Experimental and Theoretical Study of N2 Adsorption on Hydrogenated Y2C4H− and Dehydrogenated Y2C4− Cluster Anions at Room Temperature

The adsorption of atmospheric dinitrogen (N2) on transition metal sites is an important topic in chemistry, which is regarded as the prerequisite for the activation of robust N≡N bonds in biological and industrial fields. Metal hydride bonds play an important part in the adsorption of N2, while the role of hydrogen has not been comprehensively studied. Herein, we report the N2 adsorption on the well-defined Y2C4H0,1− cluster anions under mild conditions by using mass spectrometry and density functional theory calculations. The mass spectrometry results reveal that the reactivity of N2 adsorption on Y2C4H− is 50 times higher than that on Y2C4− clusters. Further analysis reveals the important role of the H atom: (1) the presence of the H atom modifies the charge distribution of the Y2C4H− anion; (2) the approach of N2 to Y2C4H− is more favorable kinetically compared to that to Y2C4−; and (3) a natural charge analysis shows that two Y atoms and one Y atom are the major electron donors in the Y2C4− and Y2C4H− anion clusters, respectively. This work provides new clues to the rational design of TM-based catalysts by efficiently doping hydrogen atoms to modulate the reactivity towards N2.


Introduction
More than 99% of the global nitrogen exists in the shape of gaseous dinitrogen (N 2 ) in the atmosphere, yet most organisms can only metabolize nitrogen-containing substances such as NH 3 rather than N 2 directly. Although N 2 is the main nitrogen source for most natural and artificial nitrogen-containing compounds, the high bond dissociation energy (9.75 eV) and the large HOMO-LUMO gap (10.8 eV) render its adsorption and activation an enormous challenge in chemistry [1][2][3][4]. Scientists regularly rely on transition metal (TM) centers to catalyze the nitrogen conversion processes [5][6][7]. The initial and critical step in the complicated reduction of dinitrogen is the adsorption of N 2 molecules at the TM center [8,9]. The fixation of nitrogen in industry is carried out at metal-based (Fe − or Ru − ) catalysts under extremely high temperatures (300-500 • C) and high pressures (100-300 atm), involving the disadvantages of large energy consumption and greenhouse gas emission [10][11][12]. Thus, it is vital to develop mild, energy-saving, and environment−friendly catalytic systems for N 2 fixation at ambient conditions. The activation of nitrogen by transition metal compounds with the involvement of hydrogen atoms is of particular interest, while the most common feature of N 2 hydrogenative cleavage is the participation of metal hydride bonds [13][14][15]. A literature survey [13] shows that metal hydride bonds have several important roles: (1) as a hydrogen source; (2) as an electron source for N 2 reduction; (3) as a powerful reducing agent for the removal of activated nitrogen atoms; and so on. As an ideal model of condensed-phase systems, gas-phase clusters can study chemical reactions and reveal related mechanisms at the strictly molecular level by simulating active sites. [16][17][18][19]. Several theoretical and experimental studies have reported the reactivity of metal species with nitrogen, however, only a few metal species such as, Sc 2 [20], Ta [28] have been characterized to cleave the N≡N triple bond completely. It can be seen that for the studies on N 2 adsorption in the gas phase, there are few metal species, and they mainly focus on the early transition metals. In the previous work, we found that a suitable number of hydrogen atoms has an influence on the reactivity of transition metal-containing clusters with N 2 [24][25][26]29,30]. Sc 3 NH 2 + [26] can effectively realize the activation of N 2 by H 2 , which is based on the regulation of N 2 reduction by two H atoms. Ta 3 N 3 H 0.1 − is an example that highlights the importance of the assisted reactivity of a single hydrogen atom, and the reactivity of Ta 3 N 3 H − is higher by a factor of five compared with that of Ta 3 N 3 − due to the hydrogen atom changing the charge distribution and geometry [25]. How can hydrogen atoms be efficiently doped to modulate the reactivity of TM-containing systems towards N 2 at the molecular scale? Considering the previous exploration of the Sc systems and the fact that Sc and Y belong to the same group, Y 2 C 4 − and Y 2 C 4 H − cluster anions were synthesized, and the reactivity towards N 2 was investigated by mass spectrometry and DFT calculations, to answer this question. This work clearly revealed that Y 2 C 4 H 0,1 − anions can adsorb N 2 , and the hydrogen atom greatly enhances the reactivity of Y 2 C 4 H − towards N 2 .

Results and Discussion
The time-of-flight (TOF) mass spectra of laser ablation-generated, further mass-elected Y 2 C 4 − and Y 2 C 4 H − cluster anions reacting with N 2 under thermal collision conditions in a linear ion trap (LIT) reactor are shown in Figure 1. The mass spectra for the generation of Y 2 C 4 H 0,1 − clusters has been given (Supplementary Figure S1). Upon the interactions of Y 2 C 4 − and Y 2 C 4 H − with N 2 , two adsorbed complexes that are assigned as Y 2 C 4 N 2 − and Y 2 C 4 HN 2 − are observed (Figure 1b,d), suggesting the following channels in Equations (1) and (2): Compared with Y 2 C 4 − , Y 2 C 4 H − shows a higher reactivity towards N 2 under the same reaction conditions in Figure 1f. Besides the major products, two weak peaks in Figure 1 are assigned to Y 2 C 4 OH − and Y 2 C 4 O 2 H − , generated from the reaction of Y 2 C 4 H 0,1 − anions with water impurities in the LIT. The pseudo-first-order rate constants (k 1 ) for the reactions one and two are estimated to be (3.7 ± 0.8) × 10 −12 cm 3 molecule −1 s −1 and (6.2 ± 1.3) × 10 −14 cm 3 molecule −1 s −1 , which are based on a least-square fitting procedure, corresponding to reaction efficiencies (Φ) [31,32] of 0.6% and 0.01%, respectively. Additionally, the signal dependence of product Y 2 C 4 H 0,1 N 2 − ions on N 2 pressures was obtained, which are derived and fitted with the mass spectrometry experimental data (Supplementary Figure S2). BPW91 calculations are performed to investigate the structures of reactant Y 2 C 4 H 0,1 − anion clusters (Supplementary Figure S3), as well as the reaction mechanisms between Y 2 C 4 H 0,1 − and N 2 . The lowest-energy isomer of Y 2 C 4 − (doublet, 2 IA1, Supplementary Figure S3), which is 0.08 eV lower than its quartet isomer, is a C s −symmetric six−membered ring, with the Y-Y bond as the symmetry axis and two C 2 ligands bonded to the two Y atoms. Moreover, the most stable isomer of Y 2 C 4 H − ( 1 IA2) has a hydrogen atom binding to the Y1 atom in the six-membered ring, similar to the Y 2 C 4 − ( 2 IA1), and it is 0.07 eV lower than the triplet state in energy (Supplementary Figure S3). Since the energies of the isomers are very close, their reaction paths are calculated. The results show that, in the reaction coordinates, the energies of the doublet and singlet stationary points and the products in the Y 2 C 4 − /N 2 and Y 2 C 4 H − /N 2 systems are lower than those of the corresponding quartet and triplet analogues, respectively (Supplementary Figure S4). Enthalpy and Gibbs free energies along with electronic and zero-point correction energies are added (Supplementary  Table S1). The concentration of dinitrogen adducts in the gas phase is relatively low, so it is difficult to collect and continue to measure Raman spectra. Currently, it is difficult to characterize structures due to technical and instrumental limitations. Infrared multiple photon dissociation may be applied to reveal such types of anions. We have added the calculated infrared spectra (Supplementary Figure S5), and the vibrational frequencies may be used for future experimental identification of these clusters.
spectrometry and DFT calculations, to answer this question. This work clearly revealed that Y2C4H0,1 − anions can adsorb N2, and the hydrogen atom greatly enhances the reactivity of Y2C4H − towards N2.

Results and Discussion
The time-of-flight (TOF) mass spectra of laser ablation-generated, further masselected Y2C4 − and Y2C4H − cluster anions reacting with N2 under thermal collision conditions in a linear ion trap (LIT) reactor are shown in Figure 1. The mass spectra for the generation of Y2C4H0,1 − clusters has been given (Supplementary Figure S1). Upon the interactions of Y2C4 − and Y2C4H − with N2, two adsorbed complexes that are assigned as Y2C4N2 − and Y2C4HN2 − are observed (Figure 1b,d), suggesting the following channels in Equations (1) and (2): The potential energy surfaces (PESs) of the most favorable reaction pathways are given in Figure 2. The N 2 molecule is initially captured by the Y1 atom in both Y 2 C 4 − and Y 2 C 4 H − to form the end-on-coordinated complexes 2 I1 and 1 I4. Notably, 2 I1 (−0.71 eV) in Figure 2a is as stable as 1 I4 (−0.70 eV) in Figure 2b, suggesting that the N 2 −adsorbed intermediates 2 I1 and 1 I4 are not the final products in the Y 2 C 4 -/N 2 and Y 2 C 4 H -/N 2 systems. As for the Y 2 C 4 − /N 2 system, the coordination mode of N 2 is further changed from η 1 in 2 I1 to η 2 in 2 I2 via 2 TS1. During this process, the N-N bond length is elongated from 110 pm in free N 2 to 119 pm in 2 I2. Subsequently, the adsorbed N 2 unit is anchored by two Y atoms via 2 TS2, forming a Y-N-N-Y bridge; at the same time, a longer N-N bond of 123 pm is generated in 2 P1. Note that the rupture of the N-N bonds encounters a high energy barrier ( 2 TS3, +2.46 eV with respect to the separated reactants), so that further activation of N 2 is hampered in this system.
The reaction of Y 2 C 4 H − /N 2 (Figure 2b) follows the similar mechanism. The complex is coordinated laterally to form a Y-N-N-Y bridge like 2 P1 by overcoming a negligible barrier 1 TS4, and the activation energy (∆E a , i.e., the energy difference between the encounter complex and the transition state) is lower than that of 2 I2 → 2 TS2 (∆E a = 0.23 eV) in Y 2 C 4 − . In the step of 1 I4 → 1 P2, an elongation of the N-N bond from 115 to 121 pm occurs. Further cleavage of N-N is also hindered due to the positive energy barrier of 4.89 eV ( 1 TS5). In addition, another adsorption of N 2 on the Y2 atom (Supplementary Figure S6) that is not bonded with the hydrogen atom can be eventually trapped in 1 P2 by generating the η 2 -mode intermediate 1 I7. In conclusion, the reactions of Y 2 C 4 H − and Y 2 C 4 − with N 2 result in the formation of bridging adsorption products 2 P1 and 1 P2, and the adsorbed N 2 molecules are in the η 1 :η 2 mode. As shown in Figure 3, the potential energy curves reveal that the adsorption process of Y 2 C 4 H − /N 2 is more favorable kinetically compared to that of Y 2 C 4 − /N 2 , since it is barrier−free for Y 2 C 4 H − /N 2 . A small barrier exists in the shallow entrance channels when N 2 approaches Y 2 C 4 − , which further explains the experimental observed low reaction rate constant for the dehydrogenated Y 2 C 4 − /N 2 . The reaction of Y2C4H − /N2 (Figure 2b) follows the similar mechanism. The complex is coordinated laterally to form a Y-N-N-Y bridge like 2 P1 by overcoming a negligible barrier 1 TS4, and the activation energy (ΔEa, i.e., the energy difference between the encounter complex and the transition state) is lower than that of 2 I2 → 2 TS2 (ΔEa = 0.23 eV) in Y2C4 − . In the step of 1 I4 → 1 P2, an elongation of the N-N bond from 115 to 121 pm occurs. Further cleavage of N-N is also hindered due to the positive energy barrier of 4.89 eV ( 1 TS5). In addition, another adsorption of N2 on the Y2 atom (Supplementary Figure S6) that is not bonded with the hydrogen atom can be eventually trapped in 1 P2 by generating the η 2mode intermediate 1 I7. In conclusion, the reactions of Y2C4H − and Y2C4 − with N2 result in the formation of bridging adsorption products 2 P1 and 1 P2, and the adsorbed N2 molecules are in the η 1 :η 2 mode. As shown in Figure 3, the potential energy curves reveal that the adsorption process of Y2C4H − /N2 is more favorable kinetically compared to that of Y2C4 − /N2, since it is barrier−free for Y2C4H − /N2. A small barrier exists in the shallow entrance channels when N2 approaches Y2C4 − , which further explains the experimental observed low reaction rate constant for the dehydrogenated Y2C4 − /N2. Frontier orbital analysis shows that the immobilization of the N2 ligand, as well as the formation of 2 P1 and 1 P2, involve d-electrons transfer from the single-occupied molecular orbital-1 (SOMO-1) of Y2C4 − and the HOMO orbital of Y2C4H − to the antibonding π*-  The reaction of Y2C4H − /N2 (Figure 2b) follows the similar mechanism. The complex is coordinated laterally to form a Y-N-N-Y bridge like 2 P1 by overcoming a negligible barrier 1 TS4, and the activation energy (ΔEa, i.e., the energy difference between the encounter complex and the transition state) is lower than that of 2 I2 → 2 TS2 (ΔEa = 0.23 eV) in Y2C4 − . In the step of 1 I4 → 1 P2, an elongation of the N-N bond from 115 to 121 pm occurs. Further cleavage of N-N is also hindered due to the positive energy barrier of 4.89 eV ( 1 TS5). In addition, another adsorption of N2 on the Y2 atom (Supplementary Figure S6) that is not bonded with the hydrogen atom can be eventually trapped in 1 P2 by generating the η 2mode intermediate 1 I7. In conclusion, the reactions of Y2C4H − and Y2C4 − with N2 result in the formation of bridging adsorption products 2 P1 and 1 P2, and the adsorbed N2 molecules are in the η 1 :η 2 mode. As shown in Figure 3, the potential energy curves reveal that the adsorption process of Y2C4H − /N2 is more favorable kinetically compared to that of Y2C4 − /N2, since it is barrier−free for Y2C4H − /N2. A small barrier exists in the shallow entrance channels when N2 approaches Y2C4 − , which further explains the experimental observed low reaction rate constant for the dehydrogenated Y2C4 − /N2. Frontier orbital analysis shows that the immobilization of the N2 ligand, as well as the formation of 2 P1 and 1 P2, involve d-electrons transfer from the single-occupied molecular orbital-1 (SOMO-1) of Y2C4 − and the HOMO orbital of Y2C4H − to the antibonding π*- Frontier orbital analysis shows that the immobilization of the N 2 ligand, as well as the formation of 2 P1 and 1 P2, involve d-electrons transfer from the single-occupied molecular orbital-1 (SOMO-1) of Y 2 C 4 − and the HOMO orbital of Y 2 C 4 H − to the antibonding π*orbitals of N 2 (Supplementary Figure S7). The presence of hydrogen atoms enhances the reactivity of the cluster cations toward N 2 since it changes the charge distribution. As shown in Figure 4a, the Y1 linked to the hydrogen atom on the Y 2 C 4 H − cluster has more negative charges compared to Y 2 C 4 − , and it promotes π-back-donation. Note that the energy differences between the transition states and the separated reactants, which is the apparent barrier (∆E ‡ ), matters in gas−phase studies. The apparent barrier for Y 2 C 4 H − /N 2 (∆E ‡ = −0.70 eV) is lower than that of Y 2 C 4 − /N 2 (∆E ‡ = −0.48 eV), and the energy of 1 P2 is lower than that of 2 P1 (−1.61 eV vs. −1.35 eV). According to the Rice-Ramsperger-Kassel-Marcus (RRKM) theory [33], the internal conversion rate of I4 → TS4 (8.49 × 10 11 s −1 ) is 32 times larger than that of I2 → TS2 (2.65 × 10 10 s −1 ). These theoretical results are consistent with the experiments.
cessible to the Y2C4H − cluster anions. The values of bond orders of Y-Y bond in Y2C4H0,1 − anions are an important indicator for the ability of storing electrons, which increases from 0.55 in Y2C4 − ( 2 IA1) to 0.66 in Y2C4H − ( 1 IA2). Therefore, although hydrogen appears to be a bystander in N2 adsorption, its presence indeed stores more electrons in the Y-Y bond and facilitates N2 adsorption. It can be concluded that the hydrogen atom in the Y2C4H − cluster significantly affects the charge distribution and electronic structure, and a suitable number of hydrogen atoms can enhance the reactivity towards N2.  To further improve the understanding of Y 2 C 4 H 0,1 − /N 2 systems, NBO analysis along reaction coordinates was performed (Figure 4b,c). The charge details were added (Supplementary Table S2). In the adsorption processes IA1 → I1 and IA2 → I4 of Y 2 C 4 H 0,1 − /N 2 , the yttrium atoms transfer 0.37 e and 0.29 e to the N1 atom, respectively, leading to the formation of the Y-N1 bonds, while two N2 atoms in Y2C4 − and Y 2 C 4 H − only increase by 0.11 e. In the subsequent steps I2 → P1 and I4 → P2 for the formation of the N2-Y2 bonds, more electrons are stored in the two nitrogen atoms, resulting in the gradual elongation of the N-N bonds. Overall, the electrons required for the N 2 adsorption by the Y 2 C 4 − and Y 2 C 4 H − clusters are mainly provided by Y atoms with total transferred amounts of 0.88 e and 0.78 e, respectively. Differently, two and one Y atoms are the electron donors in Y 2 C 4 − and Y 2 C 4 H − , respectively. The active Y1 atom in Y 2 C 4 − (IA1) has more 5s electron occupancies (5s 1.10 4d 1.03 ), which causes an unfavorable approach and a high σ-repulsion on the N 2 molecule. When one hydrogen atom on the Y 2 C 4 − ( 2 IA1) cluster bonds to form Y 2 C 4 H − ( 1 IA2), the natural charge on the Y1 increases from 0.79 e to 1.48 e; at the same time, more 4d and less 5s electron occupancies are located (5s 0.38 4d 1.12 ), which can make N 2 more accessible to the Y 2 C 4 H − cluster anions. The values of bond orders of Y-Y bond in Y 2 C 4 H 0,1 − anions are an important indicator for the ability of storing electrons, which increases from 0.55 in Y 2 C 4 − ( 2 IA1) to 0.66 in Y 2 C 4 H − ( 1 IA2). Therefore, although hydrogen appears to be a bystander in N 2 adsorption, its presence indeed stores more electrons in the Y-Y bond and facilitates N 2 adsorption. It can be concluded that the hydrogen atom in the Y 2 C 4 H − cluster significantly affects the charge distribution and electronic structure, and a suitable number of hydrogen atoms can enhance the reactivity towards N 2 .

Experimental Methods
The metal carbide clusters were generated by laser ablation metal target (made of pure yttrium powder) (Jiangxi Ketai New Materials Co. Ltd, Jiangxi, China) seeded at 2‰ CH 4 (Beijing Huatong Jingke Gas Chemical Co. Ltd, Beijing, China) in a helium carrier gas (backing pressure 4 atm). The pulsed laser is a 532 nm laser with 5-8 mJ/energy pulses and 10 Hz repetition rate (140 Baytech Drive, San Jose, CA, USA). Y 2 C 4 − and Y 2 C 4 H − anion clusters were mass-selected by a quadrupole mass filter (QMF) (China Academy of Engineering Physics, Mianyang, Sichuan, China) [34] and subsequently entered into a linear ion trap (LIT) reactor (homemade) [35]. After being confined and thermalized by the pulsed gas He for about 2 ms, they interacted with N 2 for about 6 ms and 14 ms, at room temperature, respectively. The anion clusters were ejected from the LIT and then detected by a reflection time-of-flight mass spectrometer (TOF-MS) [36]. The rate constants of the reactions between Y 2 C 4 H 0,1 − cluster anions and N 2 were described [37]. A schematic diagram of the experimental apparatus is shown in ref [34].

Computational Methods
All DFT [38] calculations were formed using the Gaussian 09 [39] program package to explore the structures of reactant clusters Y 2 C 4 H 0,1 − and the mechanistic details of Y 2 C 4 H 0,1 − with N 2 . To give the best interpretation of the experimental data, we calculated the dissociation energies of the Y-Y Y-C, N-N and C-C (Supplementary Table S3) bonds using 20 methods. The results show that BPW91 functional [40][41][42] performs very well. For application of basis sets in reaction systems, the def2-TZVP [43] basis set was used for the Y atom, and the 6 − 311 + G * basis sets [44,45] were selected for the C, H, and N atoms, which were applied in other systems containing these elements [24,27,46]. The zero-point vibration corrected energies (∆H 0K in eV) in unit of eV are reported. Vibrational frequency calculations must be performed for the geometric optimization of the reaction intermediates (IMs) and transition states (TSs) [47]. Intrinsic reaction coordinate [48] calculations were employed to ensure whether each TS was connected to two appropriate local minima. DFT-D3 correction for the complexes were contained in the system. Natural population analysis was performed using NBO 6.0 [49], and the orbital composition was analyzed by the method of natural atomic orbitals employing the Multiwfn program [50].

Conclusions
In summary, the reactions of Y 2 C 4 H − and dehydrogenated Y 2 C 4 − cluster anions with N 2 have been investigated experimentally and theoretically. The experimental results indicate that the reaction rate constant of Y 2 C 4 H − /N 2 is higher by a factor of 50 compared with that of Y 2 C 4 − /N 2 . DFT calculations indicate that the differences are caused by the different charge distributions and the bonding of the additional hydrogen atom to the yttrium atom in the Y 2 C 4 H − cluster, resulting in more 4d electron occupancies and thus more efficient π-back-donation bonding with N 2 molecules. The electron donor atoms of Y 2 C 4 − and Y 2 C 4 H − anion clusters are different, for Y 2 C 4 − , two Y atoms donate electrons, while only one Y atom donates electrons in Y 2 C 4 H − . Storing more electrons in the Y-Y bond is also an important influence of the hydrogen atom on the reactivity of Y 2 C 4 H − to N 2 . This study clearly reveals the significance of hydrogen-assisted reactions in N 2 adsorption processes. Attaching an appropriate number of hydrogen atoms on active sites can enhance the N 2 adsorption rates, providing a new strategic direction for the rational design of TM-based energy-efficient nitrogen fixation catalysts.