR) with minimal emissions of CO
2 stands as a sustainable and environmentally friendly approach for N
2 fixation under ambient conditions, facilitating the revolution for NH
3 synthesis [
3-
7]. However, NRR encounters obstacles attributed to the concurrent hydrogen evolution reaction (HER) and the formidable N≡N bond [
8]. In contrast, electrochemical reduction of NO to NH
3 (NORR) offers a more energy-efficient and practical pathway for sustainable NH
3 synthesis [
9-
14]. Additionally, NO itself serves as a prevalent atmospheric pollutant generated by vehicle exhaust and incomplete combustion of fossil fuels. NORR emerges as an efficient and eco-friendly technique for simultaneous NO treatment and NH
3 electro-synthesis. Nevertheless, NORR presents a multi-step process involving intricate five-electron mechanism, resulting in slow reaction kinetics [
15-
17]. Consequently, the exploration of highly performant catalysts becomes imperative to enhance the NO-to-NH
3 conversion efficiency.
Single-atom catalysts (SACs) have garnered increasing attention in heterogeneous catalysis, owing to the exceptional potential in delivering high activity and selectivity for a diverse range of chemical reactions [
18-
26]. To date, a variety of transition metal-based SACs are developed as NORR electrocatalysts, attributed to the ability of transition metal to augment NO adsorption and activation facilitated by the partially occupied d-orbitals [
27-
35]. Nevertheless, it has been noted that these d-orbitals also promote the binding of
*H on catalyst surfaces, subsequently promoting the undesired HER and thus impeding NH
3 synthesis [
36]. Consequently, the exploration of alternative elements as active sites for SACs, beyond transition metal atoms, remains a topic of ongoing research. Promisingly, main-group metals (Al, Ga, In, Sn, Sb, and Bi,
etc.) with partially occupied p-orbitals exhibit robust ability in restraining adsorption of
*H, and can overcome the shortcoming of transition metal-based SACs [
37-
41]. Hence, main-group metals are expected to serve as promising SACs for NORR [
42-
45]. For instance, Chen
et al. [
42] experimentally reported that Sb single atoms confined in MoO
3 were highly efficient for the NORR with the NH
3 yield rate of 273.5 µmol h
−1 cm
−2 and Faradaic efficiency of 91.7%. Wu
et al. [
45] computationally predicted that single Mg/Al/Ga atom supported on N-doped graphene can act as active centers for direct NO-to-NH
3 conversion. Despite the great potential, main-group metal-based SACs for NORR have not been explored extensively, and it is meaningful to extend the scope both from theoretical and experimental aspects.
Recently, graphene-like boron carbide (BC
3) monolayer, a semiconductor with a bandgap spanning between 0.46 eV and 0.73 eV, garners enormous attention due to its superior mechanical strength, optical characteristics, and thermal stability [
46,
47]. These exceptional properties enable it to hold great potential in various fields, including hydrogen storage [
48,
49], sensor technology [
50,
51], lithium-ion batteries [
52], and catalytic applications [
53-
57]. Particularly, the graphitic BC
3 nanosheet has been successfully synthesized in experiment
via the reaction of benzene with sodium borohydride at 800 ℃ and hold great potential to host SACs for achieving superior electrocatalytic performance [
58]. Herein, we systematically investigate the potential of nine main-group metals (
i.e., Mg, Ca, Al, Ga, In, Ge, Sn, Sb, and Bi) embedded BC
3 monolayer as SACs for the NORR toward NH
3 synthesis. Through first-principles calculations, we screen out Al@V
B, Ga@V
B, and Ge@V
C systems with outstanding NORR activity and high selectivity for NH
3 production from a pool of 18 candidate systems. Importantly, the NORR process is found to occur spontaneously in the three screened systems. Our research results provide a valuable guidance for further developing in NORR catalysts, as well as demonstrate the great potential of main-group metals as SACs.
All spin-polarized first-principles calculations are performed using the Vienna Ab initio Simulation Package (VASP) [
59] with the projector augmented wave (PAW) approach [
60] and the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional [
61]. Additionally, the van der Waals weak interactions are further treated using Grimme’s semi-empirical correction scheme (DFT-D3) [
62]. A plane-wave energy cutoff of 500 eV is applied in all calculations. To minimize interactions between two periodic images, a vacuum layer of 20.0 Å is set in the z-direction. For structure optimization and electronic property calculations of a 3 × 3 × 1 supercell, 2 × 2 × 1 and 4 × 4 × 1 k-point grids centered at the Γ point are used following the Monkhorst-Pack scheme [
63], respectively. The convergence thresholds are set to be 10
−5 eV for the total energy and 0.02 eV/Å for atomic forces, respectively. More other computational details are provided in Supporting information.
The pristine unit cell of BC
3 monolayer composed of two boron (B) atoms and six carbon (C) atoms shows a graphene-like planar structure with the space group of
P6/
mmm (No. 191), in which each B atom is coordinated to three neighboring C atoms. The optimized lattice parameters are
a =
b = 5.17 Å, and the lengths of C–C and C-B bond are 1.42 and 1.57 Å, respectively, which are consistent with previous reports [
64-
66]. To simulate the main-group metal-embedded BC
3 monolayer, we construct a 3 × 3 × 1 supercell and consider four kinds of vacancies, including B monovacancy (V
B), C monovacancy (V
C), BC divacancy (V
BC), and CC divacancy (V
CC). To evaluate the formation possibility of the four vacancies, we calculate their respective formation energies (
Ef), which are 5.37 eV for V
B, 4.44 eV for V
C, 7.47 eV for V
BC, and 6.05 eV for V
CC, indicating the thermodynamically favorable formation of monovacancies (V
B and V
C) rather than divacancies (V
BC and V
CC). Compared to that of monovacancy in graphene (7.45 eV) [
67-
69], the lower
Ef values of V
B and V
C monovacancies suggest the high feasibility for experimental synthesis. Consequently, we concentrate on V
B and V
C monovacancies for main-group metal-based SACs.
As shown in
Figs. 1a and
b, nine main-group metals (
i.e., Mg, Ca, Al, Ga, In, Ge, Sn, Sb, and Bi) are chosen and embedded into V
B and V
C monovacancies, forming the structures of M@V
B and M@V
C, respectively. To quantitatively evaluate the stability of these M@V
B and M@V
C systems, we calculate the binding energy (
Eb) as shown in
Fig. 1c, and find that all systems exhibit negative
Eb values ranging from −6.13 eV to −0.67 eV, indicating the strong binding interactions between main-group metals and BC
3 monolayer. Additionally, considering the large cohesive energy (
Ecoh) of metal bulk, the experimental feasibility of main-group metal-based SACs needs to be further assessed. Our results shown in
Fig. 1c reveal that nine candidate systems (
i.e., Al@V
B, Al@V
C, Ga@V
B, Ga@V
C, In@V
B, Ge@V
B, Ge@V
C, Sn@V
B, and Sb@V
B) have smaller
Eb in comparison with their corresponding
Ecoh, and thus satisfy the requirement
Eb -
Ecoh < 0, indicating that the main-group metals in these systems prefer to be stably anchored on BC
3 monolayer, without aggregating into metal clusters. To further examine the thermal stabilities of the nine systems, we also perform ab initio molecular dynamics (AIMD) simulations in the NVT ensemble, with a total time scale of 5 ps at 500 K. As shown in Fig. S1 (Supporting information), the total energies of the nine systems oscillate slightly, and their corresponding geometric structures remain intact without deformation at the end of the AIMD simulation, verifying the good thermodynamic stability. Therefore, Al@V
B, Al@V
C, Ga@V
B, Ga@V
C, In@V
B, Ge@V
B, Ge@V
C, Sn@V
B, and Sb@V
B are selected as potential candidates for the subsequent NORR.
It is well-established that NO adsorption and activation, the first step of the NORR process, plays a crucial role in the following reaction pathways [
70]. Therefore, we consider three initial NO adsorption configurations (
i.e., N-end, O-end, and side-on) for the above screened nine systems, and calculate their adsorption free energies (Δ
G*NO) as shown in
Fig. 2a. For all candidate systems, we can find that NO molecule is favored to adsorb
via the N-end configuration, while the side-on adsorption is unstable and will be transformed into the N-end configuration after fully structural relaxation. To ensure the effective activation of adsorbed NO molecule, spontaneous chemisorption (Δ
G*NO < 0) is considered a necessary prerequisite for the NORR. Evidently, the Sb@V
B system should be filter out from the NORR candidates because of the weak physisorption of NO (Δ
G*NO = 0.28 eV). The remaining eight candidates present sufficient negative Δ
G*NO values, ranging from −0.96 eV to −3.37 eV for the N-end configuration, which indicates that NO molecule would spontaneously chemisorb on these systems. Meanwhile, as shown in
Fig. 2b, the chemisorbed NO molecules have considerable N–O bond elongation from 1.17 Å (free NO molecule) to 1.19 - 1.21 Å, indicating the strong activation of NO molecule. To evaluate the competitive effects of NO adsorption, we further calculate the adsorption free energies of N
2 (Δ
G*N2) and NH
3 (Δ
G*NH3), and compare with NO (Δ
G*NO). As shown in Fig. S2 (Supporting information), Δ
G*NO is more negative than Δ
G*N2 and Δ
G*NH3 for the M@V
B and M@V
C systems, indicating the NO molecule could be preferentially adsorbed on them.
To elucidate the underlying mechanism of NO adsorption and activation on these main-group metal-embedded BC
3 systems, we further calculate the partial density of states (PDOS) and charge density differences to investigate the interaction between main-group metal atom and NO molecule. Take for instance the Al@V
B system illustrated in
Fig. 3a. When a NO molecule chemisorbs on it, the
π* states of NO would interact with the Al-3p states, resulting in the appearance of occupied and unoccupied p-
π* states around the Fermi level (
EF). Additionally, due to the hybridization between the occupied Al-3p states and NO-
σ states below the
EF, the occupied p-
σ states can be observed. As displayed in
Fig. 3b, we can see not only a significant charge transfer with 0.45 |e| but also bidirectional charge transfer, namely, both electron accumulation (yellow color) and depletion (cyan color) on the NO molecule and Al atom. Thus, as illustrated in
Fig. 3c, the interaction nature between Al and NO can be attributed to the "acceptance-donation" of electrons analogous to that in NRR [
71,
72]. The Al atom can accept lone-pair electrons of NO and simultaneously donate electrons back to the empty
π* orbitals of NO Similarly, as shown in Figs. S3-S9 (Supporting information), such electrons transfer mechanism can also be verified in other main-group metal-embedded BC
3 systems.
Having confirmed the effective activation of NO on the eight main-group metal-embedded BC
3 systems, we then move on to examine their catalytic performances of NORR toward NH
3 synthesis. Depending on the different adsorption configuration of NO, three possible reaction pathways shown in
Fig. 4a, named as N-end, O-end, and side-on, are taken into consideration based on previous studies [
73,
74]. Noted that we only consider the N-end pathway for these systems since the N-end adsorption configuration is energetically more favorable. Subsequently, we optimize the geometric configurations of all reaction intermediates during the NORR process, and calculate the Gibbs free energy changes of each elementary step. The obtained NORR free energy diagrams for the eight main-group metal-embedded BC
3 systems are illustrated in
Fig. 4 and Fig. S10 (Supporting information), and the corresponding intermediate structures are given in Figs. S11-S18 (Supporting information). There are two possible intermediates (
i.e.,
*HNO and
*NOH) for the first hydrogenation elementary step. For all of the eight main-group metal-embedded BC
3 systems,
*HNO is preferred with respect to
*NOH. Subsequently, these systems exhibit distinct NORR paths and potential-determining steps (PDS). Taking Al@V
B, Ga@V
B, and Ge@V
C as example, the optimum NORR pathway for Al@V
B shown in
Fig. 4b can be summarized as NO(g) →
*NO →
*HNO →
*HNOH →
*H
2NOH →
*NH
2 →
*NH
3, with the PDS being the second hydrogenation step of
*HNO →
*HNOH. In terms of Ga@V
B,
Fig. 4c shows that the second hydrogenation step prefers to form
*H
2NO instead of
*HNOH, and the PDS is identified as the hydrogenation of
*HNO to
*H
2NO. While for Ge@V
C shown in
Fig. 4d, the reaction path NO(g) →
*NO →
*HNO →
*HNOH →
*NH →
*NH
2 →
*NH
3 is predicted to be the preferable one with the PDS being the hydrogenation of NO to
*HNO.
As an important index, the maximum free energy change (Δ
Gmax-NH3) along the most favorable reaction pathway, are usually used to estimate the NORR activity [
45]. Here, Δ
Gmax-NH3 < 0 is selected as screening benchmark for targeting at finding potential NORR catalysts. As shown in
Fig. 4e, the Δ
Gmax-NH3 values for Al@V
C and Sn@V
B are 0.04 and 0.10 eV, respectively. Thus, we rule out them from NORR candidates. It is quite interesting that the remaining six systems, including Al@V
B (−0.23 eV), Ga@V
B (−0.28 eV), Ga@V
C (−0.16 eV), In@V
B (−0.03 eV), Ge@V
B (−0.12 eV), and Ge@V
C (−0.25 eV), have significantly negative Δ
Gmax-NH3 values, indicating that the NORR process can proceed spontaneously on them without using any applied voltage. Furthermore, taking Al@V
B as an example, we investigate the impact of solvation correction on the free energy during the NORR process by using an implicit solvation model as implemented in VASPsol [
75]. As shown in Fig. S19 (Supporting information), the PDS and Δ
Gmax-NH3 of Al@V
B are insensitive to the solvation effects. In this regard, the solvation effects are not considered in this work.
In addition to the catalytic activity, the NH
3 selectivity is also an essential factor for achieving high catalytic performance [
76,
77]. At high NO coverage, N
2O and N
2 could be produced as main byproducts during the NORR process. To this end, we further investigate the electrochemical NO-to-N
2O/N
2 conversion following the reaction pathways illustrated in
Fig. 5a, which starts with the 2O-side and 2N-side modes of NO dimer (labelled as
*N
2O
2–O and *N2O
2–N), respectively. The free energy diagrams of N2O and N
2 formation on the remaining six systems (Al@V
B, Ga@V
B, Ga@V
C, In@V
B, Ge@V
B, and Ge@V
C) are shown in
Figs. 5b–
g, and the optimized atomic structures of intermediates are given in Figs. S20-S25 (Supporting information). One can see that converting
*N
2O
2–O to
*O + N
2O is the PDS in the NORR process toward N
2O synthesis, which demand the energy inputs of 0.83, 0.49, 0.58, 0.68, 0.47, and 0.27 eV for the Al@V
B, Ga@V
B, Ga@V
C, In@V
B, Ge@V
B, and Ge@V
C systems, respectively. On the other hand, the protonation of
*N
2O to
*N
2OH in the NORR process toward N
2 synthesis have considerably large Δ
G values of 0.17, 0.40, and 0.12 eV for the Al@V
B, Ga@V
B, and Ge@V
C systems, respectively. Noted that the NO-to-N
2 conversion on the Ga@V
C, In@V
B, and Ge@V
B systems is an exothermic process without any extra energy, indicative of the spontaneous N
2 formation. Hence, the Al@V
B, Ga@V
B, and Ge@V
C systems can effectively prevent the byproduct reactions of N
2O and N
2 formation, indicating high selectivity of NORR toward NH
3 synthesis.
Another competing reaction to the NH3 selectivity is the H2 production through HER, which normally affects the Faradaic efficiency of NORR. Therefore, we compare the competition between H2 and NH3 production in the NORR process for the remaining three candidates. As shown in Fig. S26 (Supporting information), the calculated ΔG*H values are −0.25 eV for Al@VB, −0.39 eV for Ga@VB, and −1.65 eV for Ge@VC. Obviously, the free energy barriers for the HER are considerable larger than the PDS barriers for the NORR. This indicates that the HER process can be greatly suppressed on the three systems, ensuring the excellent NORR selectivity. According to these calculations, the Al@VB, Ga@VB, and Ge@VC systems are finally screened out as potential main-group metal-based SACs for NORR toward NH3 synthesis due to the superior catalytic activity and selectivity.
In conclusion, by performing first-principles calculations, we comprehensively investigate main-group metal-embedded BC3 monolayer as SACs for the NORR toward NH3 synthesis. Our calculations reveal that nine systems (i.e., Al@VB, Al@VC, Ga@VB, Ga@VC, In@VB, Ge@VB, Ge@VC, Sn@VB, and Sb@VB) out of 18 candidates satisfy the relationship Eb - Ecoh < 0, implying the high thermodynamic stability. Besides the Sb@VB, NO molecule can be stably adsorbed and effectively activated on the remaining eight systems via the "acceptance-donation" mechanism of electron. Further scrutiny of the catalytic activity, six candidate systems including Al@VB, Ga@VB, Ga@VC, In@VB, Ge@VB, and Ge@VC are considered as excellent NORR electrocatalysts with ΔGmax-NH3 < 0, since the electrochemical NO-to-NH3 conversion can proceed spontaneously without any limiting potential. In addition, after estimating the suppression of byproducts including N2O, N2, and H2, Al@VB, Ga@VB, and Ge@VC are expected to possess high NH3 selectivity, and finally screened out as promising main-group metal-based SACs for NORR. Consequently, these theoretical findings provide valuable insights for designing efficient NORR electrocatalysts and expanding SACs to main-group metal-based systems.
PDF
Full
55
31
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
