Theoretical Evaluation of Electrochemical Nitrate Reduction Reaction on Graphdiyne-Supported Transition Metal Single-Atom Catalysts.

The electrochemical reaction can be applied as a powerful method to eliminate the pollution of nitrate (NO3 -) and as a feasible synthesis to enable the conversion of nitrate into ammonia (NH3) at room temperature. Herein, density functional theory calculations are applied to comprehensively analyze the electrochemical nitrate reduction reaction (NO3RR) on graphdiyne-supported transition metal single-atom catalysts (TM@GDY SACs) for the first time. It can be found that the vanadium-anchored graphdiyne (V@GDY) displays the lowest limiting potential of -0.63 V versus a reversible hydrogen electrode among the investigated systems in this work. Notably, the competing hydrogen evolution reaction is relatively restrained due to the comparatively weak adsorption of the H proton on the TM@GDY SACs. Moreover, higher energy intake is needed to overcome the energy barrier during the formation of byproducts (NO2, NO, N2O, and N2) on V@GDY without applying extra electrode potential, showing the selectivity of NH3 in the NO3RR process. The ab initio molecular dynamics simulation denotes that the V@GDY possesses excellent structure stability at the temperature of 600 K without much distortion, compared with the initial shape, indicating the promise for synthesis. This study not only offers a feasible NO3RR electrocatalyst but also paves the way for the development of the NO3RR process.

Introduction

Ammonia (NH3) is a significantly important raw material in modern industrial production, which has been widely used in the chemical industry, light industry, chemical fertilizers, pharmaceutical, synthetic fiber, and other fields.[1,2] As early as the early 20th century, the emergence of the Haber-Bosch ammonia synthesis method solved the problem of the great demand for ammonia.[3,4] Nitrogen (N2) and hydrogen (H2) are the raw materials to synthesize the product NH3 under extreme conditions (400–500 °C, 200–300 atm), which are both significantly energy-intensive and environment-malignant, responsible for ∼2% global annual energy consumption and ∼3% CO2 global emission.[4−8] It seems to go against the concept of sustainable development from the view of environmental protection, so scientists are looking forward to finding new ammonia synthesis methods that can replace the Haber-Bosch method and meet the needs of industries.[9] Over the past few decades, the electrochemical nitrogen reduction reaction (NRR) seems to be a feasible method to produce NH3 by reducing N2.[10] Nevertheless, a lot of external energy is required to break the triple bond (941 kJ/mol) in the N2 molecule to improve its reaction activity.[9,11] With the efforts of researchers and scientists, different kinds of efficient NRR electrocatalysts have already been produced to accelerate the NRR process.[12−14] However, a significant challenge of increasing the concentration of N2 gas in the solution needs to be overcome due to a relatively high kinetic barrier occurring at the solid–liquid–gas interface in the NRR process.[15] Hence, the electrochemical synthesis of NH3 using N2 as a nitrogen source has a broad prospect, but there is still a long way to go to obtain considerable NH3 yield in practical application.[16]

Nitrate (NO3–), one of the nitrogenous pollutants in wastewater, is an active compound that makes reaction easier than N2 gas.[17] Domestic wastewater and residual chemical fertilizers, as well as untreated industrial wastewater, are the main source of NO3–. Applying electrochemical methods to degrade nitrogen-containing pollutants from industrial wastewater has become an important research field in environmental protection, intending to convert active nitrogen-containing species into N2 gas instead of NH3 gas.[18] Over the past few years, several metal catalysts and alloy catalysts have been developed to convert NO3– primarily into N2 gas with excellent efficiency.[19,20] Hence, it will be a quite meaningful study to find new and efficient electrocatalysts for the selective conversion of NO3– into NH3, which is rather beneficial both from the environmental perspective and from the economic perspective.

The single-atom catalysts (SACs) possess a large atomic utilization rate and highly active centers due to their unique configurations, which leads to excellent performance in various electrochemical reactions, such as the oxygen reduction reaction (ORR), hydrogen evolution reaction (HER), oxygen evolution reaction (OER), carbon dioxide reduction reaction (CO2RR), and NRR.[21] The metal atoms in the SACs form a particular coordination configuration with the neighboring nonmetallic atoms on the substrate, whose structure often plays a significant role in the stability and activity of the SACs.[22] In other words, a special substrate needs the ability to trap the metal atoms stably to prevent the agglomeration of the adjacent metal atoms, which is fundamental for the high catalytic activity of the SACs. For instance, conventional porous two-dimensional materials can provide binding sites to anchor metal atoms to construct different kinds of SACs, such as metal oxides, graphene, and g-C3N4.[22,23]

Graphdiyne (GDY) is a rapidly rising star in the carbon family referring to a class of 2D carbon allotropes, whose theoretical model was first proposed by Haley in 1997, and the real structure was first synthesized experimentally in 2010.[24,25] The configuration of GDY is composed of the sp2 hybrid benzene ring and sp hybrid diacetylene chain, which results in its unique chemical stability and mechanical stability.[26] Especially, GDY is a promising substrate for the SACs due to its natural uniform pores being conducive to the embedding of the metal atoms.[27] It is worth mentioning that many researchers have designed different kinds of graphdiyne-supported SACs with excellent catalytic performance experimentally and computationally. According to the theoretical calculation results, Ni@GDY,[28] Cu@GDY,[28] and Rh@GDY[29] display excellent electrocatalytic activity for the CO oxidation reaction. Pt@GDY,[30] Ni@GDY,[31] and Fe@GDY[31] have already been synthesized successfully, exhibiting desired electrocatalytic performance for the HER. In addition, Feng et al. have found that some TM@GDY SACs display a positive function in both the ORR and OER from the perspective of theoretical calculations.[32]

Results and Discussion

Inspired by the above research results, a theoretical investigation of the electrochemical nitrate reduction reaction (NO3RR) on the graphdiyne-supported transition metal SACs (TM@GDY SACs) (TM = Ti–Cu, Zr–Ag, Hf–Au) through first-principles calculation is expected to be conducted. First, the thermodynamic stability of the system consisting of the TM atoms and pristine GDY should be evaluated to determine which metal atoms possess the potential to form stable SACs with the pristine GDY. Subsequently, the electrocatalytic performance of all the thermodynamically-stable TM@GDY SACs in the process of reducing NO3– to NH3 continues to be studied, and the origin of NO3RR electrocatalytic performance for the most excellent TM@GDY SAC will be analyzed from the level of the electronic structure.

A 2 × 2 × 1 supercell of the pristine GDY monolayer is chosen to be optimized, whose optimized lattice constant reaches 9.454 Å, as shown in Figure a, which can be accepted with a negligible difference compared to the experiment result.[26] Furthermore, 24 types of transition metals (Ti–Cu, Zr–Ag, Hf–Au) are chosen to be anchored on the pristine GDY, whose stability is evaluated by the energy difference between the binding energy (Eb) and cohesive energy (Ecoh), as shown in Figure d. The optimized structures of all the TM@GDY SACs (TM = Ti–Cu, Zr–Ag, Hf–Au) can be divided into two species. The first configuration displays that the TM atoms are anchored at the corner of a triangular ring composed of 18 carbon atoms, called the corner pattern, as shown in Figure b, while the other configuration denotes that the TM atoms are anchored next to the center of the sp hybrid diacetylene chain, called the center pattern, as shown in Figure c. According to the analysis of the energy difference between the binding energy and cohesive energy and the criterion (ΔE < 0.15 eV), we can conclude that the ΔE values of the TM@GDY SACs (TM = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zr, Pd, Ag, Hf, Pt) are less than 0.15 eV. Moreover, the details of the length of the TM-C bonds, binding energies, cohesive energies, and charge transfer are summarized in Table S1. These results are identical to previous studies,[33] proving the stability of the TM@GDY SACs (TM = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zr, Pd, Ag, Hf, Pt) from the perspective of theoretical calculations.

Figure 1

(a) Atomic structure of the pristine GDY monolayer with the view from the top and side. (b) Corner pattern and (c) center pattern structure of the TM@GDY SACs from the top view and side view. The gray and light blue balls represent the C and TM atoms, respectively. (d) Diagram of the energy difference (ΔE) between the binding energy (Eb) and cohesive energy (Ecoh).

The chemisorption of NO3– is the first step of the NO3RR process, which plays an important role in the whole process. To evaluate how the chemisorption of NO3– affects the electrocatalytic reaction, it is necessary to calculate the adsorption strength of NO3– on the TM@GDY SACs (TM = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zr, Pd, Ag, Hf, Pt). According to the previous studies,[34,35] there are two possible initial configurations adsorbed on the TM center called the 1-O pattern (one oxygen atom adsorbed on the TM center of TM@GDY SACs) and the 2-O pattern (two oxygen atoms adsorbed on the TM center of TM@GDY SACs), as shown in Figure S1. With theoretical analysis of the data listed in Table S2, it can be found that the 2-O pattern exhibits more negative adsorption energies and more stable adsorbed configurations than the 1-O pattern. Therefore, it is reasonable to conclude that the 2-O pattern is the major configuration for NO3– adsorbed on the TM@GDY SACs. Moreover, the adsorption strength of NO3– on the pristine GDY monolayer is investigated as well with a value of +1.87 eV, proving the catalytic activation of the TM center in the TM@GDY SACs from another aspect. Notably, there are other competitive reactions against the NO3RR, such as the HER and NRR. For a parallel comparison, the adsorption energies of NO3– (ΔG*NO), H proton (ΔG*H), and N2 molecule (ΔG*N2) on the TM@GDY SACs are shown in Figure a and Table S3, and the corresponding configurations of *H and *N2 (* denotes the TM center) can be found in Figure S2. It can be inferred that ΔG*NO values are more negative than ΔG*H values, except Co@GDY and Cu@GDY with slightly more positive values, indicating that the NO3RR is a more favorable competitor with the HER. Moreover, ΔG*NO values even display a lower tendency than ΔG*N2 values for the early 3d TM@GDY SACs, which suggests that they may become superior candidates with higher selectivity of the NO3RR toward NH3 than that of the NRR.

Figure 2

(a) Diagram of adsorption energies of NO3–, N2 molecule, and H proton on the TM@GDY SACs for parallel comparison. (b) Scaling relation between charge transfer of the TM atoms and NO3– adsorption energies.

To gain a deeper insight into the adsorption trends of NO3– on the TM@GDY SACs, the relationship between charge transfer of the TM atoms and ΔG*NO values is shown in Figure b. They show a highly linear correlation with a nice correlation coefficient (R2 = 0.96), which provides an appropriate explanation for the variation of NO3– adsorption; that is, to say, the TM center with a larger electron loss could contribute to stronger adsorption of NO3–. Taking Hf@GDY as an example, its adsorption is the strongest one with a negative value of −2.54 eV, resulting from the Hf atom with the largest charge transfer of 1.49 e–, as listed in Table S1. Moreover, the positively charged TM atoms are easy for intermediates to be adsorbed, which can accelerate the subsequent reaction of the NO3RR.

The NO3RR is a reaction with complicated mechanisms, involving eight electrons participating in the whole steps.[36] Hence, there are different kinds of byproducts (NO2, NO, N2O, and N2) potentially produced during the conversion process of NO3– to NH3. Nevertheless, NH3/NH4+ is the most thermodynamically stable product at negative electrode potential. According to the previous studies, receivable mechanisms for the NO3RR process have already been proposed by Wang et al.,[34] indicating that four different configurations of adsorption intermediate *NO (called O-end, N-end, NO-side, and NO-dimer) might exist in the NO3RR process, as shown in Figure S3, which results in four different types of the mechanisms of NO3– to NH3 and a pathway from NO3– to N2, as shown in Figure a. To simplify the computation and pick up the most suitable pathway, the adsorption energies of NO with different configurations were calculated. Based on the analysis of adsorption energy differences of these NO configurations, the N-end adsorption configuration was proved to be the most stable one. Legitimately, the N-end pathway mechanism was chosen as the most reasonable mechanism for the NO3RR process. So the N-end pathway is mainly discussed in this study, in which each elementary step involved is displayed in the Supporting Information, while other pathways are not further studied.

Figure 3

(a) Schematic diagram for detailed pathways of the NO3RR, including O-end, O-side, N-end, and N-side pathway to NH3 and the NO-dimer pathway to N2. (b) Summary of limiting potentials on the TM@GDY SACs (TM = V, Cr, Mn, Fe, Co, Ni, Cu, Ag) for the NO3RR via the most favorable N-end pathway.

The limiting potential (UL), defined as UL = −ΔGmax/e, where ΔGmax denotes the maximum of free energy changes among all the elementary steps in the N-end pathway, is used to examine the NO3RR performance on the TM@GDY SACs (TM = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zr, Pd, Ag, Hf, Pt). The detailed free energy change values of each elementary step can be found in Tables S4 and S5. Moreover, it can be inferred from Figure b that the UL values of the vanadium-anchored graphdiyne (V@GDY) and Ag@GDY are −0.63 and −0.65 V respectively, which are the lowest two among all the TM@GDY SACs, implying that they may be promising electrocatalysts for the NO3RR to produce NH3. Nevertheless, the UL values of the TM@GDY SACs (TM = Ti, Zr, Pd, Hf, Pt) are not presented in Figure b, which can be attributed to the fact that some adsorption intermediates are unable to be adsorbed effectively on the TM center. Specifically speaking, for adsorption structures of all the five TM@GDY SACs (TM = Ti, Zr, Pd, Hf, Pt) shown in Figure S4, the *NO3H intermediate displays an inactive connection with the TM center for Pd@GDY and Pt@GDY (Figure S4e,g), and an unstable connection with the TM center for Ti@GDY, Zr@GDY, and Hf@GDY (Figure S4a–c) after structure relaxation. Furthermore, for Pd@GDY and Pt@GDY, it is also difficult for the *NO2H intermediate to be adsorbed on the TM center (Figure S4d,f). Notably, the *NO3H intermediate turns into two parts consisting of *O and *NO2H after structure relaxation for adsorption on the Ti@GDY, Zr@GDY, and Hf@GDY, while it remains a complete adsorption intermediate after structure relaxation without being adsorbed on the TM center of Pd@GDY and Pt@GDY, respectively. Inferentially speaking, the early transition metal atoms (Ti, Zr, and Hf) anchored on the GDY may lead to the high activity of the TM@GDY SACs, with the instability of the *NO3H intermediate, leading to the splitting of the *NO3H intermediate after optimization. On the contrary, the late transition metal atoms (Pd and Pt) embedded in the GDY exhibit more negative activity with the adsorption intermediates (*NO2H and *NO3H), leading to the desorption of *NO2H and *NO3H to the TM center.

As plotted in Figure a, the free energy diagram of the NO3RR on V@GDY is further and mainly analyzed to validate the nitrate-to-ammonia performance, while the result of Ag@GDY is displayed in Figure S6 and not discussed in detail, whose thermodynamic stability is verified by the ab initio molecular dynamics (AIMD) simulations shown in Figure S10. The corresponding structures of NO3RR intermediates are shown in Figure b (via the N-end pathway) and Figure S7 (via the NO-dimer pathway). Moreover, the free energy diagrams of the NO3RR (via the N-end pathway) on the other TM@GDY SACs (TM = Cr, Mn, Fe, Co, Ni, Cu) are also given in Figure S5,6, for comparison. It can be inferred that NO3– can be stably adsorbed with two TM-O bonds for the first step with an energy change of −1.44 eV. However, considering that the V atom is a little bit far away from the GDY plane displayed in Figure b, the AIMD simulation is carried out to check out the system stability of the structure of NO3 adsorbed on V@GDY with the results displayed in Figure S11. Therefore, it is reasonably believable that the system of all kinds of intermediates adsorbed on V@GDY in the NO3RR process is stable. Then, the following steps can be divided into two parts, with the seventh step used as the boundary. In the former part (from the second step to the seventh step), all the *NO intermediates (x = 3, 2, 1) suffer an entirely endothermal process by the attack of the (H+ + e–) pair to form their hydrogenated products, named *NOH intermediates, with the energy change of +0.29, +0.27, and +0.63 eV, respectively. However, the corresponding *NOH intermediates (x = 3, 2, 1) then go through an exothermal process with the release of H2O, whose energies drop by 1.86, 1.87, and 1.37 eV, respectively. Afterward, the removal of an oxygen atom in *NOH species (x = 3, 2, 1) requires the attack of the (H+ + e–) pair two times, in which extra energy is needed to keep the elementary reaction on for the first attack, while with the energy release for the second attack. Subsequently, in the latter part (from the eighth step to the eleventh step), the *N intermediate is hydrogenated by the (H+ + e–) pair consecutively three times to impel the formation of *NH3 and the desorption of NH3 eventually, with the energy changes of −0.67 eV for the step of *N to *NH, −0.41 eV for the step of *NH to *NH2, 0.03 eV for the step of *NH2 to *NH3, and 1.14 eV for the desorption of NH3, respectively. The tendency of the energy changes from the *N intermediate to the *NH2 intermediate is downhill, while is uphill for the *NH2 intermediate to the *NH3 intermediate with an energy change nearly close to 0.00 eV (0.03 eV), implying the spontaneousness of the elementary reaction without much external energy. Interestingly, the potential determining step for the TM@GDY SACs (TM = Cr, Fe) is the elementary reaction of *NO3 + H+ + e– → *NO3H, while the elementary reaction of *NO + H+ + e– → *NOH is the potential step for the TM@GDY SACs (TM = V, Mn, Co, Ni, Cu, Ag). Therefore, it is reasonable to conclude that the hydrogenation of the *NO species in the NO3RR process is relatively difficult to occur, leading to more extra energy needed to accelerate the production of NH3. Moreover, no overt barrier impedes the process from *NO3 to *NO; it is thus reasonably inferred that the electroreduction of NO2–, NO2, and NO to produce NH3 can be realized as well, in agreement with recent experimental results.[11,37]

Figure 4

(a) Free energy diagram of the NO3RR via the N-end pathway on V@GDY at an potential of 0.00 eV and −0.63 eV. The pathways to the release of NO2, NO, N2O, and N2 are also plotted for comparison. (b) Corresponding structures of NO3RR intermediates (via the N-end pathway) adsorbed on V@GDY.

To inspect the selectivity of the NO3RR toward NH3, the possibility of the formation of byproducts (NO2, NO, N2O, and N2) is considered, and the corresponding free energies are also plotted for comparison in the free energy diagram of V@GDY, as shown in Figure a. The elementary reaction steps and the corresponding adsorption structures involved the abovementioned byproducts are shown in Figure S7. The energy barriers for the release of NO2, NO, and N2O gas reach up to 2.74, 2.99, and 1.44 eV, respectively, on V@GDY, respectively, indicating the considerably difficult formation of these byproducts. Although N2 shows a lower energy barrier of 0.42 eV for its release, the elementary reaction of *N2OH to *N2 possesses a high energy barrier of 2.42 eV, which is impeditive for the formation of *N2, much less for the release of N2. In contrast, the desorption of NH3 demands less energy of 1.14 eV, which indicates the high selectivity for the NO3RR toward NH3 on V@GDY at an electrode potential of 0.00 V. However, with an extra negative electrode potential of −0.63 V applied in the NO3RR process of V@GDY, we can find that the release of NO2 and NO gases becomes remarkably accelerated due to much more accumulated free energy caused by the applied electrode potential.[38] Therefore, we can conclude that the desorption process of these byproducts is easier to happen with the applied electrode potential getting more negative, which could lead to the impurity of the NH3 product. Even so, the details of how the applied electrode potential influences the NO3RR process need more experimental results to verify the computational conclusions.

The density of states (DOS), band structure, charge and spin density distribution, and charge analysis, are performed to investigate the outstanding catalytic efficiency of V@GDY. The electron density of V@GDY is presented in Figure a, where the red regions around the V atom indicate the increasing electron density, which is confirmed by the Bader charge analysis. The C1 and C2 atoms denoted in Figure a,b gain 0.10 e– and 0.36 e–, respectively, while the V atom loses 1.10 e–, serving as an active site. The spin density is primary around the V atom which can be seen in Figure b, and the V@GDY monolayer possesses a total spin moment of 2.62 μB. The band structures of the pristine GDY and the V@GDY monolayer are depicted in Figure c,d. The pristine GDY possesses a direct bandgap of 0.49 eV, which is nearly consistent with previous theoretical reports.[26] However, the V@GDY displays a metallic character with a magnetic ground state because of the introduction of the occupied d orbital of the V atom, as shown in Figure d,e. The intrinsic spin moment could be responsible for the activation of NO3– on the V@GDY monolayer and for the high catalytic efficiency, which is in agreement with the conclusion that magnetism within catalysts could improve reaction activity.[39]

Figure 5

(a) Electron density of the V@GDY with an isosurface of 0.03 e Bohr–3. (b) The spin density of the V@GDY with an isosurface of 0.002 e Bohr–3. The spin-up and spin-down states are denoted by yellow and blue regions. (c) Band structure of the pristine GDY and (d) V@GDY. (e) DOSof the V@GDY with the Fermi level shifted to 0 eV. (f) Partial DOS (PDOS) and (g) pCOHP of NO3– adsorbed on the V@GDY. The dotted lines denote the Fermi level, referring to 0 eV. (h) Charge density difference of NO3– adsorbed on the V@GDY with an isosurface of 0.002 e Bohr–3.

To gain further insights into the activation of NO3– on V@GDY, the adsorption of NO3– on V@GDY can be revealed from two sides. On one hand, the V-3d orbital is almost above the Fermi level (EF), leading to a positive d-band center (εd) of 0.97 eV for V@GDY shown in Figure f. Based on previous studies,[40] the positive εd value implies the rather strong interaction between the V atom and NO3–. On the other hand, the hybridized energy levels of the V-3d orbital and NO3-2p orbitals split into the bonding and antibonding states, and the adsorption strength is determined by the antibonding states. Namely, the adsorption strength gets weakened when the antibonding population gradually moves down below EF.[41] The strong adsorption of NO3– is thus validated in Figure g because only a few antibonding states are below EF when NO3– is adsorbed on V@GDY. Additionally, the charge transfer is observed from the V atom to NO3– with 0.65 e– for V@GDY in Figure h, providing powerful clues into the adsorption of NO3– as well. Moreover, the Co@GDY with a moderate ΔG*NO of −0.13 eV is chosen for further comparison. The left-moving antibonding states shown in Figure S8 and the diminishing charge transfer shown in Figure S9 from the Co atom to NO3– on the Co@GDY lead to the weakened NO3– adsorption.

As discussed in the former part, the V atom prefers to be anchored at the corner of a triangular ring composed of 18 carbon atoms, leading to the bonding with the C1 and C2 atoms, whose bond length reaches 2.11 Å for the V–C1 bond and 2.07 Å for the V–C2 bond, as shown in Figure a. The calculated binding energy (Eb) between the V atom and the GDY monolayer is up to 5.39 eV, resulting from the significant hybridization of V-3d orbitals with C-2p orbitals from the C1 and C2 atoms in the sp hybrid diacetylene chain. Furthermore, the energy difference (ΔE) between the binding energy (Eb) and cohesive energy (Ecoh) on the V@GDY reaches 0.003 eV, proving the stability of V@GDY. Otherwise, the CI-NEB method is introduced to examine the thermodynamic stability of V@GDY, whose result is demonstrated in Figure a–c. The intake of extra energies of 1.40, 1.54, and 1.82 eV is required to transfer the initial state to the corresponding transition state, respectively, showing the excellent thermodynamic stability of V@GDY. Additionally, the AIMD calculations are performed to demonstrate the kinetic stability of the V@GDY monolayer. The total energy oscillates near the initial condition, as shown in Figure d, and the geometric structures of V@GDY are preserved well except for a slight distortion in the morphology after 600 K AIMD for 6 ps. It is thus reasonable to conclude that the V@GDY can serve as an efficient NO3RR catalyst for selective ammonia synthesis with high stability.

Figure 6

(a) Diffusion pathway of the V atom transferring from the corner to another corner on the same triangular ring. (b) Diffusion pathway of the V atom transferring from a corner to the corner of the opposite triangular ring. (c) Diffusion pathway of the V atom transferring from a corner to the corner of the neighboring triangular ring. The gray and pink balls represent the C atom and V atom, respectively. (d) Energy and temperature evolution vs the AIMD time for the V@GDY and the final structure of V@GDY after AIMD simulation lasting for 6 ps at 600 K.

Conclusions

In summary, 24 kinds of transition metal atoms (TM = Ti–Cu, Zr–Ag, Hf–Au) are selected to be anchored on the pristine GDY monolayer to construct the NO3RR SACs. The first-principles calculations are applied to comprehensively investigate the catalytic performance and product selectivity, as well as the structural stability, of these NO3RR electrocatalysts. Taking the different NO adsorption configurations into consideration, the possible mechanisms of the NO3RR could be classified as O-end, N-end, O-side, N-side, and NO-dimer pathways. The calculation results denote that the V@GDY shows the lowest limiting potential of −0.63 V versus RHE compared with other TM@GDY SACs (TM = Cr, Mn, Fe, Co, Ni, Cu, Ag). In terms of selectivity, the NO3RR process shows a preference for the NH3 product due to the high energy barriers to the formation of byproducts NO2, NO, N2O, and N2. Meanwhile, the NO3RR is more likely to happen than the NRR and HER due to the stronger adsorption strength of NO3– than the N2 molecule and H proton on the V@GDY. In addition, the structure of the V@GDY could be retained very well up to 600 K, exhibiting the initial configuration with high thermal stability. Therefore, our theoretical study validates that the V@GDY is a promising electrocatalyst to be fabricated experimentally, which is a new way for nitrate degradation and ammonia synthesis.

Computational Methods

All the density functional theory (DFT) calculations were carried out by the Vienna ab initio simulation package (VASP) with the use of Perdew–Burke–Ernzerhof functional within generalized gradient approximation.[42,43] To describe the expansion of the electronic eigenfunctions, the projector-augmented wave method was applied with a kinetic energy cutoff of 520 eV.[44] The DFT-D3 method was adopted for the consideration of describing the van der Waals interactions.[45] The solvation effect was not taken into consideration to simplify the calculation for the fact that the ignorable energy change was witnessed.[46] Bader charge analysis was performed to understand better the charge variation quantitatively.[47] The projected crystal orbital Hamilton population (pCOHP) was calculated by using LOBSTER to analyze the interaction between the TM atom and intermediates.[48] A 2 × 2 × 1 supercell of graphdiyne was accepted, whose vacuum space was 15 Å in the z-direction to minimize interlayer interactions. The Monkhorst-Pack 2 × 2 × 1 k-point mesh was employed for the structural optimizations, while the k-point mesh was set to 5 × 5 × 1 for the DOS calculations. All atomic positions were fully relaxed until energy and force reached the tolerance of 1 × 10–5 eV and 0.02 eV Å–1, respectively. To illustrate the kinetic stability, AIMD simulations were performed with the canonical (NVT) ensemble at 600 K for 6 ps.[49] Otherwise, VASPKIT, a convenient and efficient pre-and post-processing program for the VASP code, was used among partial calculations.[50]

The energy difference between the TM@GDY SAC and the single TM atom and pristine GDY is named the binding energy (Eb), which is defined asHowever, the energy difference between an atom of the bulk metal and a single TM atom is stated as the cohesive energy (Ecoh), which is calculated by

The energy difference between Eb and Ecoh, called ΔE, will be used as a criterion to judge whether the SACs possess thermodynamic stability and is shown asEGDY, ETM@GDY, ETM, and ETM-bulk mean the energies of pristine GDY, TM@GDY SAC, an isolated TM atom, and a bulk-metal atom, respectively. It is accepted that the TM@GDY SAC can be more favorable to being a SAC if the energy difference (ΔE) is smaller than 0.15 eV, which is suggested by the previous study.[51]

The adsorption energy (Eads) of reaction intermediates on the TM@GDY SAC monolayer is determined bywhere Esum denotes the energy of the adsorbate–catalyst system, while Eadsorbate denotes the energy of the isolated adsorbate.

According to the computational hydrogen electrode model,[52] the Gibbs free energy change (ΔG) of each elementary step attached to the reaction (NO3– to NH3) is defined aswhere ΔE means the adsorption energy obtained from the DFT calculations. ΔZPE denotes the correction of the zero-point energy, while ΔS is the difference of entropy, and T is room temperature (298.15 K), respectively. The calculated ΔZPE and TΔS are summarized in Table S6.[34] With the consideration of other external factors, the effect of the applied electrode potential and the pH of the solution are included by the correction of ΔU and ΔpH, respectively. Notably, ΔU = U × e, where U is the electrode potential, and e is the electron transfer, respectively. ΔpH = KT × ln 10 × pH, where KB is the Boltzmann constant, and pH = 0 in this study.

The first step of the whole reaction is the adsorption of NO3– to the TM@GDY SACs, but it is not convenient for the VASP code to calculate the energy of charged NO3– directly. So gaseous HNO3 is chosen as a reference instead to make the calculation feasible to go on with (details in Supporting Information).[53−55] The adsorption energy of NO3– (ΔG*NO) is described aswhere G*NO, G*, GHNO, and GH denote the Gibbs free energies of NO3– adsorbed on the TM@GDY SACs, TM@GDY substrates, HNO3 molecules, and H2 molecules in the gas phase, respectively. ΔGcorrect denotes the correction of the adsorption energy which is set to 0.392 eV.[34]