First Principles Study of Double Boron Atoms Supported on Graphitic Carbon Nitride (g-C₃N₄) for Nitrogen Electroreduction

Abstract

Electrocatalytic reduction of N₂ provides a clean, sustainable way for NH₃ production. Efficient catalysts thus play a key role but remain a long-term challenge. In this study, the catalytic activity of double boron supported on graphitic carbon nitride (g-C₃N₄) for a N₂ reduction reaction (NRR) is explored by density functional theory (DFT) calculations. Our results show that double boron atoms embedded in g-C₃N₄ with coordination of four N atoms and two boron atoms exhibits an excellent NRR performance with negligible energy consumption for adding hydrogen to *N₂, while a moderate $\Delta G$ of 0.58 eV for the formation of the second NH₃ suggests this catalyst is a potential candidate for N₂ fixation.

1. Introduction

Ammonia (NH₃) is not only an essential chemical for maintaining body function, but also can be used as a source of fertilizer and renewable energy for modern industry and agriculture [1,2,3]. The conventional Haber–Bosch process is the most commonly used method for direct ammonia synthesis in industry [4,5,6]. In this process, in addition to the large energy consumption, this traditional synthetic method also produces a large number of harmful gases [7,8]. Therefore, a green and sustainable way of producing NH₃ is necessary. According to the principle of nitrogen fixation, $\left({\mathrm{N}}_2(\mathrm{g}\right)+6{\mathrm{H}}^{+}+6{\mathrm{e}}^{-}\to 2{\mathrm{NH}}_3)$, in bacteria, electrochemical nitrogen reduction reactions can be operated under ambient conditions, which is considered to be a very attractive method as it is a carbon-free energy carrier and energy-saving [9,10,11,12,13,14,15,16,17,18]. However, the inertness of N₂ creates an urgent need for a high activity and highly selective catalyst to stimulate the activity of N₂ to achieve highly efficient NRR, which is still a goal pursued by chemists [19].

Compared to homogeneous electrocatalysts, heterogeneous electrocatalysts are able to be separated from the reactants, thus facilitating the recycling/reutilization of the catalysts. Many two-dimensional (2D) materials are successfully used as catalyst substrates for metal and non-metal atoms due to their unique anisotropy and electronic properties. In two-dimensional materials, the single layer g-C₃N₄ is an efficient catalyst because of its large surface area and plentiful active sites with respect to the bulk g-C₃N₄ [20]. In single layer g-C₃N₄, well-distributed and well-sized holes provide an abundance of unpaired electrons, which can firmly stabilize the double-atom catalyst (DAC) without diffusion and aggregation [21,22,23]. In the last few years, single-atom catalysts (SACs) concerned in heterogeneous catalysis and have been applied to NRR. It is noteworthy that several studies have demonstrated that double-atom catalysts (DACs) have a more efficient catalytic performance when compared with their SAC counterparts [24]. The advantages of double-atom catalysts (DACs) overcome the limitations of SACs in that they mostly offer single-molecule adsorption sites; DACs provide multiple active sites which can change the adsorption strength between the catalyst and the adsorbed gas molecules to a greater extent, especially allowing the interactions between the adsorbed molecules and catalyst to produce N-B bonds. The strong chemical interaction between adjacent atoms stabilizes the diatomic structure, thereby forming a highly stable active center. The catalytic activity of a series of single metal atoms doped on g-C₃N₄ has been explored previously; g-C₃N₄ has a strong interaction with metal atoms (TM) and induces the accumulation of polarization charge on the surface of metal atoms [25,26,27,28,29]. Co, Ti, Mo, W, and Pt atoms were doped into g-C₃N₄, respectively, and the resulting five SACs showed good catalytic performance with an extremely low onset potential for NRR. Studies have predicted that B atoms have good catalytic activity, providing a new direction in the development of NRR. The boron atom with sp3 or sp2 electron arrangement offers empty orbitals to activate N₂ [30]. In this work, using first-principles theoretical simulations, we have investigated the feasibility of double B atoms embedded in a g-C₃N₄ monolayer (B2@C₃N₄) as an NRR electrocatalyst. Our calculations show that B2@C₃N₄ has good NRR catalytic activity with a relatively low initial potential and NH₃ desorption free energy, effectively suppressing the effect of the hydrogen evolution reaction (HER) [31]. These findings indicate that B2@C₃N₄ can be used as a promising electrochemical catalyst for the production of NH₃.

2. Materials and Method

First-principles calculations were performed using the Vienna Ab initio Simulation Package (VASP) [32,33,34], in which the electron–core interactions were described by the projector augmented wave (PAW) potential [35]. The Perdew–Burke–Ernzerhof functional (PBE) was used to describe the exchange–correlation interactions [36]. The plane-wave energy cutoff was set to 450 eV and the convergence of geometry relaxation was less than 0.02 eV/Å for the force on each atom. The occupancy of the one-electron state was calculated using Gaussian smearing with SIGMA = 0.02 eV. The electron transfer in each atom was studied by Bader charge analysis. Brillouin zones were sampled with a 2 × 2 × 1 k-point grid. The total energies were −485.256 and −485.259 eV generated with 2 × 2 × 1 and 3 × 3 × 1, respectively. The difference is negligible; therefore, we used 2 × 2 × 1 k-mesh for further calculations.

The g-C₃N₄ model was constructed in a 2 × 2 × 1 supercell, as shown in Figure 1a. A vacuum layer was set at 15 Å in the z-direction to avoid the interactions between two periodic units. Note that g-C₃N₄ is stable in both forms of planar and corrugated structures [37,38]. The planar (non-corrugated) form was used in this work; furthermore, the effect of the number of layers of g-C₃N₄ was not considered. The adsorption energy ($E_{\mathrm{ads}}$) was calculated as follows

$$E_{\mathrm{ads}}=E_{\mathrm{mol}+\mathrm{B}2@{\mathrm{C}}_3{\mathrm{N}}_4}-E_{\mathrm{B}2@{\mathrm{C}}_3{\mathrm{N}}_4}-E_{\mathrm{mol}}$$

(1)

where $E_{\mathrm{mol}+\mathrm{B}2@{\mathrm{C}}_3{\mathrm{N}}_4}$ represents the total energy of substrate and adsorbate, $E_{\mathrm{B}2@{\mathrm{C}}_3{\mathrm{N}}_4}$ is the the energy of the catalyst substrate, and $E_{\mathrm{mol}}$ represents the total energy of the isolated NRR intermediate (N₂ molecules or N_xH_y intermediates). The computational hydrogen electrode (CHE) model was used to calculate the Gibbs free energy change ($\Delta G$) for each hydrogenation step [39], which can be computed by

$$\Delta G=\Delta E_{\mathrm{DFT}}+\Delta E_{\mathrm{ZPE}}-T\Delta S+\Delta G_{\mathrm{U}}+\Delta G_{\mathrm{pH}}$$

(2)

where $\Delta E_{\mathrm{ZPE}}$ and $\Delta S$ are the zero-point energy correction and the change of entropy, respectively, which were calculated from vibrational frequencies. $T$ is the temperature, which is set to 298 K, $\Delta G_{\mathrm{U}}$ is the free energy change with respect to the applied electrode potential $U$, and $\Delta G_{\mathrm{pH}}$ is the free energy correction of the ${\mathrm{H}}^{+}$ which is expressed as $\Delta G_{\mathrm{pH}}=2.303\times k_{\mathrm{B}}T\times \mathrm{pH}$, where the value of $\mathrm{pH}$ is taken as zero. In addition, the free energy corrections of gases NH₃, H₂, and N₂ were taken from the NIST database. As an indicator of catalytic activity, the onset potential is determined by the potential-limiting step and is calculated by

$$U_{\mathrm{onset}}=-\Delta G_{\max }/e$$

(3)

3. Results and Discussion

Two B atoms were added to the six-fold cavity sites of g-C₃N₄ to form the NRR electrocatalyst, as shown in Figure 1b. Upon geometric optimization, the two B atoms (B2) were still located in the porous sites and g-C₃N₄ experienced negligible distortions, suggesting that B2@C₃N₄ is a stable catalyst system. The binding energy was −13.53 eV and the B-N bonds had a length of 1.48~1.50 Å.

The adsorption configuration and electronic structures of the adsorbed N₂ molecules have an obvious impact on the degree of activation of N₂, including the electron transfer, N-N bond length, adsorption energy ($E_{\mathrm{ads}}$) of N₂, and the hydrogenation of N₂. All of the possible models of N₂ adsorption on the catalyst were considered in our calculations. The calculated results suggest that the side-on model is favored for N₂ adsorption on the B2@C₃N₄ catalyst, as shown in Figure 2. However, the end-on model is unstable for N₂ adsorption on B2@C₃N₄, since N₂ will escape from the initial position upon relaxation. It is remarkable that the length of the N-N bond increases from 1.17 Å in the gas phase to 1.25 Å for *N₂ adsorption on B2@C₃N₄ and produces an energy of adsorption Gibbs free energy of 0.26 eV. Bader charge analysis indicates that the *N₂ gains 0.54e electrons from B2@C₃N₄, suggesting that the *N₂ molecule is effectively activated [40]. The interaction between the catalyst and the active N₂ molecule was further investigated by the electron localization function (ELF) and charge density difference (CDD) analysis. In Figure 3a, the yellow color in the CDD shows that the electrons are transferred from the catalyst to the adsorbed *N₂ molecule, and the accumulated charge promotes the dissociation of *N₂. In Figure 3b, ELF indicates that the red area is mainly concentrated between the *N and B atoms, suggesting that a strong chemical bond has been formed between the B atom and *N atoms, which is beneficial for *N₂ reduction. Furthermore, the partial density of states (PDOS) of the adsorbed *N₂ on B2@C₃N₄ shown in Figure 4 suggests that an obvious hybridization between the B-2s2 orbitals and N-2p3 orbitals of the adsorbed *N₂ occurs above the B2@C₃N₄.

N₂ reduction on the catalyst is a six-electron related reaction, i.e., ${\mathrm{N}}_2+6{\mathrm{H}}^{+}+6{\mathrm{e}}^{-}\to 2{\mathrm{NH}}_3$. Generally, the reaction may involve three possible reaction mechanisms, including distal, alternating, and enzymatic pathways. We did not discuss the distal or alternating pathways because neither of them occur via an end-on *N₂ adsorption configuration. However, *N₂ prefers to be adsorbed in the side-on configuration; therefore, the enzymatic mechanism (Figure 5a) was considered for the investigation of the NRR process occurring on B2@C₃N₄ [41]. The configurations of the key hydrogenation steps were shown in Figure 5b. Interestingly, the free energy profile (Figure 6) shows that the hydrogenation steps are downhill until the first NH₃ formation, indicating that these reactions can feasibly take place without the external electrical potential. According to previous studies, addition of the first hydrogen atom to *N₂ (${*\mathrm{N}}_2+{\mathrm{H}}^{+}+{\mathrm{e}}^{-}\to *\mathrm{NNH}$) is usually the potential-determining step with a large positive Gibbs free energy change ($\Delta G$) in the NRR process found in some of other electrocatalysts. This is because N₂ is a very stable molecule; it takes a lot of energy to break the N≡N bond. In this work, the first hydrogenation step is downhill with a $\Delta G$ of −0.22 eV to form *NNH. The formation of *NHNH via addition of the second hydrogen atom to the other *N atom results in the $\Delta G$ decreasing by 0.52 eV. Addition of the third hydrogen atom to form *NHNH₂ needs to overcome a small $\Delta G$ of 0.02 eV. Sequentially, the formation of *NH₂NH₂ and *NH₂NH₃ leads to the values of $\Delta G$ decreasing by 0.25 and 2.00 eV, respectively. Desorption of the first NH₃ molecule leads to a positive $\Delta G$ of 0.50 eV. Addition of the last hydrogen atom to form the second NH₃ is the potential-limiting step with the largest positive $\Delta G$ value (0.58 eV) among all the hydrogenation steps. In order to study the influence of the applied electric potential on the NRR process, the energy profiles under U = −0.58 V were calculated and are shown in Figure 6. The results show that the external electrical potential can decrease the uphill energy changes, and thus promote the NRR process. Furthermore, desorption of the second NH₃ molecule needs to overcome a moderate $\Delta G_{\mathrm{des}}$ of 1.15 eV, the NH₃ molecule is easily detached from the surface of the catalyst suggesting that B2@C₃N₄ possesses high-efficiency catalytic activity to reduce N₂ to NH₃ [42].

The selectivity of catalysts between NRR and the competing HER was investigated. According to the hydrogen evolution reaction $\left({\mathrm{H}}^{+}+{\mathrm{e}}^{-}\to 1/2{\mathrm{H}}_2\right)$, we evaluated the selectivity of the NRR and HER by comparing their potential-limiting step, as shown in Figure 7. The hydrogen adsorption Gibbs energy change is −2.36 eV, and $\Delta G$ (*N₂) is 1.78 eV lower than $\Delta G$ (*H), suggesting that the hydrogen atom would be firmly captured by the catalyst and serve as the hydrogen source for NRR. Thus, HER is takes place with difficulty on B2@C₃N₄.

4. Conclusions

In summary, by performing DFT computations, we explored the catalytic activity of double B atoms anchored on a g-C₃N₄ monolayer toward the electroreduction of N₂ to NH₃. Our calculations indicate that the double B atoms remain in the porous position of the g-C₃N₄ framework to form the B2@C₃N₄ system. The structure does not experience remarkable structural distortion, and thus maintains its excellent properties. N₂ prefers to be adsorbed on the catalyst in the side-on configuration, yielding a large adsorption Gibbs free energy and strong hybridization between the N₂-2p orbitals and B-sp3 orbitals. Subsequently, NRR takes place on B2@C₃N₄ with the enzymatic mechanism_. The Gibbs free energy change profile shows that B2@C₃N₄ possesses a high catalytic activity with negligible energy consumption to add hydrogen to *N₂ and possesses a moderate $\Delta G$ of 0.58 eV for the formation of the second NH₃. The catalyst also exhibits a high selectivity for NRR because of the much larger value of $\Delta G$ for HER. Our calculations suggest that the B2@C₃N₄ system might be a promising metal-free double-atom NRR electrocatalyst.