Selective Transformation of CO2 to CO at a Single Nickel Center

Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Selective Transformation of CO2 to CO at a Single Nickel Center Changho Yoo, Yeong-Eun Kim, and Yunho Lee* Department of Chemistry, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic of Korea CONSPECTUS: Carbon dioxide conversion mediated by transition metal complexes continues to attract much attention because of its future potential utilization as a nontoxic and inexpensive C1 source for the chemical industry. Given the presence of nickel in natural systems that allow for extremely efficient catalysis, albeit in an Fe cluster arrangement, studies that focus on selective CO2 conversion with synthetic nickel species are currently of considerable interest in our group. In this Account, the selective conversion of CO2 to carbon monoxide occurring at a single nickel center is discussed. The chemistry is based on a series of related nickel pincer complexes with attention to the uniqueness of the coordination geometry, which is crucial in allowing for particular reactivity toward CO2. Our research is inspired by the efficient enzymatic CO2 catalysis occurring at the active site of carbon monoxide dehydrogenase. Since the binding and reactivity toward CO2 are controlled in part by the geometry of a L3Ni scaffold, we have explored the chemistry of low-valent nickel supported by PPMeP and PNP ligands, in which a pseudotetrahedral or square-planar geometry is accommodated. Two isolated nickel−CO2 adducts, (PPMeP)Ni(η2-CO2-κC) (2) and {Na(12-C-4)2}{(PNP)Ni(η1-CO2-κC)} (7), clearly demonstrate that the geometry of the nickel ion is crucial in the binding of CO2 and its level of activation. In the case of a square-planar nickel center supported by a PNP ligand, a series of bimetallic metallacarboxylate Ni−μ-CO2-κC,O−M species (M = H, Na, Ni, Fe) were synthesized, and their structural features and reactivity were studied. Protonation cleaves the C−O bond, resulting in the formation of a nickel(II) monocarbonyl complex. By sequential reduction, the corresponding mono- and zero-valent Ni−CO species were produced. The reactivities of three nickel carbonyl species toward various iodoalkanes and CO2 were explored to address whether their corresponding reactivities could be controlled by the number of valence d electrons. In particular, a (PNP)Ni(0)−CO species (13) shows immediate reactivity toward CO2 but displays multiple product formation. By incorporation of a −CMe2− bridging unit, a structurally rigidified acriPNP ligand was newly designed and produced. This ligand modification was successful in preparing the T-shaped nickel(I) metalloradical species 9 exhibiting open-shell reactivity due to the sterically exposed nickel center possessing a half-filled dx2−y2 orbital. More importantly, the selective addition of CO2 to a nickel(0)−CO species was enabled to afford a nickel(II)−carboxylate species (22) with the expulsion of CO(g). Finally, the (acriPNP)Ni system provides a synthetic cycle in the study of the selective conversion of CO2 to CO that involves two-electron reduction of Ni−CO followed by the direct addition of CO2 to release the coordinated CO ligand. 1. 10 The intermediate species C red2 -CO 2 exhibits CO 2 coordination at the nickel site with a Ni−C bond.10 One of the oxygen atoms binds to Feu, known as the unique iron.9,10 Although a series of crystallographic snapshots of the CO2 reduction cycle have become available, crucial questions concerning (i) the role of the iron−sulfur cluster, (ii) the oxidation states of nickel and iron during CO2 binding, and (iii) the driving force for the expulsion of CO from the nickel center still need to be answered.11 These specific aspects of biological CO2 conversion and the fact that nickel is an earth-abundant metal led to our interest in single-site nickel chemistry. Ligand constraint and donor atom choice were the main design parameters of this structural and reactivity study. Going into the study, we had an interest in bringing nickel pincer chemistry to bear on the selectivity issue

1. INTRODUCTION Carbon dioxide conversion has recently been drawing much attention because of its central relevance to the environmental, energy-related, and sustainable chemical industries.1,2 As a synthetic C1 source, utilization of CO2 is being widely explored to produce industrially useful chemicals such as oxalate,3 formate,4 cyclic/polycarbonates,5 carbon monoxide,6 and methanol.4,7 The formation of such products requires multielectron equivalents, which can be effectively delivered via coordination of CO2 to a metal center.2 The main challenge in transition-metal-based catalysis is to increase the selectivity in producing a desired product, which may rely on the interaction of CO2 with a metal center.6a,8 Efficient and selective conversion of CO2 to CO occurs biologically in the active site of the metalloenzyme carbon monoxide dehydrogenase (CODH).9 According to recent X-ray crystallographic studies, the CODH active site possesses a single nickel center associated with the iron−sulfur cluster, as depicted in Figure © XXXX American Chemical Society

Received: December 23, 2017

A

DOI: 10.1021/acs.accounts.7b00634 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. Proposed mechanism of the conversion of CO2 to CO in the active site of CODH. The known intermediate species (left) Cred2-CO2, (center) Cred1-CO and (right) Cred2 are shown. Reduction steps have been omitted because of ambiguity.

Figure 2. (PEP)Ni systems (E = P, N).

on the activated CO2 moiety. Protonation of the nickel−CO2 adduct allows for cleavage of the C−O bond and affords a Ni− CO unit. In the presence of a π-acidic CO ligand, the nickel(II) carbonyl complex was reliably reduced. By the addition of one or two electrons to Ni(II)−CO, nickel(I/0) monocarbonyl complexes were generated; such carbonyl complexes show interesting structural features and reactivity. Finally, successful expulsion of CO was accomplished at a nickel(0) center supported by an acriPNP ligand, suggesting that structural rigidity can be one of the catalyst design principles in selective CO2 conversion to CO.20

having a neutral phosphine or anionic amide as the central donor moiety to direct the geometry about the nickel center, PPMeP (PPMeP = PMe[2-PiPr2−C6H4]2), PNP (PNP− = N[2PiPr2-4-Me-C 6 H3]2 −), and acriPNP ( acriPNP − = 4,5-bis(diisopropylphosphino)-2,7,9,9-tetramethyl-9H-acridin-10-ide), were systematically utilized. Although there are fairly big differences in their ligand properties, instead of three sulfur donors, P and N donor atoms were used to effectively control the geometry of nickel complexes within a range of synthetic capability and convenience. In this Account, we describe our recent results from synthesis, characterization, and reactivity studies of nickel complexes supported by PEP ligands. The central donor atom strongly affects nickel coordination, crucial in its reactivity toward CO2. Nickel(II) supported by a PNP ligand favors a square-planar structure;14−18 however, its significant distortion toward a tetrahedral geometry occurs when a neutral phosphine is used as the central donor moiety.12,13 This geometric variation subsequently impacts the nickel−CO2 binding and its conversion. We first discuss CO2 binding at a nickel center and describe the observed two-electron reduction of CO2 imposed

2. MONONUCLEAR Ni−CO2 BINDING MODES By control of the local environment of the metal center, certain CO2 coordination modes can be allowed, and ultimately the selective conversion of CO2 to desired products can be accomplished. Several examples of modes of CO2 coordination to a single metal center are known, and these geometries provide crucial mechanistic snapshots for the activation of carbon dioxide.21 The first reported transition metal−CO2 adduct is Aresta’s nickel−CO2 complex (PCy3)2Ni(η2-CO2), as depicted in Figure 3.22 This nickel complex shows an η2-CO2 binding mode in which both carbon and oxygen of CO2 are involved in the coordination.22 By the use of the ligand 1,2bis(di-tert-butylphosphino)ethane (dtbpe), the analogous nickel CO2 complex (dtbpe)Ni(η2-CO2) (1) was recently reported by the Hillhouse group.23 Both examples demonstrate η2-CO2 coordination at a low-valent nickel species when two P donors are available. In 2014, we reported the reversible binding of CO2 to the four-coordinate nickel(0)−N2 complex {(PPMeP)Ni}2(μ-N2) using the neutral PPMeP ligand, as shown in Figure 3.12 Upon addition of CO2 to {(PPMeP)Ni}2(μ-N2), the five-coordinate

in enabling the conversion of CO2 to CO. We have utilized two PEP ligand systems (E = P, N) mainly equipped with two phosphine side arms, as depicted in Figure 2.12−20 Ligands

Figure 3. Three nickel−CO2 adducts, (PCy3)2Ni(η2-CO2) and 1 (top) and 2 generated from the reaction of {(PPMeP)Ni}2(μ-N2) with CO2 (bottom). The reaction of 2 with B(C6F5)3 to give the borane adduct 3 is also shown. B


Article


Figure 4. Comparison of selected bond lengths (Å) and angles (deg) in the core structures of nickel−CO2 adducts 1, 2, 3, 6, and 7.

Table 1. Selected Experimental Bond Lengths, Angles, and Vibrational Frequencies for CO2 in 1, 2, 3, 4, 5, 6, 7, 8, and 10 (dtbpe)Ni(CO2) (1) (PPMeP)Ni(CO2) (2) (PPMeP)Ni(COOB(C6F5)) (3) (PNP)NiCOOH (4) (PNP)NiCOONa (5) (PNP)NiCOONi(PNP) (6) {Na(12-C-4)2} {(PNP)NiCOO} (7) (PNP)NiCOOFe(PNP) (8) (acriPNP)NiCOONi(acriPNP) (10)a a

νCO2 (cm−1)

Ni−C (Å)

Ni−O (Å)

C−O (Å)

O−C−O (deg)

P−Ni−P (deg)

1724 1682 1631 1582 ∼1602 1518 1620 1510 1511

1.868(2) 1.904(1) 1.923(3) 1.866(2) 1.882(1) 1.888(2) 1.911(2) 1.858(1) 1.94(2), 1.94(1)

1.904(2) 2.191(1) − − − 1.897(2) 2.614(1), 2.776(1) 2.718(1), 2.792(1) 1.942(9), 1.923(9)

1.200(3), 1.266(3) 1.218(2), 1.252(2) 1.223(4), 1.340(4) 1.269(3), 1.313(3) 1.260(1), 1.271(1) 1.240(3), 1.296(3) 1.248(2), 1.247(2) 1.269(2), 1.289(2) 1.22(1), 1.34(2), 1.24(2), 1.24(1)

138.0(2) 135.1(1) 122.9(3) 119.6(2) 124.0(1) 123.7(2) 128.4(2) 116.5(1) 122(1), 121(1)

92.80(3) 143.03(2) 154.43(4) 173.69(3) 170.61(1) 168.98(2) 169.51(2) 170.00(2) 169.95(5), 171.27(5)

Two data sets are presented because of the disorder and modeling of the CO2 moiety over two distinct positions.

Figure 5. Sequential conversion of CO2 to CO22− at a single nickel center and further interaction with a Lewis acidic Fe ion.

nickel−CO2 adduct (PPMeP)Ni(η2-CO2) (2) was immediately generated; crystallographic characterization of 2 allowed for an important comparison. It is interesting that the nickel ion supported by three P donors displays a fairly similar CO2 binding mode as in Aresta’s and Hillhouse’s complexes. The nickel−oxygen bond length in 2 is noticeably elongated. In compound 1, the Ni−C and Ni−O bond lengths are similar: 1.868(2) and 1.904(2) Å, respectively. However, compound 2 shows an elongated Ni−O bond with a length of 2.191(1) Å, while the Ni−C bond length is 1.904(1) Å. This indicates that one more P donor causes the Ni−O bond to elongate, thus allowing oxygen to be available for electrophilic attack.12 While the CO2 moiety in 1 involves both the coordinating carbon and oxygen atoms in the plane of (dtbpe)Ni, compound 2 features a CO2 moiety that is oriented differently: it is orthogonal to the (PPMeP)Ni coordination plane. According to DFT analysis, the HOMO of 1 presents clear evidence of bonding between Ni and two donor atoms, C and O.12 The HOMO of 2, however, has a significant antibonding character between Ni and O. This vulnerability is displayed in its reactivity toward Lewis acidic borane. Upon addition of tris(pentafluorophenyl)borane to 2, the nickel carboxylate species (PPMeP)Ni(COOB(C6F5)3) (3) was instantaneously produced, supporting that the oxygen atom in 2 is more accessible toward electrophilic attack.12 The degree of CO2 activation can be evaluated using infrared spectroscopic data and the bond angle (Figure 4). The CO2 asymmetric vibration in 2 was assigned as 1682 cm−1, lower than the value of 1724 cm−1 in 1 and higher than the value of 1639 cm−1 in 3. With the O−C−O bond angle of 135.1(1)°,

the central carbon atom of CO2 in 2 is clearly perturbed from sp hybridization by Ni coordination. The similar O−C−O angle of 138.0(2)° in 1 suggests that the CO2 moieties in complexes 1 and 2 are similarly activated. In the presence of borane, 3 adopts a contracted O−C−O angle of 122.9(3)°, indicating that the central carbon conclusively presents sp2 hybridization.12 Furthermore, the two C−O bond lengths in 3 are clearly different by ∼0.12 Å, meaning that one of the CO bonds has a single-bond character, while 1 and 2 show a difference of around 0.05 Å, which is not enough to show a localized C−O single bond character. The core structure of 2 is significantly changed to distorted square-planar in 3 (see Table 1). This indicates that compound 3 has a nickel(II) ion with a dianionic CO22− ligand. The five-coordinate nickel−CO2 adduct 2 shows an η2 binding mode of CO2 at a single nickel center. Such a binding mode may be related to the initial interaction of CO2 with the nickel center in the CODH active site. The following step should involve electron transfer from the nickel d orbital to the CO2 π* orbital. One can anticipate that the direct addition of a Lewis acid such as Fe can facilitate such an electron transfer process, as shown in the case of the reaction of 2 with borane. Since such redox reactions can be coupled with the geometry of nickel, the structural change from tetrahedral to square planar as depicted in Figure 5 may be coupled with the corresponding electron transfer. C


Article


3. SYNTHESIS OF Ni−CO2−M COMPLEXES By the introduction of a PNP ligand, a series of nickel(II) carboxylate complexes were generated. From the reaction of (PNP)NiOH with 1 equiv of CO(g), the nickel(II) hydroxycarbonyl complex (PNP)Ni(COOH-κC) (4) was prepared.14 The transition metal hydroxycarbonyl species is well-known in the literature as an important intermediate in the CODH reaction and electrocatalytic CO2 reduction.6,8,9,24 By deprotonation of 4, (PNP)NiCOONa (5) was generated, and its reaction with (PNP)Ni−OH produces the dinuclear nickel(II) complex (PNP)Ni−μ-COO−Ni(PNP) (6), as shown in Figure 6.14 The Ni−COO moiety in 4−6 consistently

successfully synthesized.18 In this compound, the CO2 moiety is coordinated to a square-planar nickel(II) center via a Ni−η1CO2-κC mode similar to that seen in 7. The CO2 moiety exhibits a CO vibration at 1510 cm−1 with an O−C−O angle of 116.5(1)°, which is fairly close to that in CO22− and similar to that in 4−6. Accordingly, the two C−O bond distances are slightly elongated relative to those of 7, suggesting that more reduction occurs at the CO2 moiety upon addition of the Fe moiety. The sodium-free adduct 7 has the same number of valence electrons in the Ni−CO2 moiety as in 2, but it possesses a square-planar Ni(II) ion with a CO22− moiety even without a Lewis acid. In terms of the activation of CO2, 7 has a more reduced CO2 moiety relative to that in 2. The O−C−O angle of 128.4(2)° in 7 is clearly smaller than the angle of 135.1(1)° in 2, and its C−O bond distances of 1.247(2) and 1.248(2) Å are slightly longer than those of 1.218(2) and 1.252(2) Å in 2. This trend also can be seen in the CO vibrational frequency, which appears at a lower value for 7 than for 2 (1620 vs 1682 cm−1). This difference is in part due to the different geometries supported by the PEP ligands: while the anionic PNP ligand encourages a square-planar geometry, the neutral PPMeP ligand facilitates a distorted tetrahedral geometry. More recently, we have also reported a new PNP ligand possessing a structurally rigidified backbone that allows the preparation of a T-shaped Ni(I) metalloradical species, as shown in Figure 7.19 The T-shaped nickel(I) species (acriPNP)Ni(I) (9) having a sterically exposed nickel center possessing a half-filled dx2−y2 orbital was generated from (acriPNP)NiCl. Its bond formation with various substrates is efficiently coupled with an inner-sphere single electron transfer resulting in the reduction of unsaturated molecules such as C2H4 and CO2 and homolytic cleavage of various σ bonds, including H−H, H2N− NH2 and H3C−CN. This metalloradical species employs a cooperative binuclear reaction to insert CO2 into two nickel(II) centers, producing the binuclear NiII species {(acriPNP)Ni}2(μCO2) (10). This reaction occurs instantaneously at room temperature, thus suggesting that the preorganized nickel(I) center may provide a reaction route with a lower activation barrier for the binding and reduction of CO2, and such a fast reaction may require an electron donor such as Feu to be engaged in a single electron transfer to give a Ni−μ-CO2−Fe moiety.

Figure 6. Reaction scheme for producing 7 along with other nickel(II) carboxylate species.

possesses a nickel(II) ion with a P−Ni−P angle of ∼170° and CO22− with an O−C−O angle of ∼122°. The CO vibrational frequency of the CO2 moiety is found in the range of 1550− 1580 cm−1, which is markedly lower than those of Ni−η2-CO2 adducts 1 and 2. The structural and spectroscopic data for 4−6 suggest that a Lewis acid provides additional influence to stabilize a nickel(II) ion with a dianionic CO22−. Two equivalents of 12-crown-4 (12-C-4) was added to the solution of 5 to give the Lewis acid-free adduct {Na(12-C4)2}{(PNP)Ni−CO2} (7) possessing a rare η1-CO2-κC binding mode.18 Its solid-state structure clearly displays CO2 coordination to a nickel(II) center via a Ni−C bond with a distance of 1.911(2) Å (see Table 1). Its square-planar structure can be seen by the τ4 value of 0.12, which is similar to those in 4−6 (avg τ4 = 0.1). The O−C−O angle of 128.4(2)° in 7 is slightly larger than values for other carboxylate species, but the central carbon of the CO2 moiety is sp2-hybridized. The two C−O bond distances are nearly identical, suggesting that a pair of electrons is fully delocalized over the two C−O bonds, which are found to be slightly shorter than in the analogous carboxylate complexes; this is due to the absence of a Lewis acid such as H, Na, or Ni. In fact, since nature utilizes Fe as a Lewis acid to assist a Ni center in CO2 conversion, we have also added an iron source to the (PNP)Ni−CO2 species. From the reaction of 5 with (PNP)FeCl, the heterobimetallic nickel−iron carboxylate species (PNP)Ni−μ-CO2−Fe(PNP) (8) was

4. REACTIVITIES OF Ni−CO2 AND Ni−CO SPECIES Protonation of Ni−CO2 Complexes To Generate Ni−CO Species

Protonation of nickel(II) carboxylate species Ni−COO−M (M = Na, Ni, Fe) produces (PNP)Ni(COOH-κC) (4); further protonation cleaves a C−O bond to generate the nickel(II) monocarbonyl species {(PNP)NiCO}{BF4} (11), as shown in Figure 8.14 In fact, the C−O bond can be regenerated from the reaction of 11 with hydroxide.14 Analogous alkoxycarbonyl

Figure 7. Synthesis of metalloradical species 9 and its reactivity toward CO2 to produce 10. D


Article


Figure 8. Stepwise protonation of Ni(COOM-κC) to prepare 11.

species (PNP)NiCOOR (R = Me, tBu) were also obtained by the treatment of 11 with NaOMe or NaOtBu. Such reversible transformations represent a key interconversion of CO2 to CO occurring at a square-planar nickel(II) center. While the present chemistry was studied with heavy metals such as Pt and Ir,24 the corresponding nickel chemistry was not known because of the difficulty of generating and isolating a desired nickel complex possessing a Ni(COOH-κC) moiety.14 Reduction and Reactivity of Ni−CO species

The last step in the conversion of CO2 to CO requires the elimination of CO from the metal center. The cationic species {(PNP)Ni(CO)}{BF4} (11) appears to have a weak interaction between nickel and CO, as supported by a high CO vibrational frequency (νCO = 2071 cm−1) and the short C−O bond distance of 1.133(2) Å.14 While there is a weak interaction between Ni and C, direct decarbonylation in 11 does not occur when vacuum is applied. The addition of an exogenous ligand such as alkoxide (OR − , R = H, Me, t Bu) produces (PNP)NiCOOR instead of forming (PNP)NiOR or eliminating CO.14 This result indicates that the Ni(II)−CO moiety is quite electrophilic and that CO is difficult to eliminate by addition of a nucleophile. Caulton and co-workers reported that the reversible binding of CO occurs at a monovalent nickel ion supported by a PNP′ ligand (PNP′ = (tBu2PCH2SiMe2)2N).25 Because of the antibonding character of the LUMO of the square-planar nickel(II) ion, the reduction of a Ni(II)−CO species can be a good strategy to aid CO release; an added electron occupying the dx2−y2 orbital could weaken the Ni−C bond. Thus, the corresponding mono- and zero-valent nickel monocarbonyl complexes (PNP)Ni(CO) (12) and {Na}{(PNP)NiCO} (13) were prepared by stepwise reduction of 11, as depicted in Figure 9.15 To eliminate the interaction of sodium with the anionic Ni−CO moiety of 13, 2 equiv of 12crown-4 was used, resulting in the formation of {Na(12-C4)2}{(PNP)NiCO} (13′). Furthermore, X-ray diffraction (XRD) studies of 11−13′ revealed that the different numbers of d electrons significantly affect their solid-state structures (see Table 2). According to the CO vibrational frequencies (2071 cm−1 for 11 to 1927 cm−1 for 12 to 1819 cm−1 for 13), the reduction effectively influences the back-donation from Ni to a CO π* orbital. Accordingly, the C−O bond lengths change from 1.133(2) to 1.149(2) and finally to 1.174(4) Å, respectively. A change from square-planar for Ni(II) to pyramidal for Ni(I) and further to pseudotetrahedral for Ni(0) was observed for the metal center. Interestingly, the Ni− C bond distance of 1.776(2) Å in 12 is longer than the value of 1.746(2) Å in Ni(II)−CO; the electron population of the dx2−y2 orbital was found to have antibonding character between Ni and C. Further reduction to 13 clearly reduces the Ni−C distance to 1.719(3) Å. Although the dx2−y2 orbital has antibonding character between Ni and C, its population does not release CO. Instead, further reduction enforces its binding to a low-valent nickel center.

Figure 9. Three oxidation states of (PNP)Ni(CO) species and their reactivities toward iodoalkanes.

The reactivities of the three nickel monocarbonyl species {(PNP)Ni-CO}+/0/− (11−13) toward iodoalkanes were explored to address whether the reactivity can be controlled by the number of valence d electrons (Figure 9).15 While the cationic species 11 does not show any reaction, the d10 complex 13 displays an instantaneous reaction with iodomethane to give (PNP)Ni(CH3) (15-Me) in THF at −35 °C. The same reaction conducted with neutral nickel(I)−CO complex 12, however, revealed the favorable production of the acetyl moiety in 14-Me (37%) over oxidative ligand substitution to give 15Me (13%) along with 16 (50%). This result is the first example to show the C−C bond formation of NiI−CO with an alkyl cation. To investigate the possible steric effects on acylation, various iodoalkanes were used to treat 12 and 13 at −35 °C (Figure 9).15 Interestingly, with iodomethane, 12 reveals a unique reactivity pattern compared with that of 13. To evaluate the open-shell reaction of the nickel(I) monocarbonyl with CH3I, both experimental and theoretical investigations were conducted.16 The initial alkyl radical generation occurs via iodine radical abstraction by Ni(I), producing (PNP)NiI (16) (50%). The corresponding methyl radical was confirmed by trapping experiments using Gomberg’s dimer. The detailed mechanism explored computationally suggests that the nickel− acyl formation may not include the five-coordinate intermediate species (PNP)Ni(CO)(CH3) but supports direct C−C bond formation between a methyl radical and a coordinated CO ligand at a d9 nickel center. Reactivity of Ni(0)−CO toward CO2

The elimination of a bound CO ligand from Ni may occur with the coordination of CO2. Although this key step may in fact be a stepwise reaction, the direct carboxylation of a nickel(0) monocarbonyl species is suggested.17 The initial attempt to perform the carboxylation reaction with the nickel(0)−CO species was unsuccessful. Upon addition of CO2 to {Na}{(PNP)Ni(CO)} (13), multiple products were observed, including (PNP)Ni(CO) (12), {(PNCOONaP)Ni(CO)2}4 (17), and {(PNP)Ni}2(μ-CO3-κ2O,O) (18), as shown in Figure 10. With the desired oxidative addition of CO2, only an 8.3% yield of the nickel(II) carboxylate species {(PNP)Ni}2(μ-CO2κ2C,O) (6) was detected. Since the Td-symmetric tetrameric E


Article

Accounts of Chemical Research Table 2. Nickel Monocarbonyl Species with PNP and +

{(PNP)Ni(CO)} (11) (PNP)Ni(CO) (12) {(PNP)Ni(CO)}− (13′) {(acriPNP)Ni(CO)}+ (20) (acriPNP)Ni(CO) (19) {(acriPNP)Ni(CO)}− (21)

acri

PNP Ligands

νCO2 (cm−1)

Ni−C (Å)

C−O (Å)

N−Ni−C (deg)

2071 1927 1819 2100 1931 1828

1.746(2) 1.776(2) 1.719(3), 1.713(4) 1.764(6) 1.765(1) 1.77(1)

1.133(2) 1.149(2) 1.174(4), 1.173(4) 1.131(7) 1.149(2) 1.18(1)

172.36(6) 153.25(8) 128.7(1), 128.7(1) 176.2(2) 145.69(5) 129.1(4)

Figure 10. CO2 reaction of 13.

Figure 11. Three oxidation states of (acriPNP)Ni(CO) species (19−21) and the carboxylation of 21 to generate 22.

nickel cluster 17 possessing a carbamate moiety is the major product of the reaction, CO2 addition to the amide moiety of the PNP ligand mainly occurs. With this result, we concluded that the flexibility of the ligand backbone allows the central nitrogen atom to be more nucleophilic, leading to the failure of the selective addition of CO2 to a Ni(0) center. In order to reduce the nucleophilicity of the amide group of the (PNP)Ni moiety, we recently designed and prepared acri PNP ligand possessing a structurally rigidified backbone (see Figures 2 and 7).19 Our strategy was to control the corresponding amide nucleophilicity by regulating the rigidity of the PNP ligand. In fact, since we modified the PNP ligand only by incorporating an acridane moiety in its backbone, we could continue to utilize the same reactions established with the (PNP)Ni scaffold. The structural change from square-planar to tetrahedral geometry during the redox involved reactions could be minimized, thus preventing undesired side reactions such as carbamate formation. All of the relevant reactions established with the (PNP)Ni scaffold were again performed but now with the acriPNP system.19,20 Upon addition of CO to the solution of (acriPNP)Ni (9), the corresponding nickel(I) monocarbonyl species (acriPNP)Ni(CO) (19) was generated (Figure 11). Its cyclic voltammogram (CV) revealed two reversible waves at −1.20 and −1.87 V vs the Fc/Fc+ couple, almost identical to those of (PNP)Ni(CO) (12), indicating that our strategy was on track. By chemical oxidation and reduction of 19, the cationic nickel−CO species {(acriPNP)Ni(CO)}{BF4} (20) and the anionic species {Na}{(acriPNP)Ni(CO)} (21), respectively, were prepared (see Figures 11 and 12).20 Three different oxidation states of the nickel monocarbonyl complexes supported by the acriPNP ligand were well-established via various spectroscopic techniques and X-ray crystallography. All of the physical parameters showed good agreement with those of the (PNP)Ni−CO series (see Table 2).

Figure 12. X-ray crystal structures of (a) 21 and (b) 22. Thermal ellipsoids are set at 50% probability, and the countercations, hydrogen atoms, and solvent molecules have been omitted for clarity. (c) Diagram of the CODH active site with CO and CO2 channels.

Upon addition of CO2 to 21, the nickel(II) carboxylate species (acriPNP)Ni−μ-CO2−Na (22) possessing a Ni(η1-CO2κC) moiety was formed (>80% yield; see Figure 12).20 This reaction occurs with the concomitant expulsion of CO, which was detected by GC analysis. This is a remarkable result, different from the result of the CO2 reaction of 13. While lowvalent metal complexes readily react with CO2 to give metallacarboxylates, metal carbonyl complexes, on the other hand, are relatively unreactive toward CO2 because of the stable complexation with the π-acidic CO ligand. In fact, most of the low-valent transition metal complexes are oftentimes poisoned by carbon monoxide.26 To our knowledge, the CO2 reaction of the nickel(0)−CO species 21 to give 22 is a unique example of the metallacarboxylate generation directly from a nickel−CO species. The acriPNP ligand appears to successfully accommodate a zero-valent nickel monocarbonyl moiety, allowing for a direct nickel-center-mediated two-electron reduction of CO2 to occur. F


Article


5. ESTABLISHING A SYNTHETIC CYCLE FOR CO2 REDUCTION AT (ACRIPNP)Ni While both Ni(II) and Ni(I) monocarbonyl species are unreactive toward CO2, the two-electron-reduced Ni−CO species {Na}{(acriPNP)Ni(CO)} (21) successfully shows an instantaneous reaction with CO2 to give the desired Ni(II)− carboxylate product (acriPNP)Ni−μ-CO2−Na (22).20 By utilization of the rigidified (acriPNP)Ni scaffold, upon reduction of a Ni−CO complex corresponding structural changes occur, recognized by the careful structural analysis of {(acriPNP)Ni(CO)}+/0/−. There is an interesting relocation of a coordinated CO from the plane of the (acriPNP)Ni moiety to the axial position of the nickel center, as depicted in Figure 13. While d8 Figure 14. Synthetic cycle of CO2 to CO based on the (acriPNP)Ni scaffold.

liberation of CO. (2) CO is released upon CO2 binding at the nickel(0) center. In fact, in CODH, the coordinated CO ligand might stay at the nickel center in a nickel(I) or a nickel(0) state and can be eliminated by the addition of CO2. Because of the positions of the CO and CO2 channels, as depicted in Figure 12c, the reduction of (acriPNP)Ni(CO) followed by CO2 addition may be a conceivable scenario for the CO liberation process in CODH.27 The CO channel is located at the apical site, while that for CO2 is equatorial. As discussed above, upon reduction of the nickel center, the location of a CO ligand at the four-coordinate nickel center changes from an equatorial to an apical site. Therefore, the reduction of the nickel center systematically moves the terminal CO ligand toward its CO channel, resulting in the opening of a CO2 binding site close to a CO2 channel.

Figure 13. Axial positioning of the terminal CO ligand at the (acriPNP) Ni center in Ni(II)−CO (11, blue), Ni(I)−CO (12, green), and Ni(0)−CO (13, red) based on their X-ray structures (top) and their space-filling models as top views (middle) and side views (bottom).

6. SUMMARY To understand organometallic principles underlying efficient CO2 activation and its selective conversion to CO in the CODH active site, various nickel complexes have been studied in the context of C1 chemistry. Our group has systematically studied the coordination chemistry of CO2 at a single nickel center supported by PPMeP, PNP, and acriPNP ligands. In particular, coordination of CO2 at a nickel center in different geometries was exploited. To attain reactivity for a nickel(0) center in a tetracoordinate environment, a neutral PPMeP ligand was employed. In such a nickel center, CO2 coordinates to give (PPMeP)Ni(η2-CO2) (2) possessing Ni−C and Ni−O bonds. In the case of a square-planar nickel center supported by a PNP ligand, a series of nickel−carboxylate species having a Ni−μCO2-κC,O−M moiety (M = H, Na, Ni, Fe) were generated. In fact, the successful isolation of {Na(12-C-4)2}{(PNP)NiCO2} (7) possessing a Ni(η1-CO2-κC) moiety demonstrates how the square-planar nickel(II) interacts with a dianionic CO22− ligand. Depending on nickel’s geometry (tetrahedral or square-planar), two electrons are located in the Ni d orbital or the π* orbital of CO2. Protonation of the nickel carboxylate species results in C−O bond cleavage to produce {(PNP)NiIICO}{BF4} (11), demonstrating the importance of a metal−carbon bond in the CO2 conversion to CO. The reactivities of the three nickel− monocarbonyl species with formally Ni(II), Ni(I), and Ni(0) (11−13) were also explored. In particular, an unusual C−C bond formation occurs from the reaction of the open-shell nickel(I) species 12 with iodoalkanes. Such alkylation of a nickel carbonyl species can be controlled via variation of nickel’s d electrons. The reactivity of the nickel carbonyl

Ni(II) shows a N−Ni−C angle of 176.2(2)°, this angle decreases to 145.69(5)° and 129.1(4)° as the electron count changes to d9 and d10, respectively. Instead of resulting in a twisted structure often found with the (PNP)Ni scaffold, the (acriPNP)Ni plane is slightly bent upon reduction, allowing for two important steric changes. First, the side isopropyl groups prevent access to the bottom axial site of the complex. Second, the methyl group of the acridane moiety of 21 sterically protects the central amide moiety. As a result, the reduction of 21 sterically protects undesired sites of the complex and selectively opens the binding site of CO2 in its equatorial position. As of yet, we have been unsuccessful in making a catalytic cycle with continuous turnover simply because of the incompatibility of the low-valent nickel species with acid in the presence of continuous electron delivery. Although several challenges remain to operate catalysis, the selective carboxylation in 21 enables us to complete a closed synthetic cycle for CO2 reduction to CO at a single nickel center. The C−O bond of a carboxylate species can be cleaved by addition of 2 equiv of HBF4·Et2O.20 The intermediate nickel hydroxycarbonyl species (acriPNP)NiCOOH (23) can be also isolated. A sequential twoelectron reduction of Ni(II)−CO to Ni(0)−CO completes its synthetic cycle, as depicted in Figure 14. This CO2 reduction cycle based on our nickel study has the following unique features: (1) Two-electron reduction occurs prior to the G


Article


Alternative to Photo- and Electrochemical CO2 Reduction. Chem. Rev. 2015, 115, 12936−12973. (5) Klaus, S.; Lehenmeier, M. W.; Anderson, C. E.; Rieger, B. Recent Advances in CO2/Epoxide CopolymerizationNew Strategies and Cooperative Mechanisms. Coord. Chem. Rev. 2011, 255, 1460−1479. (6) (a) Benson, E. E.; Kubiak, C. P.; Sathrum, A. J.; Smieja, J. M. Electrocatalytic and Homogeneous Approaches to Conversion of CO2 to Liquid Fuels. Chem. Soc. Rev. 2009, 38, 89−99. (b) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. A Local Proton Source Enhances CO2 Electroreduction to CO by a Molecular Fe Catalyst. Science 2012, 338, 90−94. (7) (a) Goeppert, A.; Czaun, M.; Jones, J.-P.; Prakash, G. K. S.; Olah, G. A. Recycling of Carbon Dioxide to Methanol and Derived ProductsClosing the Loop. Chem. Soc. Rev. 2014, 43, 7995−8048. (b) Rezayee, N. M.; Huff, C. A.; Sanford, M. S. Tandem Amine and Ruthenium-Catalyzed Hydrogenation of CO2 to Methanol. J. Am. Chem. Soc. 2015, 137, 1028−1031. (8) (a) Rakowski Dubois, M.; Dubois, D. L. Development of Molecular Electrocatalysts for CO2 Reduction and H2 Production/ Oxidation. Acc. Chem. Res. 2009, 42, 1974−1982. (b) Song, J.; Klein, E. L.; Neese, F.; Ye, S. The Mechanism of Homogeneous CO2 Reduction by Ni(cyclam): Product Selectivity, Concerted Proton− Electron Transfer and C−O Bond Cleavage. Inorg. Chem. 2014, 53, 7500−7507. (9) Can, M.; Armstrong, F. A.; Ragsdale, S. W. Structure, Function, and Mechanism of the Nickel Metalloenzymes, CO Dehydrogenase, and Acetyl-CoA Synthase. Chem. Rev. 2014, 114, 4149−4174. (10) (a) Fesseler, J.; Jeoung, J.-H.; Dobbek, H. How the [NiFe4S4] Cluster of CO Dehydrogenase Activates CO2 and NCO−. Angew. Chem., Int. Ed. 2015, 54, 8560−8564. (b) Jeoung, J.-H; Dobbek, H. Carbon Dioxide Activation at the Ni,Fe-Cluster of Anaerobic Carbon Monoxide Dehydrogenase. Science 2007, 318, 1461−1464. (11) (a) Amara, P.; Mouesca, J.-M.; Volbeda, A.; Fontecilla-Camps, J. C. Carbon Monoxide Dehydrogenase Reaction Mechanism: A Likely Case of Abnormal CO2 Insertion to a Ni−H− Bond. Inorg. Chem. 2011, 50, 1868−1878. (b) Chen, J.; Huang, S.; Seravalli, J.; Gutzman, H., Jr.; Swartz, D. J.; Ragsdale, S. W.; Bagley, K. A. Infrared Studies of Carbon Monoxide Binding to Carbon Monoxide Dehydrogenase/ Acetyl-CoA Synthase from Moorella thermoacetica. Biochemistry 2003, 42, 14822−14830. (c) Seravalli, J.; Kumar, M.; Lu, W.-P.; Ragsdale, S. W. Mechanism of Carbon Monoxide Oxidation by the Carbon Monoxide Dehydrogenase/Acetyl-CoA Synthase from Clostridium thermoaceticum: Kinetic Characterization of the Intermediates. Biochemistry 1997, 36, 11241−11251. (12) Kim, Y.-E.; Kim, J.; Lee, Y. Formation of a Nickel Carbon Dioxide Adduct and its Transformation Mediated by a Lewis Acid. Chem. Commun. 2014, 50, 11458−11461. (13) Kim, Y.-E.; Oh, S.; Kim, S.; Kim, O.; Kim, J.; Han, S. W.; Lee, Y. Phosphinite-Ni(0) Mediated Formation of a Phosphide−Ni(II)− OCOOMe Species via Uncommon Metal−Ligand Cooperation. J. Am. Chem. Soc. 2015, 137, 4280−4283. (14) Yoo, C.; Kim, J.; Lee, Y. Synthesis and Reactivity of Nickel(II) Hydroxycarbonyl Species, NiCOOH-κC. Organometallics 2013, 32, 7195−7203. (15) Yoo, C.; Oh, S.; Kim, J.; Lee, Y. Transmethylation of a Fourcoordinate Nickel(I) Monocarbonyl Species with Methyl Iodide. Chem. Sci. 2014, 5, 3853−3858. (16) Yoo, C.; Ajitha, M. J.; Jung, Y.; Lee, Y. Mechanistic Study on C− C Bond Formation of a Nickel(I) Monocarbonyl Species with Alkyl Iodides: Experimental and Computational Investigations. Organometallics 2015, 34, 4305−4311. (17) Yoo, C.; Lee, Y. Formation of a Tetranickel Octacarbonyl Cluster from the CO2 Reaction of a Zero-valent Nickel Monocarbonyl Species. Inorg. Chem. Front. 2016, 3, 849−855. (18) Yoo, C.; Lee, Y. Carbon Dioxide Binding at a Ni/Fe Center: Synthesis and Characterization of Ni(η1-CO2-κC) and Ni-μ-CO2κC:κ2O,O′-Fe. Chem. Sci. 2017, 8, 600−605. (19) Yoo, C.; Lee, Y. A T-Shaped Nickel(I) Metalloradical Species. Angew. Chem., Int. Ed. 2017, 56, 9502−9506.

complexes toward CO2 is clearly differentiated with their oxidation state. Only a (PNP)Ni(0)−CO species (13) reacts with CO2, but it undergoes multiple reactions. With the structurally rigidified acriPNP ligand, a T-shaped nickel(I) metalloradical species (9) was successfully stabilized. Having a sterically exposed half-filled dx2−y2 orbital, this threecoordinate nickel(I) species reveals unique open-shell reactivity. Importantly, this modification is successful in reproducing all of the reactions related to CO2 conversion established using the (PNP)Ni system. The reduction of {(acriPNP)Ni(CO)}{BF4} (20) also succeeded in generating mono- and zero-valent nickel carbonyl complexes (19 and 21). In fact, the Ni(0)−CO species successfully reveals the selective addition of CO2 to give a nickel(II)−carboxylate species (22) with the expulsion of CO. The closed synthetic cycle for reduction of CO2 to CO was finally accomplished with the (acriPNP)Ni system.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail for Y.L.: [email protected]. ORCID

Yunho Lee: 0000-0002-9113-9491 Funding

This work was supported by the C1 Gas Refinery Program (NRF 2015M3D3A1A01064880). Notes

The authors declare no competing financial interest. Biographies Changho Yoo received his B.S. and Ph.D. degrees from KAIST in South Korea under the guidance of Professor Yunho Lee. In 2017, he joined the Miller group at the University of North Carolina at Chapel Hill as a postdoctoral fellow. Yeong-Eun Kim received her B.S. and Ph.D. degrees from KAIST in South Korea under the guidance of Professor Yunho Lee. She is currently a postdoctoral fellow at KAIST. Yunho Lee received his B.S. degree from Chonbuk National University in South Korea and his M.S. and Ph.D. degrees in Inorganic Chemistry from the Johns Hopkins University in Baltimore under the guidance of Professor Kenneth D. Karlin. After receiving his Ph.D. in 2007, he was a postdoctoral fellow in the laboratory of Professor Jonas C. Peters at MIT and CalTech. In the winter of 2010, he became an Assistant Professor in the Department of Chemistry at KAIST in South Korea, where he is currently an Associate Professor after his promotion in 2015.

■

REFERENCES

(1) Aresta, M.; Dibenedetto, A.; Angelini, A. Catalysis for the Valorization of Exhaust Carbon: from CO2 to Chemicals, Materials, and Fuels. Technological Use of CO2. Chem. Rev. 2014, 114, 1709− 1742. (2) Qiao, J.; Liu, Y.; Hong, F.; Zhang, J. A Review of Catalysts for the Electroreduction of Carbon Dioxide to Produce Low-carbon Fuels. Chem. Soc. Rev. 2014, 43, 631−675. (3) Paparo, A.; Silvia, J. S.; Kefalidis, C. E.; Spaniol, T. P.; Maron, L.; Okuda, J.; Cummins, C. C. A Dimetalloxycarbene Bonding Mode and Reductive Coupling Mechanism for Oxalate Formation from CO2. Angew. Chem., Int. Ed. 2015, 54, 9115−9119. (4) Wang, W.-H.; Himeda, Y.; Muckerman, J. T.; Manbeck, G. F.; Fujita, E. CO2 Hydrogenation to Formate and Methanol as an H


Article

Accounts of Chemical Research (20) Sahoo, D.; Yoo, C.; Lee, Y. Direct CO2 addition to a Ni(0)-CO species allowing the selective generation of a nickel(II) carboxylate with expulsion of CO. J. Am. Chem. Soc. 2018, 140, 2179−2185. (21) Paparo, A.; Okuda, J. Carbon Dioxide Complexes: Bonding Modes and Synthetic Methods. Coord. Chem. Rev. 2017, 334, 136− 149. (22) Aresta, M.; Nobile, C. F.; Albano, V. G.; Forni, E.; Manassero, M. New Nickel-Carbon Dioxide Complex: Synthesis, Properties, and Crystallographic Characterization of (Carbon dioxide)-bis(tricyclohexylphosphine)nickel. J. Chem. Soc., Chem. Commun. 1975, 0, 636−637. (23) Anderson, J. S.; Iluc, V. M.; Hillhouse, G. L. Reactions of CO2 and CS2 with 1,2-Bis(di-tert-butylphosphino)ethane Complexes of Nickel(0) and Nickel(I). Inorg. Chem. 2010, 49, 10203−10207. (24) (a) Chan, P. K.; Leong, W. K. Reaction of Cp*Ir(CO)2 with Activated Perfluoroaromatic Compounds: Formation of Metallocarboxylic Acids via Aromatic Nucleophilic Substitution. Organometallics 2008, 27, 1247−1253. (b) Johansson, R.; Wendt, O. F. Synthesis and Reactivity of (PCP) Palladium Hydroxy Carbonyl and Related Complexes toward CO2 and Phenylacetylene. Organometallics 2007, 26, 2426−2430. (c) Elliott, P. I. P.; Haslam, C. E.; Spey, S. E.; Haynes, A. Formation and Reactivity of Ir(III) Hydroxycarbonyl Complexes. Inorg. Chem. 2006, 45, 6269−6275. (25) Ingleson, M. J.; Fullmer, B. C.; Buschhorn, D. T.; Fan, H.; Pink, M.; Huffman, J. C.; Caulton, K. G. Influence of the d-Electron Count on CO Binding by Three-Coordinate [(tBu2PCH2SiMe2)2N]Fe, -Co, and -Ni. Inorg. Chem. 2008, 47, 407−409. (26) (a) Froehlich, J. D.; Kubiak, C. P. The Homogeneous Reduction of CO2 by [Ni(cyclam)]+: Increased Catalytic Rates with the Addition of a CO Scavenger. J. Am. Chem. Soc. 2015, 137, 3565−3573. (b) Akhade, S. A.; Luo, W.; Nie, X.; Bernstein, N. J.; Asthagiri, A.; Janik, M. J. Poisoning Effect of Adsorbed CO during CO 2 Electroreduction on Late Transition Metals. Phys. Chem. Chem. Phys. 2014, 16, 20429−20435. (27) Wang, P.-H.; Bruschi, M.; De Gioia, L.; Blumberger, J. Uncovering a Dynamically Formed Substrate Access Tunnel in Carbon Monoxide Dehydrogenase/Acetyl-CoA Synthase. J. Am. Chem. Soc. 2013, 135, 9493−9502.

I


Selective Transformation of CO2 to CO at a Single Nickel Center

Recommend Documents