Ta₂⁺-mediated ammonia synthesis from N₂ and H₂ at ambient temperature

Significance

A combined experimental/computational approach provides deep mechanistic insight into an unprecedented cluster-mediated N−H coupling mimicking the industrially extremely important ammonia synthesis from N₂ and H₂ (the “Haber–Bosch” process) at room temperature. Crucial steps were identified for both the forward reactions (i.e., the activation of N₂) and the backward process (i.e., the Ta₂⁺-mediated decomposition of NH₃). The central intermediate for either path corresponds to Ta₂N₂⁺, a four-membered ring with alternating Ta and N atoms. The root cause of tantalum’s ability to bring about nitrogen fixation and its coupling with H₂ under mild conditions has been identified by state-of-the-art quantum chemical calculations.

Abstract

In a full catalytic cycle, bare Ta₂⁺ in the highly diluted gas phase is able to mediate the formation of ammonia in a Haber–Bosch-like process starting from N₂ and H₂ at ambient temperature. This finding is the result of extensive quantum chemical calculations supported by experiments using Fourier transform ion cyclotron resonance MS. The planar Ta₂N₂⁺, consisting of a four-membered ring of alternating Ta and N atoms, proved to be a key intermediate. It is formed in a highly exothermic process either by the reaction of Ta₂⁺ with N₂ from the educt side or with two molecules of NH₃ from the product side. In the thermal reaction of Ta₂⁺ with N₂, the N≡N triple bond of dinitrogen is entirely broken. A detailed analysis of the frontier orbitals involved in the rate-determining step shows that this unexpected reaction is accomplished by the interplay of vacant and doubly occupied d-orbitals, which serve as both electron acceptors and electron donors during the cleavage of the triple bond of N≡N by the ditantalum center. The ability of Ta₂⁺ to serve as a multipurpose tool is further shown by splitting the single bond of H₂ in a less exothermic reaction as well. The insight into the microscopic mechanisms obtained may provide guidance for the rational design of polymetallic catalysts to bring about ammonia formation by the activation of molecular nitrogen and hydrogen at ambient conditions.

The direct use of molecular nitrogen with its thermodynamically stable and kinetically inert triple bond as one of the very few commodities that are freely available worldwide and in almost unlimited quantities is essential for life on Earth (1–3). Nature utilizes nitrogen-binding enzymes, the nitrogenases, to catalyze the conversion of nitrogen to ammonia at ambient conditions (4, 5). In contrast, its industrial production still relies on the highly energy-demanding Haber–Bosch process to bring about the challenging chemical marriage of N₂ and H₂ to form NH₃ (3, 6–8), which consumes ca. 1–2% of the world’s energy production (8–11). In addition, presently about 1.5 tons of the greenhouse gas carbon dioxide are produced per ton of ammonia (9). To slow down global warming (12), it would, therefore, be sensible, in addition to numerous other measures, to find a process for producing ammonia on an industrial scale from the molecular feedstock nitrogen and hydrogen in an economically viable and environmentally benign way.

The greatest obstacle to the production of ammonia from N₂ corresponds to the cleavage of the N≡N triple bond, which with a bond energy of 945 kJ mol⁻¹ (13), constitutes one of the strongest chemical bonds. While some progress has been made on the daunting road to artificial nitrogen activation, the number of well-defined complexes that bind N₂ and ultimately, lead to a complete cleavage of the N≡N triple bond is rather limited so far. For instance, complexes with single (14–32) and multiple (15, 33–39) transition metal centers, small metal clusters (40–48), and also, main group compounds (49–52) have been found to be able to split dinitrogen. The activation of N–N bonds by oriented external electric fields (53, 54) followed by insertion of, for example, a nitrogen atom in the C–C bonds of alkanes has been reported as well (55, 56). Another promising approach is based on the electrochemical cleavage of N₂ to produce ammonia (11, 31, 32, 51, 57–60). There are indications that the cooperative activation of N₂ by several transition metal atoms holds promise as well (28, 34–39, 45, 61–64). Furthermore, the reactivity of ditantalum complexes of the type ([NPN]Ta(μ-H))₂N₂ ([NPN] = PhP(CH₂SiMe₂NPh)₂) with dinitrogen (65) has also been extensively investigated in the past (28, 34, 61–71).

Mechanistically, the catalytic activity of the transition metals is based on the interplay of vacant and filled d-orbitals during multielectron rearrangements along the reaction coordinate. Here, the vacant orbitals of the metal center receive electrons from N₂ and simultaneously weaken (or cleave) the triple bond of dinitrogen by donating electron density from the filled d-orbitals into the antibonding π*-orbitals of N₂ (49, 50). According to the conceptual framework outlined by Fryzuk and coworkers (67), it was recognized that the ability of Ta compounds to store two electrons in a Ta–Ta bond is a prerequisite for subsequent reductive transformations and is of paramount importance to split the N≡N triple bond completely.

While the impressive progress made in recent decades is undeniable, a deep and comprehensive understanding of the various mechanistic details related to either fixation or activation of N₂ to ammonia is far from being complete. This also applies to a consistent description of the elementary steps involved, with the exception of the elegant elucidation of the mechanism of the Haber–Bosch process by Ertl and coworkers (7, 8, 72, 73).

As has been shown time and again, gas-phase experiments provide an ideal arena for tackling many challenging mechanistic issues at a strictly molecular level, such as investigating the detailed course of chemical reactions, including those that are industrially relevant (74–88). Structurally properly characterized gas-phase clusters have been chosen as prototypical models to probe the active sites in (including but not limited to) heterogeneous catalysis aimed at a better understanding of the intrinsic factors that govern reactivity patterns in the condensed phase (89, 90). In addition, it has been shown that fundamental questions can be addressed when complementing the experimental findings by quantum chemical (QC) calculations (91, 92).

Given the enormous importance of the Haber–Bosch process, the study of metal clusters capable of activating dinitrogen so as to synthesize ammonia represents a worthy undertaking (39), not to mention the fundamental problems being addressed. Herein, we describe our findings on the unexpected, mechanistically unique ammonia synthesis from its elements at ambient temperature mediated by the cationic tantalum dimer Ta₂⁺ in the highly diluted gas phase using advanced MS complemented by QC calculations.

Results and Discussion

Ta₂⁺ Cleaves the N≡N Triple Bond to Form Ta₂N₂⁺.

The spectra in Fig. 1 have been obtained by using Fourier transform ion cyclotron resonance (FT-ICR) MS (details are in Experimental Details) and show the results of the reactions of mass-selected Ta₂⁺ ions (m/z = 362) (refs. 93 and 94 have details) with ¹⁴N₂, ¹⁵N₂, and a 1:1 mixture of ¹⁴N₂ and ¹⁵N₂. To properly thermalize the precursor ion Ta₂⁺, it was allowed to interact with pulsed-in argon (ca. 2 × 10⁻⁶ mbar) before reacting with molecular nitrogen. A temperature of 298 K for the thermalized clusters was assumed. Spectra resulting from the reactions with background impurities as well as with argon, serving as an inert substrate, have been recorded as well (Fig. 1A).

Fig. 1.

Mass spectra for the thermal reactions of Ta₂⁺ with Ar (A), ¹⁴N₂ (B), ¹⁵N₂ (C), and a 1:1 mixture of ¹⁴N₂ and ¹⁵N₂ (D) at a pressure of ca. 2.0 × 10⁻⁷ mbar after a reaction time of 2 s. All x axes are scaled in m/z, and the y axes are normalized relative ion abundances.

As displayed in Fig. 1A, when only argon was admitted to the ion cyclotron resonance (ICR) cell, a signal B with Δm = +16 relative to the precursor ion Ta₂⁺ appears; this corresponds to the product ion Ta₂O⁺ generated by reactions with background gases. On leaking N₂ into the ICR cell, in Fig. 1B, a new signal C with Δm = +28 appears, which has been identified as Ta₂N₂⁺ (Eq. 1). By using isotope-labeled ¹⁵N₂, signal C from Fig. 1B is shifted by two mass units on the mass scale and shows up as peak D in Fig. 1C (Ta₂¹⁵N₂⁺) (Eq. 2). If Ta₂⁺ is exposed to a 1:1 mixture of ¹⁴N₂ and ¹⁵N₂, signals for both Ta₂¹⁴N₂⁺ and Ta₂¹⁵N₂⁺ are observed, but there are none containing both nitrogen isotopes Ta₂¹⁴N¹⁵N⁺ (Eqs. 3 and 4). Mass-selected and properly thermalized Ta₂¹⁴N₂⁺, when exposed to ¹⁵N₂, does not react following one of the degenerate exchange reactions 3 and 4:

$${\mathrm{T}{\mathrm{a}}_2}^{\mathrm{+}}+{\mathrm{N}}_2\to {\mathrm{T}{\mathrm{a}}_2{\mathrm{N}}_2}^{\mathrm{+}}$$ [1]

$${\mathrm{T}{\mathrm{a}}_2}^{\mathrm{+}}+{}^{15}\text{N}_2\to \mathrm{T}{\mathrm{a}}_2{{}^{15}\text{N}_2}^{\mathrm{+}}$$ [2]

$$\mathrm{T}{\mathrm{a}}_2{{}^{14}\text{N}_2}^{\mathrm{+}}+{}^{15}\text{N}_2\nrightarrow {\text{T}\mathrm{a}}_2{{}^{15}\text{N}_2}^{\mathrm{+}}+{}^{14}\text{N}_2$$ [3]

$$\mathrm{T}{\mathrm{a}}_2{{}^{14}\text{N}_2}^{\mathrm{+}}+{}^{15}\text{N}_2\nrightarrow {\text{T}\mathrm{a}}_2{}^{14}\text{N}{{}^{15}\text{N}}^{+}+{}^{14}\text{N}{}^{15}\text{N}.$$ [4]

The rate constant k(Ta₂⁺/N₂) for reaction 1 is estimated to 5.1 × 10⁻¹² cm³ molecule⁻¹ s⁻¹; this corresponds to a collision efficiency of ϕ = 0.8%. Owing to the uncertainty in the determination of the absolute N₂ pressure, an error of ±30% is associated with these measurements. In addition to the labeling experiments, the elementary compositions of the charged particles have been confirmed by exact mass measurements. Since ¹⁴N/¹⁵N kinetic isotope effects (KIEs) are expected to be quite small (95) and as it was not possible to reproducibly adjust the pressure in the ICR cell to the required accuracy to obtain meaningful data, we have renounced the determination of the ¹⁴N/¹⁵N KIE for the formation of Ta₂N₂⁺.

The simplified 2D potential energy surface (PES) of the most favorable pathway as well as selected structural parameters of key species (Fig. 2) reveals insight into the mechanism of the Ta₂⁺-mediated activation of the N≡N triple bond at a molecular level. A Fortran-based genetic algorithm (96) to generate initial guess structures of Ta₂N₂⁺ followed by optimizations at the level of the Becke-3–Lee–Yang–Par functional including the def2–triple-zeta valence basis set with one set of polarization functions (B3LYP/def2-TZVP) could only identify the intermediates 1–3 as the most stable species. (Additional computational details are provided in SI Appendix.)

Fig. 2.

Simplified PES (ΔH_298K) for the reactions of Ta₂⁺ with N₂. The calculations were done at the B3LYP/def2-QZVPP//B3LYP/def2-TZVP level of theory. Key ground-state structures with selected geometric parameters are also provided. Charges are omitted for the sake of clarity, bond lengths are given in Å, and relative energies are in kJ mol⁻¹.

Dinitrogen approaches the positively charged Ta₂⁺, which in its ground state, is a doublet (²Δ_g) (97), through the known side-on/end-on binding mode (μ-η¹:η²-N₂) (34, 63, 65–67) to form intermediate 1 (−93 kJ mol⁻¹, C_s symmetry) in a barrier-free process. The end-on bonded nitrogen atom (N^a) in 1 has an N^a−Ta^a bond length of only 1.82 Å, while Ta^b binds side on to N^a (2.08 Å) and end on to N^b (1.90 Å). The N–N and Ta–Ta bond distances are elongated by 0.23 and 0.31 Å, respectively, compared with the isolated reactants; thus, on interacting with Ta₂⁺, the N≡N triple bond is already weakened, but the N₂ unit as such is still intact. These geometric properties closely resemble those reported for the crystal structure of ([NPN]Ta(μ-H))₂N₂ (65). During the next step, atom N^b, while still connected to Ta^b, approaches Ta^a, eventually binding to Ta^a via transition state 1/2 (−55 kJ mol⁻¹). This leads to the formation of intermediate 2 (−104 kJ mol⁻¹) having C_2v symmetry; 2 displays a double side-on (μ-η²:η²-N₂) “butterfly” geometry with an already significantly elongated N–N bond distance amounting to 1.57 Å, clearly indicating additional activation of the N₂ molecule. Finally, the remaining bonding interaction between N^a and N^b is disrupted entirely by passing through the low barrier of 2/3 (−94 kJ mol⁻¹), thus giving rise to the global minimum 3 (−453 kJ mol⁻¹); 3 as well as its neutral counterpart (98) exhibit a slightly distorted planar square with D_2h symmetry. It has a cyclic structure consisting of alternating Ta and N atoms. N^a and N^b in 3 are 2.56 Å apart from each other, indicating that the direct interaction between the two N atoms is negligibly small; this is confirmed by a Mayer bond order (99, 100) of less than 0.1. The energy requirements for dissociating 3 into various couples of, for example, TaN⁺/TaN, Ta₂N⁺/N, TaN₂⁺/Ta, and Ta⁺/TaN₂ are located 247, 288, 416, and 409 kJ mol⁻¹, respectively, above the entrance level. To enable these reactions, external energy must be supplied, for instance, by collisional activation (101–104). Without external energy supply, intermediate 3 can only return to the reactants, or its lifetime can be increased by IR photon emission (105) or collisional cooling (106).

To further substantiate the claim that, in Ta₂N₂⁺, the binding interaction between the two nitrogen atoms originally tied together by a triple bond has indeed completely disappeared, we considered the following. If it were 1 that had been generated, a degenerate exchange of the N₂ unit, according to Fig. 3, should be possible, as all relevant species are located well below the entrance asymptote. However, this computational finding is in conflict with the experimental results (Eq. 3). Thus, the experimentally generated Ta₂N₂⁺ does not have structure 1. Furthermore, 2 is also not likely to be long lived along the reaction coordinate. As an isolated species, it cannot dissipate its internal energy of ca. 104 kJ mol⁻¹ and will rather easily surmount transition state 2/3 to form the global minimum 3. For this ion, in agreement with the experiments (Eq. 4), extensive calculations (SI Appendix, Fig. S1) reveal that, for the exchange reactions with N₂, prohibitive barriers are encountered. Finally, collisional activation (101–104) of Ta₂N₂⁺ with argon leads to the loss of N₂ only at rather high excitation energies [E_(coll.,CM) > 4.8 eV]. This indicates that strong chemical bonds must exist between Ta₂⁺ and N₂. Although the direct dissociation of 3 to TaN⁺/TaN by cycloreversion may have an entropic advantage, obviously this process cannot compete with the energetically favored multistep dissociation back to the starting reactants (3 → → → Ta₂⁺/N₂).

Fig. 3.

Simplified PES (ΔH_298K) for the degenerate exchange reactions of 1 with N₂. Details are in Fig. 2.

Thus, the cleavage of the N≡N triple bond of N₂ by Ta₂⁺ forms 3 and proceeds via the rate-limiting transition state 1/2. As the rate-limiting transition state 1/2 is located below the entrance asymptote, activation of N₂ is accessible at ambient temperature.

Establishing a Working Hypothesis—a Brief Detour.

Recently, Arakawa et al. (107) have shown that Ta₂N₂⁺ can also be generated in the gas phase by the reaction of Ta₂⁺ with NH₃ at 298 K (Eq. 5). In our experiments, we found that Ta₂⁺ reacts with N₂ to form Ta₂N₂⁺ as well (Eq. 1). Combining the two reactions (Eq. 6) represents the reverse of the Haber–Bosch ammonia synthesis (Eq. 7). If the structures of the key intermediates Ta₂N₂⁺ generated in reactions 1 and 5 are identical, the principal of microscopic reversibility requires that there must be a way by which N₂ and H₂, mediated by Ta₂⁺, can form NH₃ at ambient temperature (Eq. 7), since the reactions 1, 5, and 7 are exothermic (108):

$${{\mathrm{T}\mathrm{a}}_2}^{+}+\mathrm{2N}{\mathrm{H}}_{\mathrm{3}}\to {\mathrm{T}\mathrm{a}}_2{{\mathrm{N}}_2}^{\mathrm{+}}\mathrm{+}{3\mathrm{H}}_2$$ [5]

$$\mathrm{2N}{\mathrm{H}}_{\mathrm{3}}\to {\mathrm{N}}_2+\mathrm{3}{\mathrm{H}}_2$$ [6]

$${\mathrm{N}}_2\mathrm{+3}{\mathrm{H}}_2\to \mathrm{2N}{\mathrm{H}}_{\mathrm{3}}.$$ [7]

Crucial Intermediates Along the N₂ ⇄ NH₃ Reaction Coordinates.

Next, we consider computationally (Fig. 4) the elementary steps associated with the Ta₂⁺-mediated reactions in Eq. 7. It is important to stress that these processes turned out to be mechanistically rather complex, and as they do not form the main target of our investigations, they are not discussed here in detail. Furthermore, we are well aware that most likely not all conceivable intermediates and sideways have been included that could be involved in the ammonia synthesis from molecular nitrogen and hydrogen mediated by Ta₂⁺. However, the intermediates and transition states, shown in Fig. 4 (details are in SI Appendix, Fig. S6), already form a feasible road map to the formation of NH₃ out of N₂ and H₂ mediated by Ta₂⁺ at room temperature.

Fig. 4.

Simplified PES (ΔH_298K) for the reactions of Ta₂⁺ with one molecule of N₂ and three molecules of H₂ and the reverse process Ta₂⁺ + 2NH₃ → Ta₂N₂⁺ + 3H₂. Details are in Fig. 2.

As the key step, a reaction with N₂ takes place followed by the sequential uptake of two H₂ molecules. After liberation of the first NH₃ molecule and subsequent uptake of a third H₂ molecule, the second NH₃ molecule is released, and Ta₂⁺ is regenerated to be ready for another catalytic cycle. It is quite impressive to note how stable some of the intermediates are! For example, the global minimum (8, −690 kJ mol⁻¹) is arrived at after the uptake of two H₂ molecules by Ta₂N₂⁺. Ta₂N₂H₄⁺, after several intramolecular isomerizations, finally splits off one molecule of NH₃. The remaining intermediate 14, Ta₂NH⁺, is able to catch another dihydrogen molecule. After a series of additional hydrogen migrations toward the NH group, a second NH₃ is formed and finally, liberated. Note that the generation of the terminal products Ta₂⁺ and NH₃ corresponds to the rate-determining step for the whole Ta₂N₂⁺ hydrogenation–denitrogenation reactions; however and importantly, as it lies below the reaction entrance, a “Haber–Bosch” process transpires at room temperature. The theoretically obtained value for the heat of formation of two NH₃ molecules resulting at the end of the catalytic cycle shown in Fig. 4 amounts to −101 kJ mol⁻¹ and compares well with the experimental one (−91.8 kJ mol⁻¹) (108).

To test experimentally the QC predictions of this catalytic cycle, thermal reactions of Ta₂⁺ with NH₃ were examined. As can be seen from Fig. 5A, Ta₂⁺ on reacting with NH₃ gives rise to Ta₂NH⁺ as the main product (signal E), with Ta₂N₂⁺ also emerging as a weak signal C; the reaction efficiency amounts to ca. 90%. This confirms the results already obtained by Arakawa et al. (107). Next, when Ta₂NH⁺ is mass selected and further reacted with another molecule of ammonia, the only product obtained is Ta₂N₂⁺(ϕ ≈ 0.95), thus connecting 3 with 14 (Fig. 4). Starting the reaction from the product side has the benefit that part of the excess energy gained in the course of the exothermic reaction steps can be dissipated by the loss of neutral hydrogen molecules, thus increasing the lifetime of the remaining charged counterparts. We have also performed numerous experiments aimed at obtaining additional details for the forward reaction (i.e., the system Ta₂⁺/N₂/H₂). However, these experiments were not conclusive, most likely due to limited sensitivity and a lack of a sufficiently long lifetime of the intermediates. Nevertheless, the combined experimental/computational findings show the existence of a catalytic room temperature cycle in the conversion of N₂/H₂ to NH₃.

Fig. 5.

Mass spectra for the thermal reactions of (A) Ta₂⁺ with NH₃ at an NH₃ pressure of ca. 2.0 × 10⁻⁹ mbar after a reaction time of 2 s and (B) Ta₂NH⁺ with NH₃ at an NH₃ pressure of ca. 2.8 × 10⁻⁹ mbar after a reaction time of 1 s. The x axes are scaled in m/z, and the y axes are normalized relative ion abundances.

Brief View on the Reaction of Ta₂⁺ with H₂.

Although less exothermic than in the reaction with N₂, as shown computationally (Fig. 6), Ta₂⁺ also is able to react with H₂, and the rate-determining step corresponds to transition state 6.1/2 (−23 kJ mol⁻¹). In a concerted, almost barrier-free way, the C_2v-symmetric, butterfly-shaped intermediate 6.2 (−134 kJ mol⁻¹) is generated, which strongly resembles structure 2 in the Ta₂N₂⁺ system; 6.2 is able to react further with another molecule of H₂ (SI Appendix, Fig. S2). In contrast to the Haber–Bosch process where H₂ poisoning constitutes an important issue (109), in this system, for all clusters Ta₂(H₂)_x⁺ (x = 1, 2, 3) investigated, N₂ is able to displace molecular hydrogen and thus, suppress H₂ poisoning of the catalyst as shown in SI Appendix, Figs. S3 and S5.

Fig. 6.

Simplified PES (ΔH_298K) for the reactions of Ta₂⁺ with H₂. Details are in Fig. 2.

Closer Inspection of the Mechanism of the N≡N Triple-Bond Cleavage by a Frontier Orbital Analysis.

To obtain a more detailed mechanistic insight into the splitting of the N≡N triple bond as the key step of the whole reaction sequence, a frontier orbital analysis for the rate-determining step has been performed (110–117). Fig. 7 shows the detailed evolution of the electronic structures along the reaction coordinates.

Fig. 7.

Schematic orbital diagrams based on a frontier orbital analysis. Only representative orbitals are shown. The structures are oriented such that the x axis is along the Ta–Ta vector and that the y axis is confined to the Ta₂N₂ plane. The π_v- and π_v*-orbitals of the dinitrogen molecule are vertical (v) to this plane, whereas the π_p- and π_p*-orbitals are lying within this plane. The magenta borders refer to π-backdonation; the green ones represent metal centers that accept electron density from N₂. The more intense the color, the more electron density is transferred.

According to Fig. 7, the multiple bonding in the cationic tantalum dimer, having a doublet ground-state Ta₂⁺ (6s²5d⁷), comprises two doubly occupied σ-bonds [σ_(6s-6s) and σ_(dx2-dx2)], two doubly occupied π-bonds [π_(dxz-dxz) and π_(dxy-dxy)], and a singly occupied δ-bond [δ_(dyz-dyz)]. In 1, the approaching N₂ molecule lies in the same plane as the metal dimer, and the strong interaction between these two fragments mainly results from the π-backdonation from the metal centers to N₂ (i.e., electron density relocates from the metal–metal π-bonding to the vacant antibonding orbitals of the N₂ ligand); as a consequence, weakening of the N≡N triple bond occurs to a certain extent. In return, although relatively small, donation of electron density from the N–N π-bonds to the vacant metal–metal orbitals can be found as well. On its way to 2, the N₂ molecule gradually rotates perpendicular to the Ta–Ta axis while simultaneously adjusting the metal–metal orbitals. After conversion to 2, only a small barrier of 10 kJ mol⁻¹ precludes the four atoms to be trapped in the deep potential well to form 3. This process is accompanied by both π-backdonation from the metal dimer to the σ*- and π*-orbitals of N₂ and partial donation of bonding electrons from N₂ to the metal dimer. It is this electronic reorganization that eventually leads to the complete rupture of the N≡N triple bond, which is accompanied by generating strong Ta–N bonds in 3. As shown in Fig. 7, the orbital components of the antibonding orbitals of N₂ are dominant in π*_(dxz/dxz)_π_v*, π_(dxz/dxz)_σ*, and δ_(dyz/dyz)_π_v* in 3; once again, this clearly indicates the cleavage of the N≡N triple bond. Furthermore, the components of all three N₂-based bonding orbitals greatly shrink in going from 2 to 3, sharing their electrons now within the four-membered ring. In sharp contrast, for the cleavage of the single bond of a dihydrogen molecule by Ta₂⁺, the frontier orbital interaction mainly concerns the π_(dxy/dxy)-orbital of the metal dimer center and the antibonding σ*_(H–H)-orbital of H₂ (SI Appendix, Fig. S8).

An even deeper understanding of the intrinsic reactivity can be obtained by a comparison of the “naked” cationic tantalum dimer with the ligated ditantalum core in the ([NPN]Ta(μ-H))₂N₂ complex (65). The cationic Ta dimer is characterized by a Ta–Ta multiple bond with a Wiberg bond index (WBI) of 4.5. In contrast, in the ligated complex, the WBI amounts to only 1.7 due to significant binding interactions between the metal atoms and the coordinating N and P atoms of the ligands and the two bridging hydrido ligands. SI Appendix, Fig. S7 displays representative frontier orbitals dominating the metal atoms inside the complex. Only two doubly occupied orbitals are left. As the metal–metal σ-bonding orbital is not a good electron donor, the π-backdonation is expected to be rather small. More importantly, these orbitals are highly localized at one of the two metal centers, and their lobes are leaning toward the ligands; in addition, they are not regarded as potential electron acceptors, as an optimal overlap is difficult to achieve. Thus, the reactivity of the cationic ditantalum cluster toward dinitrogen can be attributed to the interplay of empty and doubly occupied d-orbitals at the metal center, which on the one hand, accepts electrons from N₂ and on the other hand, weakens the N≡N triple bond further by π-backdonation of electron density from the metal center into the antibonding orbitals of N₂ (50).

Conclusion

As shown above, our working hypothesis about a tantalum-mediated coupling of N₂ and H₂ finally has been confirmed by gas-phase experiments and QC calculations. Here, we describe the concept of an ammonia synthesis from molecular N₂ and H₂ catalyzed by a bare metal cluster cation at ambient temperature (118). The key step consists of the complete rupture of the N≡N triple bond, rendered possible by the interplay of vacant and doubly occupied d-orbitals at the ditantalum center of Ta₂⁺ to form Ta₂N₂⁺ as the central intermediate. This combined experimental/computational study further improves our knowledge about mechanistic details of the catalytic action of transition metals and emphasizes the crucial role that electron-donating and -accepting orbitals play (49, 50). These findings might serve as a base to improve or even invent “real world” catalysts to save economic and ecologic resources in the future (73).

Materials and Methods

Experimental Details.

The ion/molecule reactions were performed with a Spectrospin CMS 47X FT-ICR mass spectrometer equipped with an external ion source as described elsewhere (119–121). In brief, Ta₂⁺ was generated by laser ablation of a tantalum disk using a neodymium-doped yttrium aluminum garnet (Nd:YAG) laser operating at 532 nm and seeded in helium; the latter serves as a cooling and carrier gas. Using a series of potentials and ion lenses, the ions were transferred into the ICR cell, which is positioned in the bore of a 7.05-T superconducting magnet. To properly thermalize the precursor ion Ta₂⁺, it was allowed to take a bath in pulsed-in argon (ca. 2 × 10⁻⁶ mbar) before its reaction with dinitrogen. After thermalization, the reactions of mass-selected Ta₂⁺ were studied by introducing isotopologues of dinitrogen (N₂ and ¹⁵N₂) and ammonia via leak valves at stationary pressures. A temperature of 298 K for the thermalized clusters was assumed. Before the exchange reactions between unlabeled and ¹⁵N-labeled N₂, the precursor ion Ta₂N₂⁺ was thermalized by pulsed-in argon (ca. 2 × 10⁻⁶ mbar). In the collision-induced dissociation (102–104) experiments, mass-selected, properly thermalized Ta₂N₂⁺ ions were reacted with argon.

Computational Details.

The unbalanced treatment of static and dynamic correlation makes the transition metal chemistry hard to handle for any density functional. We, therefore, carried out extensive investigations, even using multireference perturbation theory [like the n-electron valence state perturbation theory (NEVPT2)] (122–124) as well as a comparison of many density functionals before deciding to base our conclusions on B3LYP calculations (125–128). The geometry optimization was conducted at the B3LYP/def2-TZVP level of theory. Subsequently, the electronic energies were refined by using the B3LYP functional with def2–quadruple-zeta valence basis set and two sets of polarization functions (B3LYP/def2-QZVPP//B3LYP/def2-TZVP) (129).

Additional information with regard to the computational details can be found in SI Appendix.