A Theoretical Study on Electrocatalytic Nitrogen Reduction at Boron-Doped Monolayer/Bilayer Black Phosphorene Edges

Abstract

The catalytic activity of monolayer and bilayer boron-doped edge black phosphorene nanoribbons (BPNRs) as electrocatalysts for the nitrogen reduction reaction (NRR) was investigated using first-principles calculations based on density functional theory (DFT). The results indicate that boron incorporation facilitates effective N₂ adsorption at specific BPNR edges, thereby achieving superior NRR electrocatalytic performance. Through NRR screening criteria, six candidate edges (B@ZZ₃-1, B@ZZ₄-1, B@AC₀-1, B@ZZ₀^AA-1, B@ZZ₁^AB-3, and B@ZZ₄^AA-3) were identified. Electronic property analysis revealed that boron doping significantly reduces the bandgap of BPNRs and enhances catalytic activity by promoting electron accumulation at boron sites. Free energy pathway calculations demonstrated that B@AC₀-1, B@ZZ₀^AA-1, and B@ZZ₁^AB-3 exhibit overpotentials of 0.19 V, 0.28 V, and 0.15 V, respectively, during the NRR process, outperforming other phosphorus-based catalysts in activity.

1. Introduction

Ammonia (NH₃) is one of the most produced chemical raw materials globally, extensively used in fertilizer production, industrial refrigeration, and clean energy storage. Over 180 million tons of ammonia are synthesized annually worldwide via the Haber–Bosch process from nitrogen gas, supporting approximately 50% of global food production. However, this conventional process requires operation at 400–600 °C and a high pressure of 15–25 MPa, consumes 1%–2% of global energy, emits around 450 million tons of carbon dioxide (3% of global carbon emissions), and faces thermodynamic limitations, achieving only 10%–15% single-pass nitrogen conversion efficiency, with multi-stage cycling improving this to 20%–30% [1,2,3,4,5,6,7,8,9]. Under the “dual-carbon” strategy, developing novel low-energy-consumption sustainable ammonia synthesis technologies has emerged as a major challenge. In agricultural production and related fields, converting N₂ into utilizable nitrogen-containing compounds could compensate for N fertilizer losses in agricultural systems, regulate soil nitrogen balance, and address issues such as restricted crop growth caused by nitrogen fertilizer losses [10,11]. In industrial production, the electrocatalytic nitrogen reduction reaction (NRR), which enables N₂ reduction under ambient temperature and pressure using renewable energy, is regarded as a critical pathway to overcoming the bottlenecks of traditional processes [12,13,14,15,16,17,18]. However, due to the high dissociation energy of the N≡N triple bond and competitive hydrogen evolution reactions (HERs), practical catalytic systems suffer from high overpotentials, low Faradaic efficiency, and other challenges. Mainstream noble metal catalysts are scarce and costly, while emerging catalysts face issues such as insufficient exposure of active sites and difficult desorption of intermediates [19,20,21,22,23,24,25]. Consequently, two-dimensional material systems have become a research hotspot.

Black phosphorene (BP), as a novel two-dimensional semiconductor material, exhibits exceptional carrier mobility and tunable bandgap owing to its unique wrinkled honeycomb structure [26,27,28,29,30,31,32,33]. BP is prepared using mechanical stripping, plasma etching, liquid-phase exfoliation, chemical vapor deposition, and wet chemical methods. Applicability of BP is increased with suitable surface modifications [34]. For example, Sn-doped black phosphorene has been experimentally synthesized by liquid-phase exfoliation of bulk BP, followed by Sn⁴⁺ ion adsorption, chemical reduction, and thermal annealing under inert atmosphere, yielding stable Sn-P coordination [35]. Additional studies have indicated that the intrinsic catalytic activity of pristine two-dimensional BP originates from local electron-enriched regions formed by its wrinkled structure. However, due to weak N₂ adsorption energy and high hydrogenation energy barriers, its NRR performance has not yet reached ideal levels. Wang et al. fabricated MnO₂-supported BP quantum dot hybrids with optimized interfacial interactions, significantly improving nitrogen reduction reaction (NRR) performance [36]. The research of Lai et al. also indicates that BP has the NRR ability [37]. Wang’s research group reported for the first time that the well-exfoliated few-layer BP nanosheets are a new type of metal-free NRR electrocatalyst [38]. Pei et al. significantly enhanced the NRR electrocatalytic performance of BP by constructing phosphorus vacancy defects on its surface via plasma etching technology [39]. Extensive experimental and theoretical studies demonstrate the viability of phosphorene as an efficient NRR catalyst. However, defective black phosphorene still exhibits weak adsorption of N₂ molecules and high subsequent protonation overpotential, hindering commercial applications. Studies have shown that doping boron (B) atoms can effectively break through this bottleneck. B atoms form strong σ bonds with N₂’s lone pair of electrons through their empty p orbitals, weakening the strength of the N≡N bond [40,41]. Additionally, the boundary regions of monolayer and bilayer BP, characterized by unsaturated dangling bonds and strain field effects, enable synergistic enhancement of lateral adsorption of N₂ molecules and charge transfer efficiency. This “doping-boundary” synergistic regulation strategy provides novel insights for designing NRR catalysts with high activity and stability.

This study focuses on the NRR catalytic mechanisms of boron-doped monolayer/bilayer black phosphorene boundary systems, employing first-principles calculations to investigate the structure–activity relationships among B-doping sites, boundary configurations, and electronic structures. By constructing black phosphorene nanoribbon (BPNR) boundaries with boron-doped edges, we explore the NRR catalytic performance of black phosphorene with various boron-doped edge sites, analyze the electrocatalytic nitrogen reduction mechanisms of monolayer and bilayer BPNR edges at distinct sites, and reveal the catalytic performance advantages of specific boron-doped black phosphorene boundary sites for NRR.

2. Methods and Computational Details

This study employed first-principles density functional theory (DFT) calculations, with structural optimizations performed using the VASP.5.2 software package. The projector-augmented wave (PAW) pseudopotentials were utilized to treat ion–electron interactions, the Perdew–Burke–Ernzerhof (PBE) functional was adopted for exchange-correlation energy, and the DFT-D3 method was incorporated to correct interlayer van der Waals (vdW) interactions in bilayer systems. The calculations employed a 2 × 1 × 1 k-point grid with convergence criteria of 1 × 10⁻³ eV for energy and 0.02 eV/Å for forces. The numerical uncertainty in adsorption energies and ΔG values is estimated to be within ±0.001 eV. The linear combination of atomic orbitals (LCAO) method was employed with the “Medium” basis set provided by Quantum ATK for the basis expansion, and the hybrid Heyd–Scuseria–Ernzerhof (HSE06) functional was employed for the accurate band structure calculations. Hydrogen-passivated monolayer BPNRs were used to construct the boundary catalytic systems, preventing interlayer closure phenomena caused by structural relaxation in multilayer configurations [42]. Based on our previous research, we first defined three typical pristine monolayer edge configurations: zigzag (ZZ₀), armchair (AC₀), and 54°-tilted diagonal (SD_54°) edges. Starting from ZZ₀ and AC₀ edges, we constructed five ZZ-type (ZZ_n, n = 1, 2, 3, 4, 5) and two AC-type (AC_n, n = 1, 2) reconstructed edges for monolayer boundaries. Bilayer BPNRs were constructed based on these monolayer boundaries through four stacking modes (AA/AB/AC/AD) [43]. In the side view, the AA and AB stacking modes correspond to the alignment of the same sides (e.g., top–top or bottom–bottom) of the two layers, while the AC and AD stacking modes correspond to the alignment of opposite sides (top–bottom). In the top view, the AA and AC configurations exhibit perfect vertical alignment between the two layers, whereas the AB and AD configurations are laterally staggered. For instance, the AA-stacked ZZ₀ bilayer edge is denoted as ZZ₀^AB. Other bilayer edge notations follow this convention. The lattice parameters for the three edge configurations along the periodic direction (a-axis) are ZZ edge 13.24 Å, AC edge 17.52 Å, and SD edge 10.98 Å. The width (b-axis) and vertical (c-axis) directions are uniformly set to 30 Å. Spin polarization was not considered because the whole system contains non-transition metal elements with negligible magnetic properties. Therefore, ignoring the spin polarization can help reduce computational cost without affecting the final conclusions of this work.

To compare the binding strength of B atoms doped into different BPNR edges, this study introduced the definition of binding energy ($E_{Binding}$):

$$E_{Binding}=E_{doped}-E_{BPNR}-E_B$$

(1)

where $E_{doped}$ is the energy of the B-doped BPNR, and $E_B$ is the energy of an isolated B atom.

The adsorption strength of N₂ molecules and H atoms on catalytic sites is quantified by adsorption energy ($E_{ads}$):

$$E_{ads}=E_{total}-E_{doped}-E_{Adsorbate}$$

(2)

where $E_{total}$ represents the energy of the boron-doped BPNR after adsorbing an N₂ molecule or H atom, and $E_{Adsorbate}$ denotes the energy of the N₂ molecule or H atom.

Gibbs free energy ($\Delta G$) is used to describe the energy changes in the nitrogen reduction reaction [44]:

$$\Delta G=\Delta E+\Delta E_{ZPE}-T\Delta S+\Delta G_{pH}$$

(3)

where $\Delta E$ denotes the total energy difference between the pre- and post-reaction systems, and $T$ represents the temperature of the catalytic reaction environment; $\Delta G_{pH}$ is the free energy correction term under varying pH values. In this study, $T$ and pH are set to 293.15 K and 0, respectively; $\Delta E_{ZPE }$ and $\Delta S$ correspond to the zero-point energy and entropy change of the catalytic reaction, respectively, with their formulas expressed as follows [45]:

$$E_{ZPE}={\displaystyle \frac{1}{2}}{{\displaystyle \sum }}_i\mathrm{ћ}{\gamma }_i$$

(4)

$$-TS=k_{B}T{{\displaystyle \sum }}_i\ln \left(1-e^{{\displaystyle \frac{\mathrm{ћ}{\gamma }_i}{k_{B}T}}}\right)-{{\displaystyle \sum }}_i\mathrm{ћ}{V}_i\left({\displaystyle \frac{1}{e^{\frac{\mathrm{ћ}{\gamma }_i}{k_B}}-1}}\right)$$

(5)

where $\gamma$, ћ, and k_B represent vibrational frequency, the reduced Planck constant, and the Boltzmann constant, respectively.

The overpotential ($\eta$) is defined as follows:

$$\eta =U_0-U_L$$

(6)

where $U_0$ is the equilibrium potential (approximately −0.16 eV for the electrocatalytic NRR process [44,45]), and $U_L$ denotes the limiting potential:

$$U_L=-{\displaystyle \frac{\Delta G_{max}}{e}}$$

(7)

3. Results and Discussion

3.1. Boron-Doped BPNR Edge Site Selection and Their Investigation as NRR Catalysts

First, based on our previous screening of HER catalytic active sites across numerous monolayer and bilayer edge configurations [43], we doped B atoms into these HER-active sites with excellent catalytic properties to explore their NRR catalytic performance. As shown in Figure 1, eight types of monolayer edges (ZZ, AC, and SD_54°) and ten types of bilayer edges were considered. For each edge configuration, distinct catalytic sites were examined, such as site 1 on the ZZ₁ monolayer BPNR edge (denoted as ZZ₁-1), and site 1 on the ZZ₁^AA bilayer BPNR edge (denoted as ZZ₁^AA-1). Similar notation applies to other edge sites in the figures. The binding energies of B atoms in monolayer and bilayer BPNR edges are listed in

Table 1. Binding energy of boron atoms at different sites of the monolayer BPNR edges.

Catalyst	EBPNR (eV)	Edoped (eV)	EB (eV)	EBinding (eV)
B@ZZ0-1	−277.57	−277.16	−6.19	−0.17
B@ZZ1-1	−300.47	−300.49	−6.19	−0.60
B@ZZ2-1	−299.62	−300.03	−6.19	−0.18
B@ZZ3-1	−298.46	−300.53	−6.19	−2.66
B@ZZ4-1	−322.82	−322.68	−6.19	−0.44
B@AC0-1	−333.91	−334.62	−6.19	−1.29
B@AC1-1	−333.91	-	−6.19	-
B@SD54°-1	−158.13	−158.01	−6.19	−0.46

and

Table 2. Binding energy of boron atoms at different sites of the bilayer BPNR edges.

Catalyst	EBPNR (eV)	Edoped (eV)	EB (eV)	EBinding (eV)
B@ZZ0AA-1	−560.75	−561.25	−6.19	−1.08
B@ZZ0AB-2	−561.88	−561.97	−6.19	−0.67
B@ZZ1AA-4	−606.26	−606.27	−6.19	−0.60
B@ZZ1AB-3	−605.69	−605.71	−6.19	−0.60
B@ZZ3AC-1	−608.25	−609.51	−6.19	−1.84
B@ZZ3AC-2	−608.25	−610.53	−6.19	−2.86
B@ZZ3AD-1	−609.73	−610.36	−6.19	−1.22
B@ZZ3AD-2	−609.73	−610.53	−6.19	−1.39
B@ZZ4AA-3	−650.93	−650.91	−6.19	−0.56
B@ZZ4AD-3	−650.45	−650.45	−6.19	−0.58
B@AC0AB-10	−674.50	−675.69	−6.19	−1.78
B@SD54°AB-1	−323.57	-	−6.19	-

, respectively. In monolayer BPNR edges, calculations revealed that the B binding energy at the AC₁-1 edge site was positive, indicating unstable bonding, and thus this site was excluded. For other sites, binding energies ranged from −2.66 eV to −0.17 eV, with B@ZZ₃-1 (−2.66 eV) and B@AC₀-1 (−1.29 eV) exhibiting strong bonding characteristics. In bilayer BPNR edges, except for the B@SD_54°AB-1 site where B atoms failed to form stable structures with neighboring P atoms, most sites showed binding energies between −0.56 eV and −2.86 eV, meeting stability requirements.

The efficient progression of the NRR relies on the catalyst’s capability to activate the N≡N bond. The proton-coupled electron transfer (PCET) mechanism further weakens the triple bond strength through synergistic proton–electron transfer, as reported in prior studies [44]. Based on this, we first evaluated the N₂ adsorption performance of B@BPNR. Figure 2 demonstrates that N₂ molecules adopt stable end-on adsorption configurations on most B@BPNR sites. In monolayer structures, B@ZZ₀-1 (−0.07 eV) and B@ZZ₁-1 (−0.11 eV) showed weak N₂ adsorption capability. Similarly, in bilayer BPNR edges, B@ZZ₃^AC-2 (1.55 eV) and B@AC₀^AD-10 (0.44 eV) displayed strong repulsion toward N₂, and thus these configurations were excluded from further analysis. To verify the dominance of the NRR pathway, we further analyzed the competitive relationship between hydrogen adsorption free energy (ΔG_*H) and NRR energy barriers. As shown in Figure 3, among monolayer catalysts, B@ZZ₃-1, B@ZZ₄-1, and B@AC₀-1 exhibited stronger preference for NRR. For bilayer catalysts, NRR dominated the reaction in B@ZZ₀^AA-1, B@ZZ₁^AB-3, and B@ZZ₄^AA-3. Consequently, these six configurations were selected for subsequent reaction pathway investigations.

Furthermore, superior electrical conductivity is critical for electrocatalytic reactions, but the intrinsic wide bandgap (~1.5 eV) of black phosphorene limits its charge transport capacity, leading to high NRR energy barriers. B-doping significantly optimizes the system’s conductivity. As shown in Figure 4 and Figure 5, total density of states (TDOS) analysis reveals narrowed bandgaps after doping, enhancing carrier mobility. Additional charge density difference (CDD) studies demonstrate that B atoms act as electron acceptors, inducing P→B charge transfer and forming strong covalent P-B bonds.

3.2. Reaction Pathways of N₂ on Boron-Doped Black Phosphorene Edges

The six doped edge configurations investigated in this study only stabilize N₂ adsorption via end-on configurations, thereby confining the NRR to distal or alternating mechanisms. The potential-determining step (PDS), representing the hydrogenation step with the highest free energy in the reaction pathway, determines the reaction efficiency, where lower PDS barriers favor NRR progression. This study primarily focuses on the thermodynamic properties of NRR intermediates on B-doped black phosphorene, assuming the presence of readily available active hydrogen species in the reaction environment to simplify the modeling of the reaction pathway. Figure 6 presents the NRR energy pathways and corresponding intermediate structures for N₂ on monolayer B-doped edges (B@ZZ₃-1, B@ZZ₄-1, and B@AC₀-1). Their PDS free energy barriers are 0.66 eV (B@ZZ₃-1, distal), 0.66 eV (B@ZZ₃-1, alternating), 1.18 eV (B@ZZ₄-1, distal), 1.18 eV (B@ZZ₄-1, alternating), 0.35 eV (B@AC₀-1, distal), and 0.52 eV (B@AC₀-1, alternating). For B@AC₀-1, N₂ preferentially undergoes NRR via the distal mechanism. In contrast, B@ZZ₃-1 and B@ZZ₄-1 exhibit equal PDS barriers for both mechanisms.

As shown in Figure 6a,b, for B@ZZ₃-1 via the distal mechanism, the hydrogenation from *N-NH₂ to NH₃ desorption (*N-NH₂→*N + NH₃(g)) requires overcoming an uphill free energy of 0.48 eV. In the alternating pathway, the second-highest barrier (0.59 eV) occurs during *NH-NH₂→*NH₂ + NH₂. For B@ZZ₄-1 (Figure 6c,d), under the alternating mechanism, the hydrogenation from *N-NH to *NH₃ becomes exothermic and spontaneous after overcoming the 1.18 eV PDS barrier. In the distal mechanism, the *N-NH→*N-NH₂ step shows a minor barrier of 0.12 eV. For B@AC₀-1’s distal pathway (Figure 6e), the hydrogenation from *N-N proceeds with low barriers of 0.35 eV (*N-N→*N-NH) and 0.34 eV (*N→*NH).

The NRR pathways and reaction structures for bilayer boron-doped BPNR edges are shown in Figure 7. For B@ZZ₀^AA-1 and B@ZZ₄^AA-3, when N₂ undergoes the second hydrogenation step (*N-NH→*NH-NH) via the alternating mechanism, structural relaxation causes the H atom initially bonded to the inner N atom to shift to the outer N atom, reverting to the distal reaction pathway. Thus, only the distal reaction energy pathways were calculated for these two catalyst sites. For B@ZZ₀^AA-1 (Figure 7a), its PDS corresponds to the first hydrogenation step of N₂ (*N-N→*N-NH, 0.44 eV). Subsequent hydrogenations, except for *N-NH₂→*N-NH₃ (requiring a minor barrier of 0.04 eV), proceed exothermically and spontaneously. After forming *NH₂, further hydrogenation of the N atom requires overcoming a 0.25 eV barrier. For B@ZZ₁^AB-3 (Figure 7b,c), both distal and alternating mechanisms exhibit an identical PDS at *NH₂→*NH₃ (0.31 eV), with equivalent barriers for all other steps, indicating coexistence of both mechanisms in NRR. Notably, B@ZZ₁^AB-3 demonstrates high catalytic activity, requiring only 0.08 eV (*N-N→*N-NH) and 0.31 eV (*NH₂→*NH₃) barriers, while other hydrogenation/desorption steps proceed spontaneously. For B@ZZ₄^AA-3 (Figure 7d), NRR follows the distal mechanism with a PDS at *NH₂→*NH₃ (0.80 eV). After N₂ adsorption, the initial hydrogenations require barriers of 0.25 eV (*N-N→*N-NH) and 0.07 eV (*N-NH→*N-NH₂). Subsequent steps from *N-NH₂ to *NH₂ are exothermic and spontaneous. Finally, NH₃ is generated by overcoming the 0.80 eV barrier at *NH₂→*NH₃. It should be noted that this work only focuses on the thermodynamic energy pathways of the NRR with no kinetic mechanisms calculated here.

Effective NRR catalysts require low reaction overpotentials.

Table 3. Reaction mechanisms, PDS, and overpotential (η) of NRR catalysts in this work and other studies.

System	Pathway	PDS	η (V)
B@ZZ3-1	Distal	*N-N→*N-NH	0.50
	Alternating	*N-N→*N-NH	0.50
B@ZZ4-1	Distal	*N-N→*N-NH	1.02
	Alternating	*N-N→*N-NH	1.02
B@AC0-1	Distal	*N-N→*N-NH	0.19
	Alternating	*NH-NH2→*NH2-NH2	0.36
B@ZZ0AA-1	Distal	*N-N→*N-NH	0.28
B@ZZ1AB-3	Distal	*NH2→*NH3	0.15
	Alternating	*NH2→*NH3	0.15
B@ZZ4AA-3	Distal	*NH2→*NH3	0.64
Ru@P [46]	Enzymatic	*N-NH→*NH-NH	0.70
DV-(5|8|5)@BP [39]	Alternating	*NH-NH2→*NH2-NH2	0.51
W@BP [47]	Alternating	*NH-NH2→*NH2-NH2	0.30

lists the characteristics of the six catalysts investigated in this chapter, along with comparisons to previous studies based on black phosphorene. The DV-(5|8|5)@BP system achieved an NRR overpotential of 0.51 V by utilizing internal defects in black phosphorene to create coordinatively unsaturated adsorption sites and enhance N₂ adsorption. In contrast, Ru@P and W@BP improved NRR electrocatalytic performance through metal-atom-mediated charge distribution modulation, achieving overpotentials of 0.70 V and 0.30 V, respectively. In comparison, B@ZZ₃-1 (0.50 V) and B@ZZ₄^AA-3 (0.64 V) in this study exhibit competitive overpotentials. Notably, B@AC₀-1 (0.19 V), B@ZZ₀^AA-1 (0.28 V), and B@ZZ₁^AB-3 (0.15 V) demonstrate superior NRR electrocatalytic performance with even lower overpotentials than W@BP.

4. Conclusions

Using first-principles methods based on DFT, we investigated the catalytic activity of boron-doped edge BPNRs as electrocatalysts for the NRR. Partially B-doped BPNR edges efficiently adsorb N₂ molecules. By constructing boron-doped BPNR edge models and screening based on N₂ adsorption and NRR competition criteria, six candidate edges were selected: B@ZZ₃-1, B@ZZ₄-1, B@AC₀-1, B@ZZ₀^AA-1, B@ZZ₁^AB-3, and B@ZZ₄^AA-3. Electronic structure analysis revealed that boron doping significantly reduces the bandgap of BPNRs and induces electron accumulation at the boron sites, thereby enhancing their catalytic performance. NRR free energy pathway calculations demonstrated that B@AC₀-1, B@ZZ₀^AA-1, and B@ZZ₁^AB-3 exhibit overpotentials of 0.19 V, 0.28 V, and 0.15 V during NRR, respectively, showing higher catalytic activity compared to other phosphorus-based catalysts. These findings provide a theoretical foundation for designing highly active and stable NRR catalysts.