Reaction Mechanism of Cu(I)-Mediated Reductive CO2 Coupling for

Article pubs.acs.org/IC

Reaction Mechanism of Cu(I)-Mediated Reductive CO2 Coupling for the Selective Formation of Oxalate: Cooperative CO2 Reduction To Give Mixed-Valence Cu2(CO2•−) and Nucleophilic-Like Attack Jialing Lan,† Tao Liao,† Tonghuan Zhang,†,‡ and Lung Wa Chung*,† †

Department of Chemistry, South University of Science and Technology of China, Shenzhen 518055, China Lab of Computational Chemistry and Drug Design, Key Laboratory of Chemical Genomics, Peking University Shenzhen Graduate School, Shenzhen 518055, China

‡

S Supporting Information *

ABSTRACT: A dinuclear, Cu(I)-catalyzed reductive CO2 coupling reaction was recently developed to selectively yield a metal− oxalate product through electrochemical means, instead of the usual formation of carbonate and CO (Science 2010, 327, 313). To shed light on the mechanism of this important and unusual reductive coupling reaction, extensive and systematic density functional theory (DFT) calculations on several possible pathways and spin states were performed in which a realistic system up to 164 atoms was adopted. Our calculations support the observation that oxalate formation is energetically more favorable than the formation of carbonate and CO products in this cationic Cu(I) complex. Spatial confinement of the realistic catalyst (a long metal−metal distance) was found to further destabilize the carbonate formation, whereas it slightly promotes oxalate formation. Our study does not support the proposed diradical coupling mechanism. Instead, our calculations suggest a new mechanism in which one CO2 molecule is first reduced cooperatively by two Cu(I) metals to give a new, fully delocalized mixed-valence Cu2I/II(CO2•−) radical anion intermediate (analogues to Type 4 Cu center, CuA), followed by further partial reduction of the metal-ligated CO2 molecule and (metal-mediated) nucleophilic-like attack on the carbon atom of an incoming second CO2 molecule to afford the dinuclear Cu(II)−oxalate product. Overall, our proposed reaction mechanism involves a closed-shell reactant as well as two open-shell transition states and products. The effects of size, charge, and catalyst metal on the oxalate formation were also investigated and compared.

1. INTRODUCTION Carbon dioxide (CO2) is the main greenhouse gas in the atmosphere.1 The growing atmospheric concentration of CO2 has resulted in global warming and has affected global climate change. CO2 can be a good chemical feedstock and can be used to synthesize valuable organic compounds due to the ample supply, nontoxicity, and low cost of CO2.2,3 However, CO2 is underutilized in the chemical industry because of its high thermodynamic stability and reaction barriers. Recently, the activation and functionalization of inert CO2 has become a challenging subject and research hotspot.2−4 Some metal complexes, including low-valent d-block,5−7 fblock,8,9 and s-block10 metal complexes, have been reported to reduce two CO2 molecules to form carbonate and CO (Scheme 1a). For instance, a peroxo-dicopper(II) adduct was used to react with CO2 to yield a carbonate-bridged complex.6b In contrast, examples of the selective formation of a value-added oxalate complex (Scheme 1b) from the same reaction are scarce. Notably, selective oxalate formation is more meaningful than carbonate formation because oxalate is a useful chemical feedstock for the synthesis of some useful organic compounds (e.g., methyl glycolate).11a Jones, Maron, and co-workers10 © 2017 American Chemical Society

Scheme 1. Two Possible Reductive Couplings of CO2

reported reductive coupling reactions mediated by a Mg(I) complex to give a Mg(II)−carbonate complex (as the major product) and a Mg(II)−oxalate complex (as the minor product). Their density functional theory (DFT) (B3PW91D method) calculations showed that the carbonate pathway has a lower barrier than the oxalate pathway by 10.4 kcal/mol.10b A few metal (e.g., Cu, Fe, Ir, Ni, and Yb) complexes have recently been found to selectively form metal−oxalate complexes.11−15 Bouwman and co-workers reported that the cationic dicopper(I) complex 1L promotes the reductive CO2 Received: December 22, 2016 Published: May 30, 2017 6809

DOI: 10.1021/acs.inorgchem.6b03080 Inorg. Chem. 2017, 56, 6809−6819

Article

Inorganic Chemistry coupling reaction to selectively give a Cu(II)−oxalate complex 6 and finally a Li−oxalate complex in acetonitrile solution using an electrochemical approach and a Li salt (with the formation of 12 oxalate molecules per Cu(I) complex, see Scheme 2).11a

Scheme 3. Three Model Copper Complexes Used in this Study

Scheme 2. Proposed Diradical-Coupling Mechanism for Oxalate Formation

the small-model results. To uncover the effect of the charge of the Cu(I) complex, we also used a small neutral Cu(I) complex A for the key steps (Scheme 3). Furthermore, the effect of the metal (Ni, Ag, or Au) on the energetics of oxalate formation was also briefly examined by replacing Cu in I by one of these metals. 2.2. Methods. All DFT calculations were performed with the Gaussian 09 program.19 The M06-L method and 6-31G* basis set were used to fully optimize the geometries of the reactants, intermediates, products, and transition states (TS). Unrestricted (U) DFT methods were used for calculations involving open-shell singlet, doublet, triplet, and quintet states, and restricted DFT methods were employed for closed-shell singlet calculations. UDFT methods have been widely and successfully used to study many bioinorganic and biomimetic systems with open-shell configurations.20 The Stuttgart/ Dresden ECPs and basis sets were applied for the Ag and Au metals.21 The M06-L method was reported to accurately describe the transition metal−ligand, thermochemistry, and dispersion interactions.22 Harmonic vibration frequency calculations on the optimized structures were conducted to characterize the minima (without imaginary frequency) and transition states (with one imaginary frequency) and to derive the zero-point energy (ZPE) and free-energy correction. Moreover, the intrinsic reaction coordinate (IRC) calculations of the critical transition states were performed.23 In addition, the solvent effect (acetonitrile) was included by performing single-point energy calculations with an implicit solvent SMD model24 on the gas-phase optimized geometries and using the M06-L/6-31G* and B3LYP-D3/6-31G* methods (ΔEsoln).25 We also tested the optimization of some important structures for the small model in solution. The calculated results for the optimization in solution were similar to those obtained from the single-point energy SMD calculations on the gas-phase structures (Tables S2−S5). Furthermore, as the previously reported Mg(I) complex,10b the B3PW91-D method26 was used for the single-point energy calculations in solution. Our mechanistic conclusions were not changed by the B3LYP-D3 and B3PW91-D3 methods (Tables S40−S44, S53, and S54 and Figures 1−4 and S25−S27). Several functionals (M06-L, M06, B3PW91-D3, B3LYP-D3, PBE0-D3, BHandHLYP, and ωB97XD)22,27,29 were also used to examine the critical energies and electronic structure of the key structure in solution (Tables 1 and S62). The spin contamination correction proposed by Yamaguchi was applied to the critical open-shell singlet transition state by the B3LYP-D3 and M06-L methods (the correction is minor: 0.3 and 1.1 kcal/mol, respectively).28 2.3. Entropy. The entropic contributions in the solution have long been known to be overestimated when entropy in the gas phase was used.29 Accordingly, the computed relative free energies are significantly influenced by different numbers of reactant and product molecules.29,30 The overestimated entropic contribution could be serious in this reaction: a

This reaction is highly selective possibly due to the formation of the thermodynamically favorable oxalate product.9f,11a Subsequently, two cationic Cu(I) complexes were also reported for the selective reductive coupling reaction.11c,d Unfortunately, the mechanism for this essential and selective Cu(I)-mediated reductive CO2 coupling to form the oxalate product has remained unclear. Bouwman and co-workers proposed a diradical-coupling mechanism for oxalate formation, in which the two Cu(I) metals first reduce and activate two CO2 molecules to give an intermediate with two Cu(II)(CO2•−) radical anions followed by a diradical coupling to form the C− C bond of oxalate (Scheme 2).11a To shed light on the reaction mechanism and to continue our studies on sustainable chemistry,16 extensive and systematic DFT calculations on the mechanism of this unusual and important reductive CO2 coupling reaction were performed in this study. Several possible pathways in several possible spin states17 were examined in this mechanistic study. Our computational results support the selective formation of the Cu(II)−oxalate product. In addition, our calculations propose a new mechanism: the single reduction of only one CO2 molecule by the two Cu(I) metals first proceeds to generate a new, fully delocalized mixed-valence18 Cu2(CO2•−) radical anion intermediate. Then, upon further reduction of the metalligated CO2 molecule, a metal-mediated nucleophilic-like attack on an incoming second CO2 molecule occurs, forming dicopper(II)−oxalate.

2. COMPUTATIONAL DETAILS 2.1. Model Complexes. We first used a small mononuclear cationic Cu(I) complex I (see Scheme 3) to effectively explore several possible pathways with several possible electronic configurations.17 The realistic dinuclear Cu(I) complex 1L was then adopted to examine the critical intermediates and transition states of the most probable pathways obtained from 6810


Article

Inorganic Chemistry

−2.5, −4.9, −7.2, −9.5, and −11.7 kcal/mol were applied for two-to-one, three-to-one, four-to-one, five-to-one, and six-toone transformations, respectively. The dispersion-corrected and above-mentioned entropy-corrected relative free energies in solution by the widely used B3LYP-D3 method (ΔGsoln,cor, see eq 1) were mainly used to discuss the reaction mechanism, unless stated otherwise. Electronic energy profiles for the two key paths (excluding the entropy and dispersion) are also given in Figures S31 and S32.

Table 1. Effect of Different Functionals on the Reaction Barriers of the Oxalate Pathway (Relative to Its Preceding Lowest Free-Energy Intermediate), as Well as Relative Stability of the Product 56a in Solution oxalate pathway 3

TS3a−4a a

M06-L B3LYP-D3 B3PW91-D3 PBE0-D3 M06 BHandHLYP wB97XD a

40.8 22.0c 19.6e 20.7c 40.3a 29.9c 20.3c b

oss

5

TS3a−4a a

TS5a−6a

f

39.7 /40.8 31.0c/31.3f 29.8e 30.7c 46.5a 36.2c 31.0c c

a

40.8 23.9d 31.4d 23.2d 38.5d −4.0d 24.6d

5

6a

−9.4b −55.1b −48.5b −63.7b −15.1b −92.4b −62.9b

ΔGsoln,cor = ΔEsoln + ΔGTCFE,gas + ΔEdisp + ΔGcor ‐ FVT (1)

where ΔGTCFE,gas is the “thermal correction to Gibbs free energy” and ΔEdisp is the dispersion energy.

d

Relative to 11L2s. Relative to 11L. Relative to 37. Relative to 34a. Relative to 32. fThe Yamaguchi correction was applied.

e

3. RESULTS AND DISCUSSION We first discuss the energies and structures of the cationic Cu(I) complexes and key intermediates before the reductive coupling reaction mechanism. The reaction mechanisms for the oxalate pathway (section 3.1) are then presented. Then, the mechanism of the more likely side reaction, the carbocyclic pathway (section 3.2), is reported. Finally, the effects of the charge and the catalyst metal are briefly compared in section 3.4. For the calculations with both the small and realistic catalysts, the key results for the realistic model are mainly discussed herein, and the detailed results and discussion for the small model are given in the Supporting Information.

reaction of two dinuclear Cu(I) complexes with two CO2 molecules to give only one tetracopper(II)−oxalate intermediate in the first coupling step (four-to-one transformation), as well as a reaction of two dinuclear Cu(I) complexes with four CO2 molecules to give only one tetranuclear copper(II)− dioxalate product (six-to-one transformation). We included an entropy correction to the calculated free energies based on the free-volume theory,30 which was also used in some previous computational works.31 For one-to-one transformations, no correction is needed. Free energy corrections (ΔGcor‑FVT) of

Figure 1. Calculated key structural parameters for 11L, 11L2s 11LCO2a, 1,31LCO2, and 51L2CO2a. Bond lengths, bond angles (italics), Mulliken charges, and spin densities calculated by the B3LYP-D3//M06-L method are given. Only the key atoms are shown for clarity. 6811


Article

Inorganic Chemistry

Figure 2. Dispersion- and entropy-corrected free energy profile in solution for the oxalate pathway for the realistic catalyst.17 “Solv” represents an NCMe solvent molecule. The relative free energies calculated by the B3LYP-D3//M06-L method are given.

cationic [Cu(II)(CO2•−)]2 radical anion intermediate11a is highly unstable. Surprisingly, compared to the case of the small catalyst with coordination of CO2 to one Cu(I) metal,32 the singlet−triplet energy gap in 1,31LCO2 (with coordination of CO2 to two Cu(I) metals) is very small (2.1 kcal/mol). Therefore, these computational results suggest that the two Cu metals cooperatively reduce and activate CO2 to give new (Class-III) fully delocalized mixed-valence18,33,34 Cu2(CO2•−) radical anion intermediates 31LCO2 and 33a (with spin density of 0.52−0.68 on each Cu in 31LCO2 and 33a), which have not been previously proposed and are essential for the subsequent coupling step (see below). Comparatively, a very high CO2 binding affinity with a neutral (NHC)2Ni02 complex (by the mPW1K method, ΔG = −20.0 kcal/mol; Ni−O/Ni−C = 1.90/1.80 Å) was reported by the Lin group35 because the electron-rich neutral Ni0 center promotes CO2 binding through back-donation from a filled d(Ni) orbital to a π*(CO2) orbital. 3.1. Oxalate Pathway. For the realistic catalyst 1L, the most feasible pathway involves 11L, 11L2s, 11LCO2a, 2, and 3a (or 3b) (Figure 2). The computed barrier for the first reductive coupling step via 3TS3a−4a is approximately 16.5 kcal/mol above 3 2 (22.0 kcal/mol above 37, Figures 2 and 4).36 The B3PW91D3 method gives a comparable barrier (19.6 kcal/mol above 3 37 2, Table 1). The effects of different functionals on the reaction barriers and the stability of the final products were also examined (Table 1). Generally, the reaction barriers are decreased by including Hartree−Fock (HF) exchange

For the realistic catalyst 1L, the coordination of two acetonitrile solvent molecules to two Cu(I) metals to form 1 1L2s was computed to be exergonic by 10.4 kcal/mol (Figures 1 and 2). In comparison, the coordination of the less basic CO2 to the cationic Cu(I) is weaker and becomes an endergonic step (Figures 1 and 2). One CO2 molecule can coordinate to two metals or one metal of one dicopper(I) complex 11L to form the intermediate 1,3 1LCO2 or 11LCO2a (Figure 2), respectively. 11LCO2a is lower in free energy than 1,31LCO2 by 10.8−12.9 kcal/mol. Furthermore, 1 1LCO2a (C−Cu/O−Cu: 2.1/2.5 Å) and 11LCO2 (C−Cu/O− Cu: 1.9/2.1 Å) are less stable than 11L2s by 14.6 and 25.4 kcal/ mol, respectively. Interestingly, the formal single-electron reduction of one CO2 by the two Cu(I) to form a fully delocalized mixed-valence Cu2I/II(CO2•−) radical anion intermediate 31LCO2 was found (spin density of Cu1/Cu2/CO2: 0.68/0.52/0.53, Figure 1). The acute bond angle (124°) and long C−O distances (1.26−1.27 Å) in 31LCO2 are similar to the ones for free CO2•− (133° and 1.25 Å). These results further support the radical anionic nature of 31LCO2. The formation of a formal [Cu(II)(CO2•−)]2 radical anion intermediate 51L2CO2a in a quintet state (spin density of Cu/CO2: 0.83−0.84/0.85− 0.86) was computed to be much higher in free energy than 1 1L2s, by 67.1 kcal/mol (Figures 1 and 2), partly because charge transfer from the cationic Cu(I) metals to the inert CO2 ligands (MLCT) to form the radical anion intermediate is difficult. These results suggest that the formation of the proposed 6812


Article

Inorganic Chemistry

Figure 3. (a), (c), and (d) Calculated key structural parameters for some key structures in the oxalate pathways with the realistic model catalyst. Bond lengths, bond angles (italics), Mulliken charges, and spin densities calculated by the B3LYP-D3//M06-L method are given. Only the key atoms are shown for clarity. (b) Highest occupied molecular orbital for 3TS3a−4a..

reactions to generate the final products 36a (ΔGsoln,cor = −53.7 kcal/mol) and 56a (ΔGsoln,cor = −55.1 kcal/mol) are energetically more favorable than 32 by 40.8−42.2 kcal/mol. Moreover, the reaction for the tetracopper(II)−dioxalate complex 56a with four LiClO4 molecules to form two Li−oxalate complexes is endergonic by 12.1−19.8 kcal/mol according to the M06-L, B3LYP-D3, and B3PW91-D3 methods (Scheme 4). Meanwhile, the reaction with the corresponding tetracopper(I)− dioxalate complex is energetically favorable (−32.6 to −43.3 kcal/mol). Moreover, we further investigated the mechanistic possibility of another coupling pathway through the two Cu(I) metals involving only one complex 1L. Our calculations revealed that this pathway requires a very high barrier of 61.2 kcal/mol (via 3 TS2CO2 above 11L2s), and the corresponding coupling intermediate 31L2CO2‑p was computed to be less stable than 3 4a by 32.3 kcal/mol (Figure 2), presumably due to the ring strain and larger electrostatic repulsion between the cationic metal centers (CuCu: 5.11 Å in 31L2CO2‑p vs 5.36 Å in 34a) in the former case. These results suggest that the participation of the two 1L complexes can promote the coupling reaction. For the small model (Scheme 5), an alternative oxalate pathway catalyzed by only one Cu(I) metal via 3TSXV−XVI was found to require a much higher barrier (43.9 kcal/mol above 1IIIs). These calculations demonstrate that the participation of two Cu(I) metal centers is critical to cooperatively activate and couple two CO2 molecules. In principle, there are three possible mechanistic pathways to form the observed Cu(II)−oxalate product (6a, Scheme S2)

correlation, and the stability of the final products increases with increasing the amount of the HF exchange correlation. Importantly, for the realistic catalyst, the barrier for the oxalate formation is smaller than for the carbocyclic pathway (Tables 1 and S61 and see discussion below). Likewise, the O−C−O bond angle and O−C bond lengths of the first CO2 moiety in 3 TS3a−4a (119°, 1.28−1.29 Å) are comparable to the ones in free CO2•− (133° and 1.25 Å, Figure 3). Moreover, the spin densities for 3TS3a−4a are mainly located on the first CO2 moiety (CO2: 0.32) and the two reacting Cu atoms (Cu: 0.63− 0.65, Table S62 and Figure 3). These structural features and the spin density distribution in 3TS3a−4a demonstrate the Class-III fully delocalized mixed-valence radical anion character of 3 TS3a−4a.38 The Class-III type (fully delocalized) mixed-valence radical anion character in the key precursor intermediate and coupling transition state are qualitatively supported by different functionals (Table S62). Furthermore, the computed C−C bond-forming distance in 3TS3a−4a is rather long (2.60 Å), indicating an early transition state. In addition, the formation of the resultant intermediate 34a (ΔGsoln,cor = −33.1 kcal/mol) is more stable than 32 by 20.2 kcal/mol. Furthermore, the barriers of the second reductive coupling step via 3TS5a−6a and 5TS5a−6a to form the stable final tetracopper(II)−dioxalate products 36a and 56a are slightly increased to 23.9−24.1 kcal/mol (relative to 34a), compared to the first reductive coupling step (22.0 kcal/mol). The key structural features (C−C: 2.71 Å) and spin density distributions (Cu2/Cu3, 0.62/0.65; CO2, 0.34) in 3TS5a−6a and 5TS5a−6a are also similar to the ones in 3TS3a−4a, showing the Class-III (fully delocalized) mixed-valence radical anion character. The overall 6813


Article

Inorganic Chemistry

carbocyclic pathway for the realistic model are summarized in this section (Figure 4) because this pathway is energetically more favorable than the oxo pathway (see the discussion section in the Supporting Information). Starting from the intermediate 37, the CO bond of the second CO2 could insert into the Cu4O4 (path C) or Cu2 O1 bond (path D) to give a five-membered-ring intermediate, 3 9b1 or 39b2, respectively. Overall, path C requires a slightly lower barrier than path D (with an opposite insertion direction of the incoming CO2 molecule) by 3.6 kcal/mol. The barriers for the first insertion step and the second CO cleavage step in path C via 3TS8b2−9b2 and 3TS9b2−10b2 are 21.6 and 29.0 kcal/ mol above 37, respectively. The higher barrier for the second step is attributed to the unstable carbonate product 310b2 (ΔGsoln,cor = 8.5 kcal/mol). Remarkably, the overall barrier in path C is higher than for the corresponding oxalate pathway (via 3TS3b−4b) by 5.2 kcal/mol. Unfortunately, attempts to locate an intermediate or TS in the absence of the third CO2 molecule (path E) were not successful, except for the Cu(II)−carbonate carbon monoxide product 310a (Figure 4). However, 310a is higher in free energy than 37 by 21.7 kcal/mol. Additionally, the assumed intermediate 39a was proven to be unstable because the second CO2 molecule dissociates during the geometry optimization. When the C3−O4 bond (1.57 Å as in 39b2) is fixed, 39a−fix is higher in electronic energy than 39b2 by 12.8 kcal/mol in solution (Table S54). Therefore, the overall reaction barrier (above 37) leading to 310a must be higher than 21.7 kcal/mol (Figures 2 and 4). Therefore, the current results for the realistic catalyst highlight that the oxalate pathway is both thermodynamically and kinetically more favorable than the carbocyclic pathway as calculated by the M06-L, B3LYP-D3, and B3PW91D340 methods. These results explain the preferential formation of the metal−oxalate product in the experiment.11a 3.3. Comparison of the Oxalate and Carbocyclic Pathways. We further compared the geometries of the key transition states and intermediates in the carbocyclic and oxalate pathways. First, the reacting Cu1−Cu4 distances for the realistic catalyst are long for both the former (4.8−5.2 Å) and latter (5.3−5.4 Å) pathways. Our computed Cu−Cu distance in the Cu(II)−oxalate complex 6a (5.34−5.37 Å) is also similar to the observed distance (5.43 Å, Table 2).11a Notably, the Cu− Cu distances for the oxalate pathway with the realistic catalyst are similar to the distances with the small-model catalyst (3V, 3 TSV−VI, and 3VI: 5.2−5.4 Å, Table S34), whereas these distances for the crucial steps (3TS8b2−9b2, 39b2, 310b2, and 3 10a: 5.1−5.2 Å) of the carbocyclic pathway with the realistic catalyst are longer than for the small-model catalyst (4.8−4.9

Scheme 4. Dispersion- and Entropy-Corrected Free Energetic Profiles of Two Possible Reactions between the Copper Oxalate Complexes and Lithium Perchlorate for the Realistic Catalysts in Solution

involving 51L2CO2a (the proposed radical−radical coupling shown in Scheme 2), TS3a−4a, and TS5a−6a (the sequential coupling with two molecules of the catalyst) as well as TS2CO2 (combination of the twice sequential coupling with one molecule of the catalyst with rearrangement). Only our proposed pathway via TS3a−4a and TS5a−6a has the lowest barrier than the other pathways by at least 39.2 kcal/mol. As to the possibility of a mononuclear catalyst in this system, our computed barrier for the small mononuclear Cu model is 28.5 kcal/mol via 3TSV−VI (Scheme 5), which is higher than the real dinuclear Cu catalyst by 6.5 kcal/mol and should have a lesser possibility of this mononuclear catalyst.39 These computational results also indicate that the size effect of the catalyst in this pathway is not significant, although the pathway is slightly energetically more favorable with the realistic catalyst. Moreover, our proposed reaction mechanism involves two (or multiple) different electronic structures with a closed-shell Cu(I) catalyst and two open-shell C−C coupling transition states and products (open-shell singlet/triplet for the first coupling step and open-shell triplet/quintet for the second coupling step). 3.2. Carbocyclic Pathway. The reductive coupling reaction of CO2 molecules with metal complexes is often found to preferentially give a carbonate and CO product.5−10 The carbonate can be formed through either the carbocyclic pathway or the oxo pathway.8c,9c Our key results regarding the

Scheme 5. Key Results for the Alternative Mechanism of Reductive Coupling for the Small Model

6814


Article

Inorganic Chemistry

Figure 4. Dispersion- and entropy-corrected relative free energy profile in solution by the B3LYP-D3//M06-L method for the carbocyclic pathway for the realistic catalyst.17

Table 2. Metal−Metal Distances (Å) in Some Recent Metal− Oxalate and Metal−Carbonate Complexes

Table 3. Energy Decomposition Analysis (Distortion and Interaction Energies) for the Key Oxalate and Carbocyclic Steps in Solution by the B3LYP-D3 Method (in kcal/mol)

Metal−Oxalate Complexes Cu−Cu11a 5.43 (our DFT results) (5.34−5.37) Cu−Cu11c 5.29 Cu−Cu11d 5.42−5.46 Fe−Fe13a 5.34 Ni−Ni14 5.23 Metal−Carbonate Complex Fe−Fe13b 5.01

ΔEsol

entry

Å). A longer metal−metal distance was also recently reported in crystal structures of certain metal−oxalate complexes (5.23− 5.46 Å, including two recent Cu(I) complexes for the selective formation of oxalate11c,d), compared to a recent related Fe− carbonate complex (5.01 Å, Table 2). In addition, when the size of the catalyst was increased, the Cu−S and Cu−CCO bonds were found to become significantly weaker in the carbocyclic product 310b2 (Cu−S bond distances, 2.89−3.04 Å in 3Xa and 3.39−3.64 Å in 310b2; Cu−CCO bond distances, 1.98 Å in 3Xa and 2.13 Å in 310b2, Tables S38 and S39), whereas the metal−ligand bonds in both the small-model and realistic oxalate products 3VI and 34b are similar. Furthermore, a stronger steric repulsion between the CO ligand and the carbonate part was found in 310b2 (OCO3− CCO distances: 2.78 Å in 3Xa and 2.52 Å in 310b2). Accordingly, these structural features (a larger structural difference (Cu−Cu bond), a weaker Cu−ligand interaction, and a larger steric repulsion) are involved in the carbocyclic pathway for this realistic cationic Cu(I) complex. In fact, when the size of the catalyst was increased, the carbocyclic pathway requires less distortion energy (ΔΔEsol‑dist = −0.3 kcal/mol) and has a weaker interaction energy (ΔΔEsol‑int = +16.6 kcal/mol) for the catalyst and substrates than the oxalate pathway (Table 3).41 This unfavorable interaction energy for the carbocyclic pathway should be due

A B C

3

E F G

3

TSV−VI TS3b−4b size effect

3

TSIXa−Xa TS9b2−10b2 size effect ΔΔE 3

Esol‑dist

Oxalate Pathway −7.5 73.8 −33.2 139.9 (66.1)a (−25.7)a Carbocyclic Pathway −22.3 186.5 −31.7 252.3 (65.8)b (−9.4)b c [+16.3] [−0.3]c

ΔEsol‑int −81.3 −173.1 (−91.8)a −208.8 −284.0 (−75.2)b [+16.6]c

a

Size effect evaluated from energy difference between entries B and A. Size effect evaluated from energy difference between entries F and E. c The difference in the size effect (ΔΔE) was evaluated from energy difference between entries G and C. b

to the above-mentioned structural features (a larger structural change, weaker Cu−ligand interactions, and a larger steric repulsion). Therefore, the spatial confinement of the realistic catalyst (with longer Cu−Cu distance) favors oxalate formation but disfavors carbonate formation. The unstable dicopper−carbonate monoxide complexes in our study are different from the electron-rich Ni system studied by the Lin group.35 The formation of the CO-coordinated Ni− carbonate product (with a computed exergonicity of 6.2 kcal/ mol) can be stabilized by strong back-donation from the filled d(Ni) orbital to the empty CO π* ligand.35 Maron and coworkers revealed that the oxalate pathway for the neutral Mg(I) complex must overcome a higher barrier than the carbonate pathway by 10.4 kcal/mol (B3PW91-D method).10b This difference in the energy barrier qualitatively explains why this cationic Cu(I) system does not favor the carbocyclic pathway, as well as why the neutral Ni(0) complex and neutral Mg(I) complex favor the formation of carbonate and CO. 6815


Article

Inorganic Chemistry 3.4. Effects of the Charge and Metal. The effects of both the charge and the catalyst metal were briefly examined by the M06-L method (Scheme 3, Figures S17−S21, and Tables S18− S24). When the small neutral Cu(I) complex A was used, the barrier of the oxalate pathway is slightly smaller than for the cationic catalyst I by 1.8 kcal/mol (Section 3.1). For the carbocyclic pathways, the barriers are also decreased to 26.8− 27.9 kcal/mol compared to the ones for I (30.1−36.0 kcal/ mol). Moreover, when the Cu metal was replaced in I with Ni, the electron-rich neutral Ni(0) catalyst INi is computed to have a small singlet/triplet energy gap (i.e., a strong reducing ability) and strongly favors IIINi and VINi, compared to INi.42 In contrast, when Cu was replaced by Ag or Au, the corresponding cationic metal complexes have larger singlet/triplet energy gaps (lower reducing ability), and their key intermediates and products are less stable than the ones for the Cu complex. These results demonstrate that the more electron-rich the metal complex is, the lower the barriers of the carbocyclic and oxalate pathways are, and the more favorable the carbocyclic pathway is.

indicate that oxalate formation is energetically more favorable than carbonate formation in this cationic Cu(I) complex. Spatial confinement in the realistic catalyst (with a long Cu−Cu distance) destabilizes the carbonate formation but slightly favors oxalate formation. Our calculations do not support the proposed diradical coupling mechanism (Scheme 2).11a Instead, we propose a new mechanism in which one CO2 molecule is first cooperatively reduced by two Cu(I) metals to give a new delocalized, mixed-valence Cu2(CO2•−) radical anion intermediate (analogues to Type 4 Cu center, CuA), followed by further partial reduction and a metal-mediated nucleophilic-like attack of the carbon of a second CO2 molecule to generate the dicopper(II)−oxalate product (Scheme 6 (bottom)). Additionally, different electronic structures were found to be involved in this reductive reaction (closed-shell Cu(I) reactant and openshell C−C coupling transition states, as well as oxalate products). Moreover, the positive charges of the catalyst were found to favor oxalate formation over carbonate formation. Our study offers new insights and a detailed understanding of the mechanism of this important reductive CO2 coupling reaction to form oxalate. Our mechanism may also apply to the same reaction mediated by other electron-poor metal complexes. However, we do not eliminate the possibility of the proposed diradical mechanism for electron-rich metal complexes.

4. CONCLUSION The reaction mechanism of the Cu(I)-mediated reductive CO2 coupling to the preferential and uncommon formation of the Cu(II)−oxalate product was extensively and systematically investigated by DFT calculations. Our computational results

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b03080. Discussion for the small-model catalyst, complete citations for refs 3a and 19, structural parameters for key structures, Cartesian coordinates and energies of all optimized structures, IRC calculations, key transition states (PDF)

Scheme 6. (Top) Mechanisms of the Two Side Reactions and (Bottom) Our Proposed Schematic Mechanism of the Electrocatalytic CO2 Reductive Coupling for the Selective Oxalate Formation Mediated by Two Cationic Cu(I) Complexes

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. (L. W. Chung) ORCID

Lung Wa Chung: 0000-0001-9460-7812 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge the financial support from the National Natural Science Foundation of China (21473086 and 21672096), the Shenzhen Peacock Program (KQTD20150717103157174), and the South University of Science and Technology of China. We sincerely thank Prof. Elisabeth Bouwman for helpful discussions and the reviewers for comments.

■

REFERENCES

(1) (a) Friedlingstein, P.; Andrew, R. M.; Rogelj, J.; Peters, G. P.; Canadell, J. G.; Knutti, R.; Luderer, G.; Raupach, M. R.; Schaeffer, M.; van Vuuren, D. P.; Le Quéré, C. Persistent growth of CO2 emissions and implications for reaching climate targets. Nat. Geosci. 2014, 7, 709−715. (b) Nordhaus, W. D. To slow or not to slow: the economics of the greenhouse effect. The Economic Journal 1991, 101, 920−937. (2) Aresta, M., Ed. Carbon Dioxide as a Chemical Feedstock; WileyVCH: Weinheim, Germany, 2010. 6816


Article

Inorganic Chemistry

from CO2 using reduced rhenium tricarbonyl catalysts. J. Am. Chem. Soc. 2012, 134, 5180−5186. (5) (a) Notni, J.; Schenk, S.; Görls, H.; Breitzke, H.; Anders, E. Formation of a unique zinc carbamate by CO2 fixation: implications for the reactivity of tetra-azamacrocycle ligated Zn(II) complexes. Inorg. Chem. 2008, 47, 1382−1390. (b) Sarkar, B.; Liaw, B. J.; Fang, C. S.; Liu, C. W. Phosphonate- and ester-substituted 2-cyanoethylene-1,1dithiolate clusters of zinc: aerial CO2 fixation and unusual binding patterns. Inorg. Chem. 2008, 47, 2777−2785. (6) (a) Chen, J. M.; Wei, W.; Feng, X. L.; Lu, T. B. CO2 fixation and transformation by a dinuclear copper cryptate under acidic conditions. Chem. - Asian J. 2007, 2, 710−719. (b) Sorrell, T. N.; Allen, W. E.; White, P. S. Sterically hindered [tris(imidazolyl)phosphine]copper complexes: formation and reactivity of a peroxo-dicopper(II) adduct and structure of a dinuclear carbonate-bridged complex. Inorg. Chem. 1995, 34, 952−960. (c) Haines, R. J.; Wittrig, R. E.; Kubiak, C. P. Electrocatalytic reduction of carbon dioxide by the binuclear copper complex [Cu2 (6-(diphenylphosphino-2,2′-bipyridyl) 2(MeCN)2 ][PF6]2. Inorg. Chem. 1994, 33, 4723−4728. (7) (a) Lee, C. H.; Laitar, D. S.; Mueller, P.; Sadighi, J. P. Generation of a doubly bridging CO2 ligand and deoxygenation of CO2 by an (NHC)Ni(0) complex. J. Am. Chem. Soc. 2007, 129, 13802−13803. (b) Ariafard, A.; Brookes, N. J.; Stranger, R.; Boyd, P. D.; Yates, B. F. Density functional theory study on the mechanism of the reductive cleavage of CO2 by a bis-β-diketoiminatediiron dinitrogen complex. Inorg. Chem. 2010, 49, 7773−7782. (8) (a) Evans, W. J.; Seibel, C. A.; Ziller, J. W. Organosamariummediated transformations of CO2 and COS: monoinsertion and disproportionation reactions and the reductive coupling of CO2 to [O2CCO2]2‑. Inorg. Chem. 1998, 37, 770−776. (b) Davies, N. W.; Frey, A. S. P.; Gardiner, M. G.; Wang, J. Reductive disproportionation of carbon dioxide by a Sm(II) complex: unprecedented f-block element reactivity giving a carbonate complex. Chem. Commun. 2006, 4853−4855. (c) Castro, L.; Kefalidis, C. E.; McKay, D.; Essafi, S.; Perrin, L.; Maron, L. Theoretical treatment of one electron redox transformation of a small molecule using f-element complexes. Dalton Trans. 2014, 43, 12124−12134. (9) (a) Lam, O. P.; Bart, S. C.; Kameo, H.; Heinemann, F. W.; Meyer, K. Insights into the mechanism of carbonate formation through reductive cleavage of carbon dioxide with low-valent uranium centers. Chem. Commun. 2010, 46, 3137−3139. (b) Summerscales, O. T.; Frey, A. S. P.; Cloke, F. G. N.; Hitchcock, P. B. Reductive disproportionation of carbon dioxide to carbonate and squarate products using a mixed-sandwich U(III) complex. Chem. Commun. 2008, 198−200. (c) Schmidt, A. C.; Heinemann, F. W.; Kefalidis, C. E.; Maron, L.; Roesky, P. W.; Meyer, K. Activation of SO2 and CO2 by trivalent uranium leading to sulfite/dithionite and carbonate/oxalate complexes. Chem. - Eur. J. 2014, 20, 13501−13506. (d) Tsoureas, N.; Castro, L.; Kilpatrick, A. F. R.; Cloke, F.; Geoffey, N.; Maron, L. Controlling selectivity in the reductive activation of CO2 by mixed sandwich uranium(III) complexes. Chem. Sci. 2014, 5, 3777−3788. (e) Schmidt, A. C.; Nizovtsev, A. V.; Scheurer, A.; Heinemann, F. W.; Meyer, K. Uranium-mediated reductive conversion of CO2 to CO and carbonate in a single-vessel, closed synthetic cycle. Chem. Commun. 2012, 48, 8634−8636. (f) Castro, L.; Lam, O. P.; Bart, S. C.; Meyer, K.; Maron, L. Carbonate formation from CO2 via oxo versus oxalate pathway: theoretical investigations into the mechanism of uranium-mediated carbonate formation. Organometallics 2010, 29, 5504−5510. (10) (a) Lalrempuia, R.; Stasch, A.; Jones, C. The reductive disproportionation of CO2 using a magnesium(I) complex: analogies with low valent f-block chemistry. Chem. Sci. 2013, 4, 4383−4388. (b) Kefalidis, C. E.; Stasch, A.; Jones, C.; Maron, L. On the mechanism of the reaction of a magnesium(I) complex with CO2: a concerted type of pathway. Chem. Commun. 2014, 50, 12318−12321. (11) (a) Angamuthu, R.; Byers, P.; Lutz, M.; Spek, A. L.; Bouwman, E. Electrocatalytic CO2 conversion to oxalate by a copper complex. Science 2010, 327, 313−315. (b) Farrugia, L. J.; Lopinski, S.; Lovatt, P. A.; Peacock, R. D. Fixing carbon dioxide with copper: crystal structure of [LCu(μ-C2O4)CuL][Ph4B]2 (L=N,N′,N″-Triallyl-1,4,7-triazacyclo-

(3) (a) Appel, A. M.; et al. Frontiers, opportunities, and challenges in biochemical and chemical catalysis of CO2 fixation. Chem. Rev. 2013, 113, 6621−6658. (b) Jessop, P. G.; Ikariya, T.; Noyori, R. Homogeneous hydrogenation of carbon dioxide. Chem. Rev. 1995, 95, 259−272. (c) Wang, W.; Wang, S.; Ma, X.; Gong, J. Recent advances in catalytic hydrogenation of carbon dioxide. Chem. Soc. Rev. 2011, 40, 3703−3727. (d) Oh, Y.; Hu, X. Organic molecules as mediators and catalysts for photocatalytic and electrocatalytic CO2 reduction. Chem. Soc. Rev. 2013, 42, 2253−2261. (e) Schneider, J.; Jia, H.; Muckerman, J. T.; Fujita, E. Thermodynamics and kinetics of CO2, CO, and H+ binding to the metal centre of CO2 reduction catalysts. Chem. Soc. Rev. 2012, 41, 2036−2051. (f) Liu, Q.; Wu, L.; Jackstell, R.; Beller, M. Using carbon dioxide as a building block in organic synthesis. Nat. Commun. 2015, 6, 5933. (g) Costentin, C.; Robert, M.; Savéant, J. M. Catalysis of the electrochemical reduction of carbon dioxide. Chem. Soc. Rev. 2013, 42, 2423−2436. (h) Fan, T.; Chen, X.; Lin, Z. Theoretical studies of reactions of carbon dioxide mediated and catalysed by transition metal complexes. Chem. Commun. 2012, 48, 10808−10828. (i) Windle, C. D.; Perutz, R. N. Advances in molecular photocatalytic and electrocatalytic CO2 reduction. Coord. Chem. Rev. 2012, 256, 2562−2570. (j) Nakano, R.; Ito, S.; Nozaki, K. Copolymerization of carbon dioxide and butadiene via a lactone Intermediate. Nat. Chem. 2014, 6, 325−331. (4) (a) Yin, X. L.; Moss, J. R. Recent developments in the activation of carbon dioxide by metal complexes. Coord. Chem. Rev. 1999, 181, 27−59. (b) Leitner, W. The coordination chemistry of carbon dioxide and its relevance for catalysis: a critical survey. Coord. Chem. Rev. 1996, 153, 257−284. (c) Braunstein, P.; Matt, D.; Nobel, D. Reactions of carbon dioxide with carbon-carbon bond formation catalyzed by transition-metal complexes. Chem. Rev. 1988, 88, 747−764. (d) Darensbourg, D. J. Making plastics from carbon dioxide: salen metal complexes as catalysts for the production of polycarbonates from epoxides and CO2. Chem. Rev. 2007, 107, 2388−2410. (e) Mondal, B.; Song, J.; Neese, F.; Ye, S. Bio-inspired mechanistic insights into CO2 reduction. Curr. Opin. Chem. Biol. 2015, 25, 103−109. (f) Song, J.; Klein, E. L.; Neese, F.; Ye, S. The mechanism of homogeneous CO2 reduction by Ni(cyclam): product selectivity, concerted protonelectron transfer and C−O bond cleavage. Inorg. Chem. 2014, 53, 7500−7507. (g) Hull, J. F.; Himeda, Y.; Wang, W.-H.; Hashiguchi, B.; Periana, R.; Szalda, D. J.; Muckerman, J. T.; Fujita, E. Reversible hydrogen storage using CO2 and a proton-switchable iridium catalyst in aqueous media under mild temperatures and pressures. Nat. Chem. 2012, 4, 383−388. (h) Tanaka, R.; Yamashita, M.; Nozaki, K. Catalytic hydrogenation of carbon dioxide using Ir(III)-pincer complexes. J. Am. Chem. Soc. 2009, 131, 14168−14169. (i) Fan, T.; Sheong, F. K.; Lin, Z. Y. DFT studies on copper-catalyzed hydrocarboxylation of alkynes using CO2 and hydrosilanes. Organometallics 2013, 32, 5224−5230. (j) Yuan, R.; Lin, Z. Y. Mechanism for the carboxylative coupling reaction of a terminal alkyne, CO2, and an allylic chloride catalyzed by the Cu(I) complex: a DFT study. ACS Catal. 2014, 4, 4466−4473. (k) Yang, X. Z. Hydrogenation of carbon dioxide catalyzed by PNP pincer iridium, iron, and cobalt complexes: a computational design of base metal catalysts. ACS Catal. 2011, 1, 849−854. (l) Ohnishi, Y.-Y.; Matsunaga, T.; Nakao, Y.; Sato, H.; Sakaki, S. Ruthenium(II)-catalyzed hydrogenation of carbon dioxide to formic acid. Theoretical study of real catalyst, ligand effects, and solvation effects. J. Am. Chem. Soc. 2005, 127, 4021−4032. (m) Musashi, Y.; Sakaki, S. Theoretical study of rhodium(III)-catalyzed hydrogenation of carbon dioxide into formic acid. Significant differences in reactivity among rhodium(III), rhodium(I), and ruthenium(II) complexes. J. Am. Chem. Soc. 2002, 124, 7588−7603. (n) Musashi, Y.; Sakaki, S. Theoretical study of ruthenium-catalyzed hydrogenation of carbon dioxide into formic acid. Reaction mechanism involving a new type of σ-bond metathesis. J. Am. Chem. Soc. 2000, 122, 3867−3877. (o) Zhou, Q.; Li, Y. The real role of N-heterocyclic carbene in reductive functionalization of CO2: an alternative understanding from density functional theory study. J. Am. Chem. Soc. 2015, 137, 10182−10189. (p) Agarwal, J.; Fujita, E.; Schaefer, H. F., III; Muckerman, J. T. Mechanisms for CO production 6817


Article

Inorganic Chemistry nonane). Inorg. Chem. 2001, 40, 558−559. (c) Takisawa, H.; Morishima, Y.; Soma, S.; Szilagyi, R. K.; Fujisawa, K. Conversion of carbon dioxide to oxalate by α-ketocarboxylatocopper(II) complexes. Inorg. Chem. 2014, 53, 8191−8193. (d) Pokharel, U. R.; Fronczek, F. R.; Maverick, A. W. Reduction of carbon dioxide to oxalate by a binuclear copper complex. Nat. Commun. 2014, 5, 5883. (12) Tanaka, K.; Kushi, Y.; Tsuge, K.; Toyohara, K.; Nishioka, T.; Isobe, K. Catalytic generation of oxalate through a coupling reaction of two CO2 molecules activated on [(Ir(η5-C5Me5))2(Ir(η4-C5Me5)CH2CN)(μ3-S)2]. Inorg. Chem. 1998, 37, 120−126. (13) (a) Lu, C. C.; Saouma, C. T.; Day, M. W.; Peters, J. C. Fe(I)mediated reductive cleavage and coupling of CO2: an FeII(μ-O,μCO)FeII core. J. Am. Chem. Soc. 2007, 129, 4−5. (b) Saouma, C. T.; Lu, C. C.; Day, M. W.; Peters, J. C. CO2 reduction by Fe(I): solvent control of C−O cleavage versus C−C coupling. Chem. Sci. 2013, 4, 4042−4051. (14) Horn, B.; Limberg, C.; Herwig, C.; Braun, B. Nickel(I)mediated transformations of carbon dioxide in closed synthetic cycles: reductive cleavage and coupling of CO2 generating NiICO, NiIICO3 and NiIIC2O4NiII entities. Chem. Commun. 2013, 49, 10923−10925. (15) Wong, W. K.; Zhang, L. L.; Xue, F.; Mak, C. W. Synthesis and X-ray crystal structure of an unexpected neutral oxalate-bridged ytterbium(III) porphyrinate dimer. J. Chem. Soc., Dalton Trans. 2000, 2245−2246. (16) (a) Tanaka, R.; Yamashita, M.; Chung, L. W.; Morokuma, K.; Nozaki, K. Mechanistic studies on the reversible hydrogenation of carbon dioxide catalyzed by an Ir-PNP complex. Organometallics 2011, 30, 6742−6750. (b) Chung, L. W.; Chan, T. H.; Wu, Y.-D. Theoretical study of the intrinsic reactivities of various allylmetals toward carbonyls and water. Organometallics 2005, 24, 1598−1607. (c) Miao, W.; Chung, L. W.; Wu, Y.-D; Chan, T. H. Experimental and theoretical studies of the propargyl-allenylindium system. J. Am. Chem. Soc. 2004, 126, 13326−13334. (d) Xu, L.; Chung, L. W.; Wu, Y.-D. Mechanism of Ni-NHC catalyzed hydrogenolysis of aryl ethers: roles of the excess base. ACS Catal. 2016, 6, 483−493. (17) Arabic number in superscript denotes spin multiplicity of the system (1, closed-shell singlet; 3, triplet; 5, quintet) except that “oss” in superscript represents an open-shell singlet configuration. (18) Fully delocalized mixed-valence Cu centers in a few bioinorganic systems: (a) Solomon, E. I.; Heppner, D. E.; Johnston, E. M.; Ginsbach, J. W.; Cirera, J.; Qayyum, M.; Kieber-Emmons, M. T.; Kjaergaard, C. H.; Hadt, R. G.; Tian, L. Copper active sites in biology. Chem. Rev. 2014, 114, 3659−3853. (b) Liu, J.; Chakraborty, S.; Hosseinzadeh, P.; Yu, Y.; Tian, S.; Petrik, I.; Bhagi, A.; Lu, Y. Metalloproteins containing cytochrome, iron-sulfur, or copper redox centers. Chem. Rev. 2014, 114, 4366−4469. (c) Kim, E.; Chufán, E. E.; Kamaraj, K.; Karlin, K. D. Synthetic models for heme-copper oxidases. Chem. Rev. 2004, 104, 1077−1134. (19) Frisch, M. J.; et al. Gaussian 09, revision D.01; Gaussian, Inc.: Wallingford, CT, 2009. (20) (a) Chung, L. W.; Li, X.; Hirao, H.; Morokuma, K. Comparative reactivity of ferric-superoxo and ferryl-oxo species in heme and nonheme complexes. J. Am. Chem. Soc. 2011, 133, 20076−20079. (b) Chung, L. W.; Li, X.; Sugimoto, H.; Shiro, Y.; Morokuma, K. ONIOM study on a missing piece in our understanding of heme chemistry: bacterial tryptophan 2,3-dioxygenase with dual oxidants. J. Am. Chem. Soc. 2010, 132, 11993−12005. (c) Chung, L. W.; Li, X.; Sugimoto, H.; Shiro, Y.; Morokuma, K. Density functional theory study on a missing piece in understanding of heme chemistry: the reaction mechanism for indoleamine 2,3-dioxygenase and tryptophan 2,3-dioxygenase. J. Am. Chem. Soc. 2008, 130, 12299−12309. (d) Shaik, S.; Cohen, S.; Wang, Y.; Chen, H.; Kumar, D.; Thiel, W. P450 enzymes: their structure, reactivity, and selectivity-modeled by QM/MM calculations. Chem. Rev. 2010, 110, 949−1017. (e) Blomberg, M. R. A.; Borowski, T.; Himo, F.; Liao, R.-Z.; Siegbahn, P. E. M. Quantum chemical studies of mechanisms for metalloenzymes. Chem. Rev. 2014, 114, 3601−3658. (f) Funes-Ardoiz, I.; Maseras, F. Cooperative reductive elimination: the missing piece in the

oxidative-coupling mechanistic puzzle. Angew. Chem., Int. Ed. 2016, 55, 2764−2767. (21) (a) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. Energy-adjusted ab initio pseudopotentials for the first row transition elements. J. Chem. Phys. 1987, 86, 866−872. (b) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Energy-adjustedab initio pseudopotentials for the second and third row transition elements. Theor. Chim. Acta. 1990, 77, 123−141. (22) (a) Zhao, Y.; Truhlar, D. G. Density functionals with broad applicability in chemistry. Acc. Chem. Res. 2008, 41, 157−167. (b) Zhao, Y.; Truhlar, D. G. Attractive noncovalent interactions in the mechanism of Grubbs second-generation Ru catalysts for olefin metathesis. Org. Lett. 2007, 9, 1967−1970. (c) Averkiev, B. B.; Truhlar, D. G. Free energy of reaction by density functional theory: oxidative addition of ammonia by an iridium complex with PCP pincer ligands. Catal. Sci. Technol. 2011, 1, 1526−1529. (d) Minenkov, Y.; Occhipinti, G.; Jensen, V. R. Metal-phosphine bond strengths of the transition metals: a challenge for DFT. J. Phys. Chem. A 2009, 113, 11833− 11844. (e) Gusev, D. G. Assessing the accuracy of M06-L organometallic thermochemistry. Organometallics 2013, 32, 4239− 4243. (23) Fukui, K. Formulation of the reaction coordinate. J. Phys. Chem. 1970, 74, 4161−4163. (24) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions. J. Phys. Chem. B 2009, 113, 6378−6396. (25) (a) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 2010, 132, 154104. (b) Grimme, S.; Hansen, A.; Brandenburg, J. G.; Bannwarth, C. Dispersion-corrected mean-field electronic structure methods. Chem. Rev. 2016, 116, 5105−5154. (26) (a) Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Atoms, molecules, solids, and surfaces: applications of the generalized gradient approximation for exchange and correlation. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 46, 6671−6687. (b) Perdew, J. P.; Wang, Y. Accurate and simple analytic representation of the electron-gas correlation energy. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 45, 13244−13249. (c) Perdew, J. P.; Burke, K.; Wang, Y. Generalized gradient approximation for the exchange-correlation hole of a many-electron system. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 16533− 16539. (27) (a) Chai, J.-D.; Head-Gordon, M. Systematic optimization of long-range corrected hybrid density functionals. J. Chem. Phys. 2008, 128, 084106−15. (b) Adamo, C.; Barone, V. Toward reliable density functional methods without adjustable parameters: The PBE0 model. J. Chem. Phys. 1999, 110, 6158−6169. (c) Ernzerhof, M.; Scuseria, G. E. Assessment of the Perdew-Burke-Ernzerhof exchange-correlation functional. J. Chem. Phys. 1999, 110, 5029−5036. (d) Becke, A. D. A new mixing of Hartree-Fock and local density-functional theories. J. Chem. Phys. 1993, 98, 1372−1377. (e) Becke, A. D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (f) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (g) Vosko, S. H.; Wilk, L.; Nusair, M. Accurate spindependent electron liquid correlation energies for local spin density calculations: a critical analysis. Can. J. Phys. 1980, 58, 1200−1211. (h) Becke, A. D. Density-functional exchange-energy approximation with correct asymptotic behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (28) Yamaguchi, K.; Jensen, F.; Dorigo, A.; Houk, K. N. A spin correction procedure for unrestricted Hartree-Fock and Møller-Plesset wavefunctions for singlet diradicals and polyradicals. Chem. Phys. Lett. 1988, 149, 537−542. (29) (a) Cheng, G.-J.; Zhang, X.; Chung, L. W.; Xu, L.; Wu, Y.-D. Computational organic chemistry: bridging theory and experiment in 6818


Article

Inorganic Chemistry establishing the mechanisms of chemical reactions. J. Am. Chem. Soc. 2015, 137, 1706−1725. (b) Plata, R. E.; Singleton, D. A. A case study of the mechanism of alcohol-mediated Morita Baylis-Hillman reactions. The importance of experimental observations. J. Am. Chem. Soc. 2015, 137, 3811−3826. (c) Liang, Y.; Liu, S.; Xia, Y.; Li, Y.; Yu, Z.-X. Mechanism, regioselectivity, and the kinetics of phosphine-catalyzed [3 + 2] cycloaddition reactions of allenoates and electron-deficient alkenes. Chem. - Eur. J. 2008, 14, 4361−4373. (30) (a) Benson, S. W. The Foundations of Chemical Kinetics; Krieger: Malabar, FL, 1982. (b) Okuno, Y. Theoretical investigation of the mechanism of the Baeyer-Villiger reaction in nonpolar solvents. Chem. - Eur. J. 1997, 3, 212−218. (31) (a) Ardura, D.; López, R.; Sordo, T. L. Relative Gibbs energies in solution through continuum models: effect of the loss of translational degrees of freedom in bimolecular reactions on Gibbs energy barriers. J. Phys. Chem. B 2005, 109, 23618−23623. (b) Schoenebeck, F.; Houk, K. N. Ligand-controlled regioselectivity in palladium-catalyzed cross coupling reactions. J. Am. Chem. Soc. 2010, 132, 2496−2497. (c) Liu, Q.; Lan, Y.; Liu, J.; Li, G.; Wu, Y.-D.; Lei, A. Revealing a second transmetalation step in the Negishi coupling and its competition with reductive elimination: improvement in the interpretation of the mechanism of biaryl syntheses. J. Am. Chem. Soc. 2009, 131, 10201−10210. (d) Wang, M. Y.; Fan, T.; Lin, Z. Y. DFT studies on copper-catalyzed arylation of aromatic C−H bonds. Organometallics 2012, 31, 560−569. (e) Ariafard, A.; Brookes, N. J.; Stranger, R.; Yates, B. F. DFT study on the mechanism of the activation and cleavage of CO2 by (NHC)CuEPh3 (E= Si, Ge, Sn). Organometallics 2011, 30, 1340−1349. (32) A formal Cu(II)(CO2•−) radical anion intermediate 3II is formed (spin density of Cu/CO2: 0.86/0.85), but 3II is much higher in free energy than 1II by 25.0 kcal/mol. (33) (a) Zhang, S.; Fallah, H.; Gardner, E. J.; Kundu, S.; Bertke, J. A.; Cundari, T. R.; Warren, T. H. A dinitrogen dicopper(I) complex via a mixed-valence dicopper hydride. Angew. Chem., Int. Ed. 2016, 55, 9927−9931. (b) Houser, R. P.; Young, V. G., Jr.; Tolman, W. B. A thiolate-bridged, fully delocalized mixed-valence dicopper(I, II) complex that models the CuA biological electron-transfer site. J. Am. Chem. Soc. 1996, 118, 2101−2102. (c) Harding, C.; McKee, V.; Nelson, J. Highly delocalized Cu(I)/Cu(II): a copper-copper bond? J. Am. Chem. Soc. 1991, 113, 9684−9685. (d) Barr, M. E.; Smith, P. H.; Antholine, W.; Spencer, B. Crystallographic, spectroscopic and theoretical studies of an electron-deiocalized Cu(1.5)−Cu(1.5) complex. J. Chem. Soc., Chem. Commun. 1993, 1649−1652. (e) Harding, C.; Nelson, J.; Symons, M. C. R.; Wyatt, J. A coppercopper bond by intent. J. Chem. Soc., Chem. Commun. 1994, 2499− 2500. (f) LeCloux, D. D.; Davydov, R.; Lippard, S. J. Synthesis and characterization of spin-delocalized carboxylate-bridged Cu(I)-Cu(II) mixed-valence complexes having only oxygen donor ligands. J. Am. Chem. Soc. 1998, 120, 6810−6811. (g) Ziegler, M. S.; Levine, D. S.; Lakshmi, K. V.; Tilley, T. D. Aryl group transfer from tetraarylborato anions to an electrophilic dicopper(I) center and mixed-valence μ-aryl dicopper(I, II) complexes. J. Am. Chem. Soc. 2016, 138, 6484−6491. (h) Kababya, S.; Nelson, J.; Calle, C.; Neese, F.; Goldfarb, D. Electronic structure of binuclear mixed valence copper azacryptates derived from integrated advanced EPR and DFT calculations. J. Am. Chem. Soc. 2006, 128, 2017−2029. (34) Robin, M. B.; Day, P. Mixed valence chemistry-a survey and classification. Adv. Inorg. Chem. Radiochem. 1968, 10, 247−422. (35) Li, J.; Lin, Z. Y. Density Functional Theory Studies on the Reduction of CO2 to CO by a (NHC)Ni0 Complex. Organometallics 2009, 28, 4231−4234. (36) Moreover, we investigated two alternative possible paths for the oxalate pathway for the realistic catalyst (Figure S22) by adding one more CO2 molecule to the two nonreactive Cu sites (paths SC and SD). The computed barrier for path SD is very high (50.3 kcal/mol). (37) The relative free energy for 32 is the lowest (ΔGsoln,cor‑B3PW91‑D3 = −10.7 kcal/mol) at the B3PW91-D3//M06-L level, which is lower than 11L2s (−8.0 kcal/mol, Table S43) by 2.7 kcal/mol. Thus, the reference point should be changed to 32 instead of 11L2s.

(38) The charge and spin analysis on the small-model catalyst further elucidate that the C1 atom undergoes a (metal-mediated) nucleophilelike attack of the C2 atom in the second CO2 moiety in this oxalate pathway (Figure S1). In addition, the coupling step in the closed-shell singlet state should not be operative (Tables S1, S2, and S40). (39) (a) Personal communication with Prof. Elisabeth Bouwman. Her group prepared several mononuclear Cu compounds of similar ligands, but they did not observe the oxalate formation. They found that the bis-copper(I) disulfide complex has potentially different conformations, which may affect its reactivity. (b) Ording-Wenker, E. C. M.; van der Plas, M.; Siegler, M. A.; Bonnet, S.; Bickelhaupt, F. M.; Fonseca Guerra, C.; Bouwman, E. Thermodynamics of the CuII μThiolate and CuI Disulfide Equilibrium: A Combined Experimental and Theoretical Study. Inorg. Chem. 2014, 53, 8494−8504. (c) OrdingWenker, E. C. M.; Siegler, M. A.; Lutz, M.; Bouwman, E. CuI Thiolate Reactivity with Dioxygen: The Formation of CuII Sulfinate and CuII Sulfonate Species via a CuII Thiolate Intermediate. Inorg. Chem. 2013, 52, 13113−13122. (d) However, we do not completely override the possibility of mononuclear metal catalysts in the other systems (refs 8a, 8c, 11c, and 13 and Scheme S3). (40) At the B3PW91-D3//M06-L level, the barriers for the oxalate pathway (via 3TS3a−4a, 3TS3b−4b, and 5TS3b−4b) are lower in free energy than those for the corresponding carbocyclic pathway (via 3 TS9b1−10b1 and 3TS9b2−10b2) by at least 15.0 kcal/mol in solution (Tables S43 and S44). Additionally, the dicopper(II)−oxalate complexes (34a, 34b, and 54b) are more stable in free energy than those for the corresponding carbocyclic pathway (310a, 310b1, and 3 10b2) by at least 32.3 kcal/mol in solution (Tables S43 and S44). (41) Energy decomposition analysis: (a) Morokuma, K. Why do molecules interact? The origin of electron donor-acceptor complexes, hydrogen bonding and proton affinity. Acc. Chem. Res. 1977, 10, 294− 300. (b) Nagase, S.; Morokuma, K. An ab initio molecular orbital study of organic reactions. The energy, charge, and spin decomposition analyses at the transition state and along the reaction pathway. J. Am. Chem. Soc. 1978, 100, 1666−1672. (c) Ess, D. H.; Houk, K. N. Distortion/interaction energy control of 1, 3-dipolar cycloaddition reactivity. J. Am. Chem. Soc. 2007, 129, 10646−10647. (42) The S−S bond cleavage was observed (Figure S17).

6819


Reaction Mechanism of Cu(I)-Mediated Reductive CO2 Coupling for

Recommend Documents