A 2D Ba₂N Electride for Transition Metal-Free N₂ Dissociation under Mild Conditions

Abstract

N₂ activation is a key step in the industrial synthesis of ammonia and other high-value-added N-containing chemicals, and typically is heavily reliant on transition metal (TM) sites as active centers to reduce the large activation energy barrier for N₂ dissociation. In the present work, we report that a 2D electride of Ba₂N with anionic electrons in the interlayer spacings works efficiently for TM-free N₂ dissociation under mild conditions. The interlayer electrons significantly boost N₂ dissociation with a very small activation energy of 35 kJ mol^–1, as confirmed by the N₂ isotopic exchange reaction. The reaction of anionic electrons with N₂ molecules stabilizes (N₂)^2– anions, the so-called diazenide, in the large interlayer space (∼4.5 Å) sandwiched by 2 cationic slabs of Ba₂N as the main intermediate.

Dinitrogen activation is a key step in the industrial synthesis of ammonia and many other high-value-added functional N-containing chemicals.¹⁻³ However, the high thermodynamic stability of N≡N triple bonds with kinetic inertness (945 kJ mol^–1) makes dinitrogen activation a significant challenge. Transition metal (TM)-based solid catalysts or complexes are generally necessary for N₂ activation to reduce the high activation energy barrier.^3,4 For example, N₂ activation for ammonia synthesis is well-known to occur on the surface of TMs such as Fe and Ru under high pressures and temperatures.^3,5 The development of efficient supports or promoters that enhance electron transfer to the TM sites to facilitate N≡N bond weakening through metal-to-N₂ π-back-donation^6,7 has been the common approach to decrease the N₂ activation barrier. For instance, a 12CaO·7Al₂O₃ (C12A7:e^–) electride with a very low work function (2.4 eV) can promote electron transfer to a Ru catalyst, and thus boost N₂ dissociation on the Ru surface with relatively low activation energy.^8,9 Although these new strategies can effectively decrease the N₂ activation energy barrier, TMs are still irreplaceable as active centers in most cases. The development of methods for TM-free N₂ activation under mild conditions has been a long-standing target due to the potential to realize an environmentally friendly alternative process for N₂-activation related to the industrial synthesis of chemicals.^10,11

Alkaline earth metal subnitrides, denoted as Ae₂N (Ae = Ba, Sr, and Ca), are 2D electride materials with anionic electrons sandwiched by cationic slabs of [Ae₂N]²⁺ with large interlayer spaces (3.9–4.5 Å).¹²⁻¹⁸ The low work function of Ae₂N (<3.0 eV along the (100) direction)¹⁴ endows them with high electron transport ability, so that they have high potential as catalyst supports to obtain negatively charged TM catalysts for the activation of adsorbed molecules.^11,19−23 However, the direct activation of molecules such as N₂ by electride materials in the absence of TM sites has not been reported to date.

In the present work, we report that a 2D Ba₂N electride works as an efficient catalyst for TM-free N₂ dissociation with a very small activation energy of 35 kJ mol^–1, as confirmed by the N₂ isotopic exchange reaction (N₂–IER). The (N₂)^2– anion is formed in the large interlayer space of Ba₂N (∼4.5 Å) by the reaction of the interlayer electron with N₂, which is confirmed by Raman spectroscopy and density functional theory (DFT) calculations.

The alkaline earth metal subnitride of Ba₂N was easily prepared by heating pure Ba metal in a N₂ flow under ambient pressure at elevated temperatures (see Supporting Information). The X-ray diffraction (XRD) pattern for Ba₂N synthesized at 700 °C is well assigned to a calculated Ba₂N phase, and the phase crystallizes with a strong (001) orientation (Figure 1a and Figure S1); i.e., the prepared Ba₂N crystal was oriented along the plane perpendicular to the c-axis. This Ba₂N phase began to form at temperatures as low as 300 °C and further crystallized at higher temperatures over 800 °C without the formation of other barium nitride phases (Figure S1), which suggests that the barium subnitride is the most stable phase in the Ba–N system under ambient N₂ flow. The microstructure of the as-prepared Ba₂N powder observed by using scanning electron microscopy (SEM) revealed a typical layered structure (Figure 1b). The sample was a black color due to the metallic nature that originates from itinerant anionic electrons. A single phase of Sr₂N powder can be prepared by the same synthesis procedure with only Ba replaced by Sr metal (Figure S2). In contrast to Ba₂N and Sr₂N, Ca₂N could not be obtained by the direct nitridation of Ca metal under the same synthesis conditions; instead, Ca₃N₂ with a deep red color was obtained (Figure S3), which indicates that Ca₃N₂ is the most stable phase of the calcium nitrides. Ca₂N powder could be prepared by the reduction of Ca₃N₂ with Ca metal at 800 °C^12,19 (Figure S4).

Figure 1.

(a) XRD pattern for as-prepared Ba₂N synthesized at 700 °C. The inset shows the photo of collected black Ba₂N powder. (b) SEM image of as-prepared Ba₂N.

As a 2D electride, Ba₂N is first shown to have high efficiency for N₂ dissociation as confirmed by N₂–IER under 4 kPa of ¹⁵N₂ and 16 kPa of ¹⁴N₂ (Figure 2a). The mass signals at m/z = 28, 29, and 30 were monitored with the reaction time at various reaction temperatures (Figures S5–S11). Alkaline earth metal nitrides, such as Sr₃N₂, Ca₃N₂, and Mg₃N₂, as nonelectride references, showed no activity for the N₂–IER, even at temperatures up to 500 °C, which indicates that they cannot dissociate dinitrogen molecules. In contrast, the Ba₂N alkaline earth metal subnitride functioned as an efficient catalyst for N₂ dissociation with outstanding activity above 200 °C (0.7 mmol g^–1 h^–1) and reached 23.1 mmol g^–1 h^–1 at 400 °C, which is far beyond the activity of the conventional Ru/MgO catalyst. Sr₂N also acted efficiently for the N₂–IER above 200 °C (0.6 mmol g^–1 h^–1) and reached 9.5 mmol g^–1 h^–1 at 400 °C. In contrast to Ba₂N and Sr₂N, Ca₂N exhibited much lower activity for the N₂–IER. Isotopic N₂ exchange over Ca₂N began above 375 °C (0.3 mmol g^–1 h^–1) and reached 4.1 mmol g^–1 h^–1 at 500 °C. The activation energy (E_a) of Ba₂N and Sr₂N for the N₂–IER was as low as 34.3 and 38.8 kJ mol^–1, respectively, which is much lower than that of Ca₂N (68.3 kJ mol^–1), the conventional Ru/MgO catalyst (108.3 kJ mol^–1) (Figure 2b), and even lower than that (58 kJ mol^–1) of the Ru/C12A7:e^– catalyst.⁸ These results demonstrate that 2D alkaline earth metal subnitrides of Ba₂N and Sr₂N can dissociate N₂ with a very small energy barrier in the absence of TM sites.

Figure 2.

(a) N₂–IER rate for Ae₂N, Ae₃N₂ and the conventional Ru/MgO catalyst at various temperatures. (b) Corresponding Arrhenius plots for Ae₂N and Ru/MgO.

To clarify the N₂ activation dynamics with the Ae₂N electrides, pure ¹⁵N₂ (20 kPa) was used to react with these electrides at 400 °C, and the mass signal at m/z = 29 (¹⁴N¹⁵N) was monitored with the reaction time (Figure 3a and Figures S12–S14). For Ba₂N and Sr₂N, the m/z = 29 signal increased quickly with reaction time, which suggests their lattice nitrogen species attended the nitrogen dissociation process and exchange with molecular nitrogen of ¹⁵N₂. The mass signal intensity at m/z = 29 for Ca₂N did not increase with reaction time under the same reaction conditions, which suggests that Ca₂N follows different N₂ dissociation pathways than Ba₂N and Sr₂N. The rates for m/z = 29 formation under isotopic N₂–IER conditions and pure ¹⁵N₂ treatment conditions over Ae₂N electrides are summarized in Figure 3b. The results clarify that the N₂ dissociation activity of both Ba₂N and Sr₂N is mainly induced by the exchange of lattice nitrogen species during the N₂ dissociation process, while Ca₂N only follows the direct N₂ dissociation process.

Figure 3.

(a) Monitoring of mass signal intensities at m/z = 28, 29, and 30 as a function of the reaction time when Ba₂N reacts with pure ¹⁵N₂ (20 kPa) at 400 °C. (b) Comparison of the N₂–IER rates measured under different gas atmospheres at 400 °C. Red columns: 16 kPa ¹⁴N₂ + 4 kPa ¹⁵N₂; Blue columns: 20 kPa ¹⁵N₂. (c) TPD of N₂ desorption (c) and Raman spectra (d) for Ba₂N collected by 20 kPa of pure ¹⁴N₂ (bottom) or ¹⁵N₂ (top) treatment at 400 °C for 2 h.

Temperature-programmed desorption (TPD) of N₂ from Ba₂N collected after treatment with pure ¹⁴N₂ or ¹⁵N₂ (20 kPa) was performed (Figure 3c) to determine the intermediates produced on/in Ba₂N during N₂ dissociation. When the as-prepared Ba₂N was pretreated in pure ¹⁴N₂, N₂ desorption peaks at low temperatures (290–570 °C) were observed. This is different from Ca₂N, which has only N₂ desorption from lattice N^3– ions at over 600 °C (Figure S15). After Ba₂N was heat treated in ¹⁵N₂, the m/z = 28 (¹⁴N₂) signal intensity almost disappeared, and instead, strong signals at m/z = 30 and 29 with the same peak shape emerged in the same temperature range. The fraction of exchanged ¹⁴N with ¹⁵N is estimated to be 14.4%. Similar results were also observed for Sr₂N, but not for Ca₂N (Figures S15 and S16). Raman analysis was performed to identify these activated nitrogen species (Figure 3d and Figure S17). The feature signals that appeared at 1554 cm^–1 for all of the tested samples were due to atmospheric O₂. The sample collected after pure ¹⁴N₂ treatment showed a main signal centered at 1447 cm^–1. After the ¹⁵N₂ treatment, the signal was split into two peaks at 1397 and 1418 cm^–1, which agreed well with the isotopic effect induced by the change of ¹⁴N₂ to ¹⁵N₂ and ¹⁵N¹⁴N. These signals can be attributed to stretching vibrations of (N₂)^2– anions, so-called diazenide, as identified in high-pressure stabilized inorganic materials and organic complexes.^1,24−27 Moreover, the Raman signal for (N₂)^2 – anions almost disappeared when Ba₂N was further heated in He at 600 °C (Figure S18). Therefore, N₂ desorption in the low-temperature range (290–570 °C) is confirmed to result from the release of (N₂)^2– anions that are formed as intermediates of N₂ dissociation on/in Ba₂N.

DFT calculations were performed to determine the structure of the incorporated (N₂)^2– anions in Ba₂N (see Supporting Information, Figure S19 and Table S1). When pure Ba₂N reacts with N₂ molecules (Ba₂₄N₁₂+3N₂) (Figure 4), (N₂)^2– anions are captured in the interlayers of Ba₂N through reaction with the anionic electrons. The stabilization of (N₂)^2– anions in Ba₂N proceeds with the formation energy of –2.92 eV, which indicates that N₂ molecules can be readily incorporated into the interlayers of Ba₂N by binding with the Ba ions present in the upper and lower layers. The bond length for the stabilized dinitrogen was 1.23–1.29 Å (Table S2), which is much longer than that for free N₂ (1.09 Å) but similar to the (N₂)^2– (∼1.20–1.35 Å).¹ The corresponding N₂ vibration was calculated to be 1424–1448 cm^–1 (Table S2), which is very close to the Raman analysis result (∼1446 cm^–1). The electron localization function calculation for Ba₂N with and without (N₂)^2– anions is shown in Figure 4 (right side), and the Bader charge analysis of an (N₂)^2– anion in the model of Ba₂₄N₁₂+3N₂ is shown in Figure S20. The high-density anionic electron layers are located in the interlayer spaces of pure Ba₂N. After the formation of an (N₂)^2– anion within an interlayer, the electron concentration in the interlayer was distinctly decreased compared to that for pure Ba₂N. This is because the N₂ molecules were activated as (N₂)^2– ions by accepting electrons from the interlayer, and the resultant (N₂)^2– ions are stabilized electrostatically by coordination with Ba²⁺ ions. The electron concentration of Ba₂N (0.002 mol g^–1) was evaluated by a iodometric titration (Table S3),²⁸ which suggests the molar ratio for Ba₂N:e^– is 1.0:0.6. This ratio is significantly lower than the theoretical ratio of 1.0:1.0 for Ba₂N, which confirms that the (N₂)^2 – anions were formed in the interlayers through the consumption of a portion of the electrons.

Figure 4.

Schematic illustration of (N₂)^2– anion formation in the interlayers of Ba₂N as demonstrated by DFT calculations.

Ca₂N shows a significantly weaker ability for N₂ activation than Ba₂N and Sr₂N. According to the DFT calculation results, the stabilized (N₂)^2 – anions are likely to be dissociated at the interlayer space of Ba₂N (Figure S21). The work functions of Ba₂N and Ca₂N for the (104) plane are calculated as 2.05 and 2.55 eV, respectively. As the work function of Ba₂N is close to or slightly lower than the lowest unoccupied molecular orbital (LUMO) energy level of the N₂ molecule (∼2.1 eV), electron donation from the Ba₂N to N₂ molecule, therefore, easily occurs (Figure S22). However, the work function of Ca₂N is a little bit larger, making it difficult to donate its interlayer electron to the N₂ molecule. This is why Ca₂N shows a much larger activation energy for N₂–IER compared to that of Ba₂N.

In conclusion, we have demonstrated that an easily obtained 2D Ba₂N electride works efficiently for TM-free N₂ dissociation with a very small activation energy of 35 kJ mol^–1. Isotopically labeled TPD and Raman analyses, and DFT calculation results clearly confirmed the formation of the (N₂)^2– anion, so-called diazenide, in the interlayer space of Ba₂N as the main intermediate. This study thus demonstrated a new way to realize TM-free N₂ activation under mild conditions.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c09362.

• The experimental section, which includes the synthesis of Ae₂N, N₂–IER measurement methods, TPD of N₂, Raman observations, and DFT calculation methods can be found in the Supporting Information. (PDF)

Author Contributions

^§ Z.Z and Y.J. contributed equally to this work.

The authors declare no competing financial interest.