Nb2BN2 cluster anions reduce four carbon dioxide molecules: reactivity enhancement by ligands

Hai-Yan Zhou , Ming Wang , Yong-Qi Ding and Jia-Bi Ma *
Key Laboratory of Cluster Science of Ministry of Education, Beijing Key Laboratory of Photoelectronic/Electrophotonic Conversion Materials, School of Chemistry and Chemical Engineering, Beijing Institute of Technology, 100081, Beijing, People's Republic of China. E-mail: majiabi@bit.edu.cn

Received 30th July 2020 , Accepted 2nd September 2020

First published on 3rd September 2020

The thermal gas-phase reactions of Nb2BN2 cluster anions with carbon dioxide have been explored by using the art of time-of-flight mass spectrometry and density functional theory calculations. Four CO2 molecules can be consecutively reduced by Nb2BN2, resulting in the formation of Nb2BN2O1–4 anions and the release of one CO molecule each time. To illustrate the role of ligands in Nb2BN2, the reactivities of Nb2N2 and Nb2B toward CO2 were also investigated; two and three CO2 molecules are activated, respectively, and the rate constants are slower than that of Nb2BN2/CO2 systems. This comparison indicates that metal–metal multiple bonds and appropriate ligands, such as B, are important factors for CO2 reduction. The synergy between a transition metal atom (Nb) and a main-group atom (B) in CO2 reduction mediated by gas-phase clusters is revealed for the first time. To the best of our knowledge, Nb2BN2 anions are gas-phase clusters that reduce the largest number of CO2 molecules. A fundamental understanding of the efficient reduction of carbon dioxide molecules may shed light on the rational design of active sites on supported transition metal/boron nitride catalysts.


As an abundant, nontoxic, easily available, and renewable carbon source, CO2 reduction is of utmost importance to achieving energy and environmental goals.1 Because CO2 is the most oxidized state of carbon and the reaction CO2 → CO + O is 526.15 kJ mol−1 endothermic at 0 K, its transformation under ambient conditions is still quite challenging. The transitioning of CO2 to CO occupies a pivotal stage, and CO2 could be converted to various liquid fuels (gasoline, diesel, and so on) by hydrogenation of the CO intermediate. For instance, reverse water gas shift (RWGS, CO2 + H2 → CO + H2O) reaction is one of the most promising processes for CO2 conversion from both fundamental and practical points of view.2,3 In the CO2 reduction reactions, transition-metal-based oxide catalysts display exceptional reactivity,2–5 and one of the proposed mechanisms for CO formation is via a redox mechanism of active metal centers (CO2 + Mn+ → MOn+ + CO).4 In addition, boron nitride (BN) materials, as an analogue of graphene, have gained increasing attention because of their inherent advantages such as high chemical and thermal stabilities.6–8 Transition metal/BN nanomaterials have been explored for several catalytic reactions.9–13 The future development of new catalysts with high efficiency and reactivity for CO2 conversion under ambient conditions requires a fundamental understanding of the structure–reactivity relationships of active sites and reaction mechanisms.

Gas-phase studies provide an ideal arena for investigating elementary steps at a strictly molecular level, such as CO2 reduction.14–23 Among the reported three fundamental reactions of CO2 transformation in the gas phase,19,24–30 most of the ions or clusters that can perform oxygen-atom transfer with the generation of CO are positively charged, while the negatively charged species, being reactive toward CO2, are quite limited,17 including MoxOy, WxOy,31 MoxWyOz,32 Nin,33 and ReO2.27 The channel of CO2 adsorption was also observed in other reported anion systems.18,20,33 All these reactive cations and anions can reduce at most one CO2 molecule. Quite recently, Schwarz and co-workers reported that the V2+ cations can consecutively react with two CO2 molecules, liberating one CO molecule each time.23 Based on the reported reactions of the reduction of CO2 to CO by metal-containing systems, the oxygen-atom affinities of transition metals, the spin changes associated with the oxidation of M+ to MO+, and the oxidation numbers of the reactive ions are important factors.19

Herein, we report that niobium nitride anions doped with a single boron atom, Nb2BN2, can surprisingly promote four CO2 reductions, producing the final oxygen-rich product Nb2BN2O4. In contrast, the homonuclear niobium nitride clusters Nb2N2 and niobium boride anions Nb2B only activate two and three CO2 reductions, respectively, under similar reaction conditions. Therefore, the presence of appropriate ligands influences the reactivity dramatically. To the best of our knowledge, Nb2BN2 is the cluster that reduces the largest number of CO2 molecules and also exhibits quite a high reactivity in these kinds of reactions.


Experimental methods

Nb2BN2 ions are generated by laser ablation of a rotating and translating NbB2/Nb disk target (molar ratio: NbB2[thin space (1/6-em)]:[thin space (1/6-em)]Nb = 1[thin space (1/6-em)]:[thin space (1/6-em)]5) in the presence of about 20% N2 seeded in a He carrier gas with a backing pressure of 4 atm. NbB2 and Nb powder were pressed to the disk, and the mechanical stability is sufficient for the procedures described below. The cluster ions of interest were mass-selected by a quadrupole mass filter (QMF) and entered into a linear ion trap (LIT) reactor, where they were confined and thermalized by collisions with a pulse of He gas and then interacted with a pulse of CO2 or C18O2 for a period of time. The instantaneous gas pressure of He in the reactor was around 1.6 Pa. It has been proved that clusters are thermalized to (or close to) room temperature before reactions in the previous works.34,35 A reflectron time-of-flight mass spectrometer was used to detect the cluster ions ejected from the LIT. Eqn (4) was used to estimate the pseudo-first-order rate coefficient:
image file: d0dt02680h-t1.tif(4)
in which IR is the intensity of the reactant cluster ions after the reaction, IT is the total ion intensity including product ion contribution, T is the temperature, Peffective is the effective gas pressure in the trap, tR is the reaction time, and k is the Boltzmann constant. More details about the method to derive k1 can be found in ref. 36.

Theoretical methods

All DFT calculations were performed using the Gaussian 0937 program package employing the hybrid M06L exchange–correlation functional.38–40 For all the reaction pathways, the def2-TZVPD basis sets41 were used for the Nb atom, and the aug-cc-pVTZ basis sets42 were selected for the B, N, C, O atoms. This level of theory with a moderate computational cost is tested to give reasonably good results for Nb2BN2 systems. Among the 20 tested methods, M06L gives the best interpretation of Nb–C, Nb–O, Nb–Nb, B–O, C–O and N–O bond-dissociation energies, as listed in Table S1, ESI. A Fortran code based on a genetic algorithm43 was used to generate the initial guess structures of Nb2BN2O0–4, Nb2N2O0–2, and Nb2BO0–3. The reaction mechanism calculations involve the geometrical optimization of reaction intermediates (IMs) and transition states (TSs). Vibrational frequency calculations were performed to ensure that the IMs and TSs have zero and only one imaginary frequency, respectively. Intrinsic reaction coordinate (IRC)44,45 calculations were carried out to make sure that a TS connects two appropriate minima. The reported energies (ΔH0 K in eV) are corrected with zero-point vibrations. Natural bond orbital (NBO) analysis was performed using NBO 6.046 implemented in Gaussian 09, and the program Multiwfn47 was employed to analyze orbital compositions by a natural atomic orbital method.


Experimental results

The time-of-flight (TOF) mass spectra for the reactions of laser-ablation-generated, mass-selected, and thermalized Nb2BN2 clusters with CO2 are given in Fig. 1. On the reactions with CO2, new signals of Nb2BN2O, Nb2BN2O2 and Nb2BN2O3 increase (Fig. 1b and c) as increasing CO2 pressures in the reactor. When the pressure of CO2 is further increased (Fig. 1d), the reactant Nb2BN2 and the first product peaks Nb2BN2O disappear; while the peak corresponding to Nb2BN2O3 has the largest intensity, and one new signal, which can be assigned to Nb2BN2O4, shows up. These results provide convincing evidence that Nb2BN2 can oxidize four CO2 molecules consecutively (eqn (1a)–(d)). The pseudo first-order rate constants (k1) for the reactions of Nb2BN2 with CO2 (reactions (1a)–(d)) were estimated on the basis of a least-squares fitting procedure (Fig. 2)36 and are (5.3 ± 1.1) × 10−10 cm3 molecule−1 s−1, (2.7 ± 0.5) × 10−10 cm3 molecule−1 s−1, (8.9 ± 1.8) × 10−11 cm3 molecule−1 s−1 and (1.3 ± 0.3) × 10−11 cm3 molecule−1 s−1, corresponding to the reaction efficiencies (Φ) of 80%, 41%, 14%, and 2%, respectively. Also, we can draw the conclusion that Nb2BN2 and Nb2BN2O1–4 have no unreactive components according to Fig. 1d and the kinetic fits (Table 1). These reaction channels were further confirmed by using the isotopic labeling experiments with C18O2 (Fig. 1e and f).
image file: d0dt02680h-f1.tif
Fig. 1 Time-of-flight (TOF) mass spectra for the reactions of mass-selected Nb2BN2 with He (a), CO2 (b–d) for 2 ms, and C18O2 (e) for 5.9 ms and (f) for 13.9 ms, and mass-selected Nb2N2 (g) as well as Nb2B (i) with CO2 (h and j). The time periods for the reactions were 2.8 ms in (g) and (h) and 8.4 ms in (i) and (j). The weak Nb2N2B16O18O, Nb2N2B16O18O2 and Nb2N2B16O18O3 signals in (e) and (f) are due to the scrambling of the 18O atom and the 16O atom in residual water. Long reaction times render these products present. More spectra are shown in Fig. S1 and S2 for the Nb2N2/CO2 and Nb2B/CO2 systems.

image file: d0dt02680h-f2.tif
Fig. 2 Variations of the relative intensities of the reactant and product anions in the reactions of Nb2BN2 with CO2 with respect to the CO2 pressures for 2.7 ms. The solid lines are fitted to the experimental data points by using the equations derived with the approximation of the pseudo-first-order reaction mechanism.
Table 1 Reaction equations, pseudo-first-order rate constants (in unit of 10−11 cm3 molecule−1 s−1) and reaction efficiencies for the reactions investigated
  Reactions   k 1 Φ/%
1 Nb2BN2 + CO2 → Nb2BN2 + CO (1a) 53 ± 11 80
Nb2BN2O + CO2 → Nb2BN2O2 + CO (1b) 27 ± 5.0 41
Nb2BN2O2 + CO2 → Nb2BN2O3 + CO (1c) 8.9 ± 1.8 14
Nb2BN2O3 + CO2 → Nb2BN2O4 + CO (1d) 1.3 ± 0.3 2
2 Nb2N2 + CO2 → Nb2N2O + CO (2a) 8.8 ± 1.8 13
Nb2N2O + CO2 → Nb2N2O2 + CO (2b) 4.0 ± 0.8 6
3 Nb2B + CO2 → Nb2BO + CO (3a) 1.6 ± 0.3 2
Nb2BO + CO2 → Nb2BO2 + CO (3b) 0.8 ± 0.17 1
Nb2BO2 + CO2 → Nb2BO3 + CO (3c) 0.06 ± 0.013 0.1

If the boron atom or the two nitrogen atoms are absent in Nb2BN2, that is Nb2N2 or Nb2B, only two or three CO2 molecules are reduced, respectively (Fig. 1h, j, and Fig. S1 and S2, ESI); also, the rate constants of Nb2N2 and Nb2B are 6 and 33 times slower than that of Nb2BN2, respectively. When Nb2B anions accept three oxygen atoms from three CO2 molecules, the reactivity of the products Nb2BO3 toward CO2 decreases dramatically, and no further oxygen-atom-transfer product is observed in our apparatus, even at a relatively high pressure of CO2 (up to 225 mPa) and a rather long reaction time (up to 8.6 ms). The significance of the ligands B and N atoms in the Nb2BN2 anions is quite obvious.

Theoretical results

The reaction pathways of Nb2BN2 with CO2. To investigate the key role of the boron atom and the intrinsic properties of these clusters as well as mechanistic details of the reactions, density functional theory (DFT) calculations were performed. As shown in Fig. 3a and S3, the lowest-energy structure of Nb2BN2 (IA1) has a planar C2v geometry, consisting of one Nb–Nb multiple bond, and the B atom bridges the two N atoms. The energy difference between the singlet state 1IA1 and the triplet state 3IA1 is 0.02 eV. Compared with these two structures, the major distinction is the bond order of Nb–Nb bond, and the Wiberg bond indexes48,49 (WBIs) amount to 3.39 and 2.43 for 1IA1 and 3IA1, respectively. The orbital analysis indicates that due to the elongated bond length of Nb–Nb bond in 3IA1, the highest occupied molecular orbital−1 (HOMO−1) with paired electrons in 1IA1 is divided into two orbitals containing unpaired d electrons of Nb atoms; most of the other orbitals are the same (Fig. S4).
image file: d0dt02680h-f3.tif
Fig. 3 (a) DFT-calculated structures and relative energies of Nb2BN2, Nb2N2 and Nb2B at the M06L/def2-TZVPD and aug-cc-pVTZ level of theory. The point group is given under each structure. Bond lengths are given in pm. (b) DFT-calculated potential energy surfaces for the reactions of Nb2BN2 with CO2. The zero-point vibration corrected energies (ΔH0 K in eV) of the reaction intermediates, transition states, and products with respect to the separated reactants are given. Superscripts denote the spin multiplicities.

The DFT-calculated singlet potential energy surface (PES) for reaction (1a) is shown in Fig. 3b. An insertion–elimination mechanism is operative. In detail, CO2 approaches the Nb–Nb bond site to form intermediate I1, and the CO2 molecule is bent from its linear configuration in a free CO2 molecule to a ∠OCO angle of 134° in I1. Meanwhile, both C–O bonds are elongated to 127 pm from their gas phase value of 117 pm. ViaTS1, one C–O bond in CO2 is broken, resulting in a Nb–Ot (t = terminal) bond and a CO unit in I2. Then, I2 is converted into I3 by transforming the Ot in I2 into an Ob atom (b = bridge) in I3. By overcoming TS3, intermediate I4 is generated; subsequent liberation of CO produces the product Nb2BN2O. In the following consecutive reactions, three more CO2 molecules are reduced; Nb2BN2O2, Nb2BN2O3, and Nb2BN2O4 are produced in turn, as shown in Scheme 1. Full details of structures and PESs are given in Fig. S5. The subsequently added three additional oxygen atoms in Nb2BN2O2–4 are bonded as Ot, Ob, and Ot, respectively. As the oxygen atoms increase, the Nb–Nb multiple bonds are elongated from 229 pm in 1IA1 to 279 pm in Nb2BN2O4 (1P4); at the same time, the bond lengths of Nb–N and Nb–B are also increased. This tendency can be reflected from the decreased WBIs for Nb–N bond, Nb–N and Nb–B bonds, as given in Table 2. Along the whole reactions (1a)–(d), the five atoms in the [Nb2BN2] unit are almost in a plane. The final product Nb2BN2O4 is with C2v symmetry, and two Ot as well as two Ob atoms are bonded with two Nb atoms; the electronic ground state is singlet. The triplet PESs for the reactions of Nb2BN2 with CO2 have also been explored (Fig. 3b); the singlet and triplet PESs are similar, and it turns out that the CO2 reduction in the triplet state is also achievable. If the singlet and triplet isomers of Nb2BN2 coexist in the cluster source, they both can react with CO2 efficiently and there is no inert component, which is consistent with the experimental data. In Nb2BN2, another isomer, with one N atom bridgingly bonded with two Nb atoms (3IS2 in Fig. 3a and S6), may also exist since this isomer is 0.18 eV higher in energy than the ground-state 3IA1, based on the DFT calculations, and its reactions toward CO2 molecules can occur, based on the DFT calculations (Fig. S7). No matter which structure of Nb2BN2 (3IA1, 1IA1, or 3IS2 in Fig. 3a) is reacted with CO2, the final product is the singlet Nb2BN2O4 (1P4 in Scheme 1), which is uniform in structure and in the spin state.

image file: d0dt02680h-s1.tif
Scheme 1 Schematic diagram of the DFT calculated key structures along the PESs for the (a) Nb2BN2O0–3/CO2, (b) Nb2N2O0,1/CO2 and (c) Nb2BO0–2/CO2 systems. The superscripts denote the spin multiplicities.
Table 2 WBI for selected bonds
  Nb–Nb Nb–N Nb–B B–N
Nb2BN2 (1IA1) 3.39 1.33 0.43 1.31
Nb2BN2O (1P1) 2.78 1.16 0.25 1.41
Nb2BN2O2 (1P2) 1.12 1.34 0.35 1.47
Nb2BN2O3 (1P3) 0.34 1.20 0.24 1.46
Nb2BN2O4 (1P4) 0.05 1.04 0.09 1.47
Nb2N2 (1IA2) 0.68 1.45
Nb2N2O (1P5) 0.35 1.02
Nb2N2O2 (1P6) 0.14 0.64
Nb2B (2IA3) 3.91 1.62
Nb2BO (2P7) 2.90 1.45
Nb2BO (2P8) 4.42 0.54
Nb2BO2−2(2P9) 3.28 0.57
Nb2BO3 (2P10) 3.37 0.42

The reaction pathways of Nb2N2 and Nb2B with CO2. Fig. 1a shows the lowest-energy structures for Nb2N2 and Nb2B. Scheme 1 shows the schematic diagram of reactions (2a), (2b) and (3a)–(c). The detailed PESs are given in the ESI. As shown in Scheme 1b and Fig. S8, the PESs for the reactions of Nb2N2 with CO2 (reactions (2a) and (2b)) commence with the formation of encounter complexes I8 and I9 by coordinating the incoming CO2 molecules to one of the Nb atoms in Nb2N2 successively. Subsequently, one O–C bond in the adsorbed CO2 moiety is broken via a transition state. Accompanied by the release of one CO molecule each time, the intermediate product Nb2N2O and the final product Nb2N2O2 are generated.

The reaction pathways of Nb2B with CO2 are given in Scheme 1c and Fig. S9. There are two coordination modes for the encounter complexes. One mode is the C atom in CO2 coordinates with two Nb atoms (I10, path I in Scheme 1c), and the other mode involves the C and one O atoms in CO2, which interact with the B and one Nb in Nb2B, respectively (I12, path II). These two intermediates lead to different structures of Nb2BO (P7 and P8), but to the same Nb2BO2 (P9). For the third CO2 activation reaction, CO2 is adsorbed by the B atom and the other Nb site (I14), leading to the final product Nb2BO3 (P10). All the product structures given in Scheme 1, except for P3, (P1, P2, and P4–P10), obtained from the reactions are the same with the DFT-calculated lowest-lying structures shown in Fig. S6, S10 and S11, ESI. For the reaction of Nb2BN2O2 with CO2, both of 1P3 (Ot–Nb–(Ob)2–Nb) and 1P3′ (Ot–Nb–Ob–Nb–Ot, the lowest-energy structure of Nb2BN2O3, IS44 in Fig. S6) may co-exist as products, since 1P3 is only 0.25 eV higher in energy than 1P3′, based on DFT calculations (Fig. S5); the former anions react with the fourth CO2 molecule much faster than the latter system because of the existence of one exposed Nb active site to catch the incoming CO2.


The importance of M–M multiple bond

When the B atom is removed from the cluster, the reactivity of Nb2N2 toward CO2 is decreased dramatically, which is reflected by the reduced number of CO2 molecules and the slower rate constants. It is certain that the boron atom in Nb2BN2 is important, and finding out the key role of the B atom is quite necessary to reveal the high reactivity of Nb2BN2. The major differences arising from the modification of the B atom consist of two aspects: (1) From the structural point of view, the presence of one B atom in Nb2BN2 shortens the Nb–Nb bond length (229 pm for the singlet 1IA1) remarkably, compared to that in Nb2N2 (273 pm for 2IA2); therefore, more electrons can be stored in the Nb–Nb bond of the former system and used in the following CO2 reduction reactions. The changes in bond lengths and the values of WBIs (Table 2) along the reaction coordinates indicate that the electrons stored in the Nb–Nb and Nb–X (X = B, N) bonds are released to participate in CO2 reduction and to bond with oxygen atoms. The frontier orbital analysis, as schematically shown in Fig. S12, also clearly exhibits how the Nb–Nb and Nb–B bonds are gradually weakened in the Nb2BN2O0–4 series with an increase in the number of attached oxygen atoms. (2) From the mechanistic point of view, CO2 molecules are captured and activated over two Nb atoms in Nb2BN2 and Nb2BN2O; thus, active sites are two Nb atoms. As the reaction proceeds, CO2 is adsorbed on one Nb atom in Nb2BN2O2, but activated by two Nb atoms. For Nb2BN2O3, only one Nb site can be used to react with CO2. In contrast, due to the longer Nb–Nb distance in Nb2N2 (WBI (Nb–Nb) = 0.68), the two Nb atoms cannot synergistically achieve the reduction of CO2 molecules. Combined with the experimental results that the rate constants for the four consecutive reactions (1a)–(d) gradually diminish and those for Nb2N2/CO2 systems (reactions (2a) and (b)) are much slower, we can draw the conclusion that the M–M multiple bond in clusters are vital for efficient CO2 reduction reactions. If more electrons are stored in the Nb–Nb multiple bond, the backdonation of electron density from the occupied d orbitals of the Nb centers to the vacant antibonding π* orbitals of CO2 becomes stronger. The synergistic effect between the two Nb atoms in Nb2BN2 also has a decided advantage for the rapid activation of CO2 compared to the single Nb reactive site in Nb2N2 or in Nb2BN2O3.

The B → CO2 π backdonation

It is known that an empty orbital in boron is similar to the d orbital of transition metal atoms, but boron offers p-block chemistry. The relationship between the states of boron and CO2 reduction performance is essential for designing new catalysts and revealing the factors influencing CO2 reactivity. A frontier orbital analysis of Nb2BO0,1/CO2 systems has been performed to provide further details and insights. As shown in Fig. 4a, the two p orbitals and two sp hybrid orbitals of B interact with the d-orbitals of two Nb atoms to form bonding π orbitals in Nb2B. When CO2 molecules are adsorbed on Nb2B (I12 in Scheme 1) and Nb2BO (I11), the values of WBI(B–C) are 0.94, indicating that B–C single bonds are formed in I11 and I12. In Fig. 4b, the B (mainly p orbitals and a small portion of sp hybrid orbitals) → CO2 π backdonation is formed, and the overlap of the unoccupied antibonding π*-orbitals of CO2 with the p and sp orbitals of B is comparable to that with the d orbital of Nb. This electron transfer process induces a bending of the linear O[double bond, length as m-dash]C[double bond, length as m-dash]O bond in a free CO2 molecule to 124° in I12 and 122° in I11, and the formation of CO2 units (−0.87e in I12vs. −0.89e in I11). In the literature, most CO2 activation is based on transition metal atoms.20 Notably, oxygen-atom abstraction from CO2 by main-group containing anions such as H3Si, HSiO, and HSiS has been pioneered by the groups of Damrauer, Bowie, and Depuy,50–52 and the Si atoms in these anions are the active sites for oxygen atom transfer. Weber's work also shows the critical importance of moving electron density (negative charge) to CO2 by bending the CO2 and activating it.14 To the best of our knowledge, there is no report about the synergy between a transition metal atom and a main-group atom in CO2 reduction mediated by gas-phase clusters. In the reactions of Nb2B with CO2, since the B atom has a higher electronegativity than the Nb atom (2.04 vs. 1.60) and the B–O bond energy is larger than that of Nb–O by 0.386 eV (Table S1, ESI), path II, which is involved with the B atom as an electron donor, is more prone to CO2 activation by electrophilic attack. Therefore, the product P8 in Scheme 1 is more stable than P7 by 0.39 eV, indicating that path II originating from the B site is more favourable thermodynamically than those involved with two Nb atoms (path I). The Rice–Ramsperger–Kassel–Marcus (RRKM)53 theory estimates that the absolute rate constants of traversing the rate-determining steps in the first CO2 reduction steps for path I and path II are comparable (2 × 1012 s−1vs. 9 × 1011 s−1). Then we can draw the conclusion that path II is more favourable than path I. Based on these data, doping boron atoms into metal-containing catalysts may serve as a useful method to reduce the used amount of metal. However, if path II is compared with the reactions of Nb2BN2 with CO2, in which CO2 molecules are reduced by Nb–Nb sites, the later systems have higher reactivity. Therefore, the chemical environment in clusters is quite important for the activation of stable molecules.
image file: d0dt02680h-f4.tif
Fig. 4 Schematic molecular orbital diagrams for (a) 1IA2, (b) 1I12 and (c) 1I11. The parts of the components of some key molecular orbitals are given. Natural orbital partial occupation numbers are given. See Fig. S13 for a complete collection of orbitals.

Correlation with the condensed phase system

One of the inherent advantages of BN is its superior oxidation resistance.54 Also, in the reported metal atoms supported on BN catalysts, B-vacancy and N-vacancy are introduced into BN materials.7,55 These vacancies can endow BN nanosheets with higher chemical activity56,57 and stronger interaction between metal atoms and BN substrates;58 thus, BN materials with these point defects are excellent substrates to support metal particles for various catalytic reactions. In the CO2 reduction mediated by the Nb2BN2 anions, the WBIs for the B–N bonds remain almost unchanged from Nb2BN2 to Nb2BN2O4 (Table 2), suggesting that the B–N bonds are stable and are not involved in the reactions. This property may correspond to the oxidation resistance of BN materials. In addition, the bond between the Nb and N atoms in Nb2BN2 is not affected by these reactions, which is reflected by WBI(Nb–N); therefore, a strong interaction between the “substrate” BN2 moiety and the Nb atoms exists. This Nb2BN2 cluster may serve as a model for one type of active site on supported Nb/BN with B-vacancies. Therefore, our gas-phase study suggests that the BN substrate with B-vacancies may be more appropriate than separate B and N substrates in the CO2 reduction reactions.


In conclusion, we have studied the reactivities of Nb2BN2, Nb2N2 and Nb2B anions toward carbon dioxide by applying mass spectrometry experiments and theoretical calculations. Consecutive oxygen atom transfer processes by the release of one CO molecule each time were identified in these reactions. Compared to Nb2N2 and Nb2B, the ternary nitride clusters Nb2BN2 have a higher reactivity, and four CO2 molecules are reduced consecutively by Nb2BN2. Under similar reaction conditions, the maximum number of activated CO2 molecules mediated by Nb2B and Nb2N2 is one and two fewer than that by Nb2BN2, respectively. Detailed comparisons indicate that the root causes of the high reactivity of Nb2BN2 are the presence of a Nb–Nb multiple bond, in which electrons are stored and back donated to the π* orbitals of CO2 and appropriate ligands such as B. In addition, the synergy effect between a transition metal atom (Nb) and a main-group atom (B) in CO2 reduction mediated by gas-phase clusters has also been discussed in the Nb2B/CO2 system. The observed high reactivity of Nb2BN2 toward CO2 can provide molecular-level insights into the design of efficient active sites on supported transition metal/BN materials to activate carbon dioxide.

Conflicts of interest

There are no conflicts to declare.


This work was supported by the National Natural Science Foundation of China (No. 91961122), National Key R&D Program of China (No. 2016YFC0203000), and the Fundamental Research Funds for the Central Universities (No. 2017CX01008).

Notes and references

  1. M.-Y. He, Y.-H. Sun and B.-X. Han, Angew. Chem., Int. Ed., 2013, 52, 9620–9633 Search PubMed.
  2. M. Yang, L. F. Allard and M. Flytzani-Stephanopoulos, J. Am. Chem. Soc., 2013, 135, 3768–3771 Search PubMed.
  3. S.-W. Li, Y. Xu, Y.-F. Chen, W.-Z. Li, L.-L. Lin, M.-Z. Li, Y.-C. Deng, X.-P. Wang, B.-H. Ge, C. Yang, S.-Y. Yao, J.-L. Xie, Y.-W. Li, X. Liu and D. Ma, Angew. Chem., Int. Ed., 2017, 56, 10761–10765 Search PubMed.
  4. W. Wang, S.-P. Wang, X.-B. Ma and J.-L. Gong, Chem. Soc. Rev., 2011, 40, 3703–3727 Search PubMed.
  5. X. Su, X.-F. Yang, Y.-Q. Huang, B. Liu and T. Zhang, Acc. Chem. Res., 2019, 52, 656–664 Search PubMed.
  6. D. Golberg, Y. Bando, Y. Huang, T. Terao, M. Mitome, C.-C. Tang and C.-Y. Zhi, ACS Nano, 2010, 4, 2979–2993 Search PubMed.
  7. Y. Lin and J. W. Connell, Nanoscale, 2012, 4, 6908–6939 Search PubMed.
  8. Y. Lin, C. E. Bunker, K. A. S. Fernando and J. W. Connell, ACS Appl. Mater. Interfaces, 2012, 4, 1110–1117 Search PubMed.
  9. C.-J. Huang, C. Chen, X.-X. Ye, W.-Q. Ye, J.-L. Hu, C. Xu and X.-Q. Qiu, J. Mater. Chem. A, 2013, 1, 12192–12197 Search PubMed.
  10. K. Uosaki, G. Elumalai, H. Noguchi, T. Masuda, A. Lyalin, A. Nakayama and T. Taketsugu, J. Am. Chem. Soc., 2014, 136, 6542–6545 Search PubMed.
  11. Q.-R. Fu, Y. Meng, Z.-L. Fang, Q.-Q. Hu, L. Xu, W.-H. Gao, X.-C. Huang, Q. Xue, Y.-P. Sun and F.-S. Lu, ACS Appl. Mater. Interfaces, 2017, 9, 2469–2476 Search PubMed.
  12. S. Back and Y. S. Jung, ACS Energy Lett., 2017, 2, 969–975 Search PubMed.
  13. S. Zhou, X.-W. Yang, X. Xu, S.-X. Dou, Y. Du and J.-J. Zhao, J. Am. Chem. Soc., 2020, 142, 308–317 Search PubMed.
  14. L. G. Dodson, M. C. Thompson and J. M. Weber, Annu. Rev. Phys. Chem., 2018, 69, 231–252 Search PubMed , and references therein.
  15. G. K. Koyanagi and D. K. Bohme, J. Phys. Chem. A, 2006, 110, 1232–1241 Search PubMed.
  16. M.-F. Zhou, Z.-J. Zhou, J. Zhuang, Z.-H. Li, K.-N. Fan, Y.-Y. Zhao and X.-M. Zheng, J. Phys. Chem. A, 2011, 115, 14361–14369 Search PubMed.
  17. Y.-X. Zhao, Q.-Y. Liu, M.-Q. Zhang and S.-G. He, Dalton Trans., 2016, 45, 11471–11495 Search PubMed.
  18. X.-X. Zhang, G.-X. Liu, K.-H. Meiwes-Broer, G. Ganteför and K. Bowen, Angew. Chem., Int. Ed., 2016, 55, 9644–9647 Search PubMed.
  19. H. Schwarz, Coord. Chem. Rev., 2017, 334, 112–123 Search PubMed.
  20. L.-X. Jiang, C.-Y. Zhao, X.-N. Li, H. Chen and S.-G. He, Angew. Chem., Int. Ed., 2017, 56, 4187–4191 Search PubMed.
  21. S.-D. Zhou, J.-L. Li, M. Firouzbakht, M. Schlangen and H. Schwarz, J. Am. Chem. Soc., 2017, 139, 6169–6176 Search PubMed.
  22. M. Wang, C.-X. Sun, J.-T. Cui, Y.-H. Zhang and J.-B. Ma, J. Phys. Chem. A, 2019, 123, 5762–5767 Search PubMed.
  23. J.-L. Li, C.-Y. Geng, T. Weiske and H. Schwarz, Angew. Chem., Int. Ed., 2020, 59, 12308–12314 Search PubMed.
  24. R. Wesendrup and H. Schwarz, Angew. Chem., Int. Ed. Engl., 1995, 34, 2033–2035 Search PubMed.
  25. N. Sändig and W. Koch, Organometallics, 1998, 17, 2344–2351 Search PubMed.
  26. J. Zhuang, Z.-H. Li, K.-N. Fan and M.-F. Zhou, J. Phys. Chem. A, 2012, 116, 3388–3395 Search PubMed.
  27. V. Canale, R. Robinson, A. Zavras, G. N. Khairallah, N. d'Alessandro, B. F. Yates and R. A. J. O'Hair, J. Phys. Chem. Lett., 2016, 7, 1934–1938 Search PubMed.
  28. M. Firouzbakht, M. Schlangen, M. Kaupp and H. Schwarz, J. Catal., 2016, 343, 68–74 Search PubMed.
  29. M. Firouzbakht, N. J. Rijs, M. Schlangen, M. Kaupp and H. Schwarz, Top. Catal., 2018, 61, 575–584 Search PubMed.
  30. Q. Chen, Y.-X. Zhao, L.-X. Jiang, J.-J. Chen and S.-G. He, Angew. Chem., Int. Ed., 2018, 57, 14134–14138 Search PubMed.
  31. E. Hossain, D. W. Rothgeb and C. C. Jarrold, J. Chem. Phys., 2010, 133, 024305 Search PubMed.
  32. D. W. Rothgeb, E. Hossain, J. E. Mann and C. C. Jarrold, J. Chem. Phys., 2010, 132, 064302 Search PubMed.
  33. M. Ervin, J. Chem. Phys., 1995, 103, 7897–7906 Search PubMed.
  34. Z. Yuan, Z.-Y. Li, Z.-X. Zhou, Q.-Y. Liu, Y.-X. Zhao and S.-G. He, J. Phys. Chem. C, 2014, 118, 14967–14976 Search PubMed.
  35. Y.-X. Zhao, Z.-Y. Li, Z. Yuan, X.-N. Li and S.-G. He, Angew. Chem., Int. Ed., 2014, 53, 9482–9486 Search PubMed.
  36. Z.-Y. Li, Z. Yuan, X.-N. Li, Y.-X. Zhao and S.-G. He, J. Am. Chem. Soc., 2014, 136, 14307–14313 Search PubMed.
  37. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2013 Search PubMed.
  38. C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789 Search PubMed.
  39. A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100 Search PubMed.
  40. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652 Search PubMed.
  41. D. Andrae, U. Haussermann, M. Dolg, H. Stoll and H. Preuss, Theor. Chem. Acc., 1990, 77, 123–141 Search PubMed.
  42. T. H. Dunning, J. Chem. Phys., 1989, 90, 1007–1023 Search PubMed.
  43. X.-L. Ding, Z.-Y. Li, J.-H. Meng, Y.-X. Zhao and S.-G. He, J. Chem. Phys., 2012, 137, 214311 Search PubMed.
  44. C. Gonzalez and H. B. Schlegel, J. Chem. Phys., 1989, 90, 2154–2161 Search PubMed.
  45. C. Gonzalez and H. B. Schlegel, J. Phys. Chem., 1990, 94, 5523–5527 Search PubMed.
  46. E. D. Glendening, J. K. Badenhoop, A. E. Reed, J. E. Carpenter, J. A. Bohmann, C. M. Morales, C. R. Landis and F. Weinhold, NBO 6.0, Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, 2013, http://nbo6.chem.wisc.edu/ Search PubMed.
  47. T. Lu and F.-W. Chen, J. Comput. Chem., 2012, 33, 580–592 Search PubMed.
  48. K. B. Wiberg, J. Am. Chem. Soc., 1968, 90, 59–63 Search PubMed.
  49. K. B. Wiberg, Tetrahedron, 1968, 24, 1083–1096 Search PubMed.
  50. J. C. Sheldon, J. H. Bowie, C. H. Depuy and R. Damrauer, J. Am. Chem. Soc., 1986, 108, 6794–6800 Search PubMed.
  51. S. Gronert, R. A. J. O'Hair, S. Prodnuk, D. Suelzle, R. Damrauer and C. H. Depuy, J. Am. Chem. Soc., 1990, 112, 997–1003 Search PubMed.
  52. R. Damrauer, M. Krempp and R. A. J. O'Hair, J. Am. Chem. Soc., 1993, 115, 1998–2005 Search PubMed.
  53. J. I. teinfeld, J. S. Francisco and W. L. Hase, Chemical Kinetics and Dynamics, Prentice-Hall, New Jersey, 1999, p. 231 Search PubMed.
  54. L. Ma and X.-C. Zeng, Nano Lett., 2017, 17, 3208–3214 Search PubMed.
  55. C.-H. Jin, F. Lin, K. Suenaga and S. Iijima, Phys. Rev. Lett., 2009, 102, 195505 Search PubMed.
  56. W.-L. Sun, Y. Meng, Q.-R. Fu, F. Wang, G.-J. Wang, W.-H. Gao, X.-C. Huang and F.-S. Lu, ACS Appl. Mater. Interfaces, 2016, 8, 9881–9888 Search PubMed.
  57. J.-H. Wu, L.-C. Wang, B.-L. Lv and J.-G. Chen, ACS Appl. Mater. Interfaces, 2017, 9, 14319–14327 Search PubMed.
  58. S. Lin, X.-X. Ye, R. S. Johnson and H. Guo, J. Phys. Chem. C, 2013, 117, 17319–17326 Search PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/d0dt02680h

This journal is © The Royal Society of Chemistry 2020