Cobalt N-Heterocyclic Phosphenium Complexes ... - ACS Publications

(a) Evers-McGregor , D. A.; Bezpalko , M. W.; Foxman , B. M.; Thomas , C. M. ...... Kyle G. PearceAndryj M. BorysEwan R. ClarkHelena J. Shepherd. Inor...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IC

Cobalt N‑Heterocyclic Phosphenium Complexes Stabilized by a Chelating Framework: Synthesis and Redox Properties Mark W. Bezpalko, Andrew M. Poitras, Bruce M. Foxman, and Christine M. Thomas* Department of Chemistry, Brandeis University, 415 South Street, Waltham, Massachusetts 02454, United States S Supporting Information *

ABSTRACT: Two cobalt complexes containing coordinated Nheterocyclic phosphenium (NHP+) ligands are synthesized using a bidentate NHP+/phosphine chelating ligand, [PP]+. Treatment of Na[Co(CO)4] with the chlorophosphine precursor [PP]Cl (1) affords [PP]Co(CO)2 (2), which features a planar geometry at the NHP+ phosphorus center and a short Co−P distance [1.9922(4) Å] indicative of a CoP double bond. The more electron-rich complex [PP]Co(PMe3)2 (3), which is synthesized in a one-pot reduction procedure with 1, CoCl2, PMe3, and KC8, has an even shorter Co−P bond [1.9455(6) Å] owing to stronger metal-to-phosphorus back-donation. The redox properties of 2 and 3 were explored using cyclic voltammetry, and oxidation of 3 was achieved to afford [[PP]Co(PMe3)2]+ (4). The electron paramagnetic resonance spectrum of complex 4 features hyperfine coupling to both 59Co and 31P, suggesting strong delocalization of the unpaired electron density in this complex. Density functional theory calculations are used to further explore the bonding and redox behavior of complexes 2−4, shedding light on the potential for redox noninnocent behavior of NHP+ ligands.



INTRODUCTION As cationic analogues of N-heterocyclic carbenes (NHCs), Nheterocyclic phosphenium cations (NHP+s) have very different donor/acceptor properties and the potential to facilitate different reactivities upon coordination to transition metals.1 Unlike NHCs, NHP+s are weak σ donors and strong π acceptors and have the unique ability to adopt either planar or pyramidal binding modes. Their electrophilic nature and susceptibility to nucleophilic attack has somewhat limited the study of the coordination chemistry of NHP+s to electron-rich metals in the absence of nucleophilic ligands. While a majority of the reported coordination compounds of NHP+ ligands reported to date feature second- and third-row transition metals,2 NHP+s have been coordinated to iron,3 cobalt,4 and nickel5 carbonyl fragments and our group recently reported a family of nickel coordination compounds (Chart 1).6 Our group has recently reported both bidentate and tridentate chelating ligands containing NHP+ cations and one or two appended phosphine side arms to enforce metal coordination.6b,8 Many of the coordination compounds of the NHP+/diphosphine [PPP]+ ligand feature pyramidal geometries about the PNHP centers (e.g., G and H in Chart 1).4a,6a,7 Through structural and computational studies, we have determined that this pyramidal geometry is best described as arising from formal electron transfer from electron-rich Mn fragments to the NHP+ ligand, resulting in a formal assignment of these compounds as NHP− phosphido complexes with metals in the Mn+2 formal oxidation state. While there are other valid interpretations of the pyramidal geometry at the PNHP center, it is intriguing to consider NHP± ligands as redoxnoninnocent analogues of nitrosyls.2f,7 Interestingly, we found © XXXX American Chemical Society

that the preference for a pyramidal NHP geometry appears to be unique to the rigid tridentate [PPP]+ framework because otherwise analogous complexes with either monodentate NHP+ ligands or our bidentate NHP+/phosphine ligand [PP]+ consistently feature planar geometries about the PNHP atom. For example, in the series of cationic [(NHP)M(PR3)3]+ complexes (M = Pd, Ni; e.g., D, H, and I in Chart 1), the NHP ligands tethered to either 0 or 1 phosphine donor have a planar geometry about the PNHP center, while the NHP ligand that is part of the rigid [PPP]+ pincer ligand is pyramidal.2f,6 Herein, the coordination chemistry of NHP+ ligands is further explored using the bidentate [PP]+ ligand specifically targeting cobalt complexes to expand the scope of transitionmetal fragments ligated by NHP+ ligands. Two [PP]Co(L)2 complexes are isolated and their redox properties explored through experimental and computational studies to further explore the redox noninnocent behavior of NHP+ ligands.



RESULTS AND DISCUSSION Synthesis and Characterization of (PP)Co(CO)2 and (PP)Co(PMe3)2. Similar to previous reports, we found that a salt metathesis route involving an anionic metal precursor and an N-heterocyclic chlorophosphine precursor was an effective metalation strategy for our bidentate NHP+/phosphine ligand.2k,l,3b−d,4 Treatment of chlorophosphine 16b with Na[Co(CO)4] in tetrahydrofuran (THF) generates a red complex, [PP]Co(CO)2 (2), with concomitant evolution of CO(g) and precipitation of NaCl (Scheme 1). The 31P NMR Received: October 4, 2016

A

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

stretches at 1967 and 1915 cm−1 for the carbonyl ligands that remain bound to cobalt. As expected from the spectroscopic features, the solid-state structure of 2 reveals a neutral cobalt complex with a coordinated bidentate NHP+/phosphine ligand and two carbonyl ligands (Figure 1 and Table 1). The Co center

Chart 1. Representative Transition-Metal NHP+ Complexes Reported by Gudat et al.,3b,4b Jones et al.,2f Paine et al.,4c and Thomas et al.4a,6b,7

Figure 1. Displacement ellipsoid (50%) representation of 2. For clarity, all H atoms and a THF solvate molecule have been omitted. Relevant interatomic distances (Å): Co1−P1, 1.9922(4); Co1−P2, 2.1615(5); Co1−C30, 1.7622(19); Co1−C31, 1.7693(18); P1−N1, 1.6449(13); P1−N2, 1.6852(13).

adopts a distorted tetrahedral coordination geometry, and the NHP+ phosphenium ligand adopts a nearly planar binding mode, with the sum of the three angles about the P atom equal to 356.0°. The PNHP−Co bond distance in 2 is quite short [1.9922(4) Å] and indicative of Co−P multiple bonding. In fact, the PNHP−Co bond distance in 2 is shorter than that observed in the similar previously reported (uRNHP)Co(CO)3 complexes (C, shown in Chart 1, where “u” denotes an unsaturated heterocycle and the “R” in RNHP represents the substituents on nitrogen; R = Mes [2.0018(4) Å] and tBu [2.0450(5) Å].4b The shorter PNHP−Co bond distance in 2 can be attributed to a more electron-rich Co center upon replacement of a carbonyl ligand with a phosphine ligand, allowing for stronger Co → PNHP π-back-bonding to occur. A weak intermolecular contact is observed between the THF solvate molecule and the PNHP center with a P−O distance of 3.220(2) Å, which is less than the sum of the van der Waals radii of the two atoms. This interaction, depicted in Figure S10,

Scheme 1

spectrum of 2 features a broad downfield signal at 241.8 ppm for the NHP ligand bound to cobalt and an upfield signal at 41.1 ppm corresponding to the cobalt-bound phosphine side arm (Table 1). The IR spectrum of 2 reveals two ν(CO)

Table 1. Experimental M−PNHP Distances, 31P NMR Chemical Shifts, and ν(CO) Stretching Frequencies, Where Applicable, of Complexes 1−3 and a Selection of Comparable Metal(NHP) Complexes compound

M−PNHP distance (Å)

PNHP 31P NMR chemical shift (δ, ppm)

[PP]Cl (1) [PP]Co(CO)2 (2) [PP]Co(PMe3)2 (3) [(PP)Ni(PMe3)2]+ (I) [PPP]Co(CO)2 (G) (uMesNHP)Co(CO)3 (C) (utBuNHP)Co(CO)3 (C) [(uDippNHP)Fe(CO)3]− (B)

N/A 1.9922(4) 1.9455(6) 1.9840(4) 2.2386(6) 2.0018(4) 2.0450(5) 1.989(1)

155.3 241.8 207.0 236.0 286.4 233.0 234.7 197.9

ν(CO) (cm−1) N/A 1967, N/A N/A 1981, 2015, 2015, 1894,

1915a

1926a 1945c 1948b 1813, 1798

reference 6b this work this work 6b 4a 4b 4b 3b

a c

The IR spectrum was measured in a THF solution using a KBr solution cell. bThe IR spectrum was measured as a Nujol mull using a KBr plate. The IR spectrum was measured in a MeCN solution using a NaCl solution cell. B

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry is a result of the electrophilic nature of the NHP+ in complex 2 and provides support for the assignment of the threecoordinate P center as an NHP+ phosphenium rather than as an NHP− phosphido ligand. The planar NHP binding mode observed in complex 2 is consistent with the nickel complexes ligated by the same bidentate NHP+/phosphine [PP]+ ligand.6b However, the planarity about the PNHP atom in 2 is in stark contrast to the pyramidal NHP binding mode reported for a similar cobalt dicarbonyl complex, [PPP]Co(CO)2 (G, shown in Chart 1), supported by a tridentate NHP/diphosphine ligand.4a In the case of G, the pyramidal geometry about the PNHP atom was attributed to formal reduction of the NHP+ fragment. As was concluded for a set of analogous Ni(NHP) complexes ligated by the bidentate and tridentate [PP]+ and [PPP]+ ligands,6 the planar NHP binding mode in 2 is a result of the less rigid geometry of the bidentate [PP]+ ligand and the ability of the Nmesityl substituent to freely rotate to an orientation perpendicular to the plane of the heterocycle. This orientation prevents delocalization of one N lone pair throughout the mesityl ring and allows more N−P π delocalization to occur in 2, stabilizing the phosphenium cation and preventing formal reduction to a pyramidal NHP− phosphido ligand as occurs in G. This hypothesis is supported by the differences in the two N−P distances in 2. The N−P distance associated with the mesityl-substituted N atom is 1.6449(13) Å, while that associated with the chelating side arm is 1.6852(13) Å. For comparison, the N−P distances in G are 1.73−1.74 Å,4a supporting that there is more N−P π bonding when at least one N-aryl ring is not coplanar with the heterocycle. While it seems logical to extend the observed trends in N−P π bonding to explain the differences in metal−P bonding between 2 and G, we note that the Co−P distances in 2 and C are very similar, suggesting that there is not a continuous variation in Co−P bonding upon a decrease in the number of coplanar N-aryl rings from two to one to zero. The symmetric and asymmetric carbonyl stretching frequencies at 1967 and 1915 cm−1 for the Co(CO)2 fragment in 2 are lower in energy than those of G [ν(CO) = 1981 and 1926 cm−1],4a indicating that there is stronger Co → CO πback-bonding in complex 2 than in G. A comparison between the two complexes may not be entirely meaningful given the different coordination numbers and geometries of the Co centers in 2 and G, but given that G has an additional phosphine donor compared to 2, the lower-frequency CO stretches for 2 are difficult to rationalize unless the differences in bonding between the metal and the NHP fragment are taken into account. If the pyramidal geometry about the PNHP atom in G is indicative of a formal NHP−/CoI configuration and the planar NHP geometry in 2 leads to an NHP+/Co−I assignment, then the difference in the Co oxidation state accounts for the increased Co → CO back-bonding in complex 2. Next we aimed to isolate the first structurally characterized carbonyl-free Co(NHP) complexes with the hope of generating complexes with labile ancillary ligands and, thus, more potential for reactivity studies. We employed a one-pot method involving coordination of the [PP]Cl ligand to a CoII source followed by reduction of the resulting complex in situ to Co−I. The addition of 3 equiv of KC8 to a 1:1 mixture of the chlorophosphine precursor 1 and CoCl2 in the presence of 2 equiv of trimethylphosphine cleanly generates the diamagnetic complex [PP]Co(PMe3)2 (3; Scheme 2). The 31P NMR spectrum of 3 features three signals at 207.0, 41.6, and 3.8 ppm in a 1:1:2

Scheme 2

integral ratio, representative of the PNHP atom, the cobaltbound phosphine side arm, and two PMe3 ligands, respectively. Similar to complex 2, the solid-state structure of 3 reveals a distorted tetrahedral coordination geometry about cobalt and the NHP+ ligand adopts a planar binding mode, with the angles about the P atom summing to 359° (Figure 2). However, the

Figure 2. Displacement ellipsoid (50%) representation of 3. For clarity, all H atoms have been omitted. Relevant interatomic distances (Å): Co1−P1, 1.9455(6); Co1−P2, 2.1116(6); Co1−P3, 2.1469(6); Co1−P4, 2.1454(7); P1−N1, 1.6816(18); P1−N2, 1.7274(18).

PNHP−Co distance in 3 [1.9455(6) Å] is 0.05 Å shorter than that in 2, owing to the strong σ donation from the PMe3 ligands to the Co center that allow it to π-back-bond more strongly with the NHP+ ligand. Stronger Co → P π-back-bonding is also apparent from the 0.04 Å longer P−N bond distances in 3 [1.6816(18) and 1.7274(18) Å] because the acceptor orbital has P−N π* character. The PNHP−Co bond distance in 3 is the shortest metal−P distance reported for a metal NHP complex, with the closest comparison being the PNHP−M bond distances of 1.989(1) and 1.9840(4) Å in [(uDippNHP)Fe(CO)3][PPh4] (B; Dipp = 2,6-diisopropylphenyl) and [(PP)Ni(PMe3)2][BPh4] (I), respectively.3b,6b Furthermore, the M−P distance in 3 is among the shortest metal−P distances reported in any compound to date (based on a 2016 search of the Cambridge Structural Database)9 and is shorter than the M−P distances in all of the reported late-first-row metal phosphinidenes, M = PR, that have bona fide MP double bonds {e.g., (dtbpe)Ni P(dmp) [2.0772(9) Å], (Cp)(PPh3)CoPMes* [2.1102(8) Å], where dtbpe = 1,2-bis(di-tert-butylphosphino)ethane, dmp = 2,6-dimesitylphenyl, and Mes* = 2,4,6-tri-tert-butylphenyl}.10 Thus, the Co−P distance in 3 is indicative of very strong Co−P π-back-bonding, resulting in a CoP double bond. To probe our hypothesis that the PMe3 ligands in complex 3 would be more labile that the CO ligands in complex 2, the reactivity of 3 with extrinsic CO was explored to see if the PMe3 ligands in 3 could be displaced to generate complex 2. The addition of 1 atm of CO to 3 resulted in an immediate C

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry reaction to afford a new product with three 31P NMR resonances in a 1:1:1 integral ratio at 243, 33, and 2 ppm (Figure S9). An additional signal of equal integration at −61 ppm indicates that 1 equiv of PMe3 was released. In concert with the diminished integration of the cobalt-bound PMe3 signal, this suggests the substitution of just one of the two PMe3 ligands with CO to form [PP]Co(CO)(PMe3). The ligand substitution process did not proceed further to generate complex 2 even with further reaction time (48 h) or elevated temperature (60 °C), but the rapid substitution of one of the two ancillary ligands in complex 3 does support the lability of at least one of the PMe3 ligands in this complex. Redox Behavior of 2 and 3. To probe the redox activity of complexes 2 and 3, cyclic voltammetry measurements were performed. The cyclic voltammograms (CVs) of 2 and 3 in THF exhibit irreversible reduction waves at Epc = −2.59 and −2.53 V, respectively (vs Fc/Fc+, where Fc = Cp2Fe; Figure 3).

The solution magnetic moment of 4 measured using Evans’ method is 1.7 μB, consistent with an S = 1/2 spin state. Electron paramagnetic resonance (EPR) spectroscopy was therefore a useful tool to evaluate localization of the unpaired electron in 4. The X-band EPR spectrum of complex 4 was collected in frozen THF at 3 K. A relatively broad signal with discernible hyperfine features was observed and simulated as a rhombic signal with a degree of g anisotropy typical of a metal-based radical (gx = 2.227, gy = 2.113, and gz = 2.028; Figure 4). The

Figure 3. CVs of 2 and 3 (2 mM in 0.2 M [nBu4N][ClO4] in THF; scan rate = 100 mV s−1).

Figure 4. Experimental (black) and simulated (red dotted) X-band (9.38 GHz) EPR spectra of complex 4 in frozen THF at 3 K. The simulation was produced using the following parameters: gx = 2.227, gy = 2.113, gz = 2.028; ACox = 110.1 MHz, ACoy = 115.4 MHz, ACoz = 162.9 MHz, APx = 111.7 MHz, APy = 245.4, APz = 175.0 MHz.

On the basis of their similarity, these features are assigned as ligand-based reductions, and this hypothesis is supported by computational studies (vide infra). In contrast, the CVs of 2 and 3 reveal quite different oxidative features (Figure 3). The CV of 2 has an irreversible oxidation at Epa = −0.16 V (vs Fc/ Fc+), while the CV of 3 contains a reversible oxidation at E1/2 = −1.26 V (vs Fc/Fc+). The ease and reversibility of the oxidation of compound 3 can be attributed to the more electron-rich Co center in this molecule. Inspired by the reversible oxidation event at mild potential observed in the CV of 3, its bulk chemical oxidation with Fc+ was investigated. The addition of a solution of 3 in THF to stoichiometric [Fc][BPh4] generates a new paramagnetic complex, [[PP]Co(PMe3)2][BPh4] (4; Scheme 3). Complex

signal can only be satisfactorily simulated by incorporating significant hyperfine coupling to both the 59Co nucleus (I = 7/2, Ax = 110.1 MHz, Ay = 115.4 MHz, and Az = 162.9 MHz) and one 31P nucleus (I = 1/2, Ax = 111.7 MHz, Ay = 245.4 MHz, and Az = 175.0 MHz), which we hypothesize is the PNHP atom. The hyperfine coupling to both 59Co and 31P suggests that the singly occupied molecular orbital is delocalized over both the Co and PNHP atoms, and computational studies support this hypothesis (vide infra). Computational Studies. To better understand the bonding and electronic structure of complexes 2 and 3, a computational investigation was carried out using density functional theory (DFT). Geometry optimizations and subsequent natural bond order (NBO) analyses were performed on untruncated molecules of 2 and 3 starting from the crystallographic coordinates. The bond distances computed for 2 and 3 are similar to those determined crystallographically (Table S2), lending merit to the choice of the functional (B3LYP) and basis set (LANL2DZ). Consistent with the short Co−P distances that indicate multiple bonding, NBO calculations of 2 and 3 each found two Co−PNHP NBOs (Figure 5). One of the Co−PNHP NBOs in each molecule is comprised of dative donation from the sphybridized P lone pair to the Co center, as indicated by the composition of the NBO (2, 30.1% Co and 69.9% P; 3, 25.8% Co and 74.2% P), as tabulated in Table 2. In addition to σ donation, an additional π-type Co−P NBO contributes to bonding in both complexes and is comprised of almost equal

Scheme 3

4 has a broad and paramagnetically shifted 1H NMR spectrum with 12 well-resolved peaks in the −15 to −21 ppm range along with three distinct resonances in the 6.5−7.5 ppm region attributed to the [BPh4]− counterion. Despite multiple attempts, single crystals of 4 suitable for X-ray diffraction could not be obtained. D

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Table 3. Computed Bond Metrics for Complexes 3−5

a

complex

M−PNHP (Å)

ΣP angles (deg)a

[PP]Co(PMe3)2 (3) [[PP]Co(PMe3)2]+ (4) [[PP]Co(PMe3)2]− (5)

2.00 2.12 2.15

359 359 337

ΣP angles = sum of angles about the PNHP atom.

To further investigate the origin of the elongated metal−P distances in 4 and 5, the pyramidal phosphorus geometry in 5, and the localization of the unpaired electrons in the two S = 1/2 molecules 4 and 5, we investigated the computed electronic structure of these molecules. Mulliken population analysis of reduced compound 5 predicts that the unpaired spin density resides on the Co atom, with little density residing on the PNHP atom [Mulliken spin densities: 0.89 (Co) and 0.13 (PNHP); Figure 6]. However, pyramidalization of the PNHP atom Figure 5. Visual representations of the calculated Co−P bonding interactions in complexes 2 (a and b) and 3 (c and d) using NBO analysis.

contributions from both the Co and P atoms (2, 45.4% Co and 54.6% P; 3, 48.4% Co and 51.6% P). This more covalent interaction involves the P pπ orbital and an orbital on Co with substantial p/d mixing. The combination of σ- and π-bonding interactions in 2 and 3 is consistent with the short Co−PNHP bond distances observed in the solid-state structures. This bonding description, along with the natural electron configurations computed for each of the Co centers, which reveal a total of 10 valence electrons (Table 2), suggests a formal oxidation state assignment of Co−I with an NHP+ phosphenium ligand. However, it is worth noting that the metal−P bonding in these molecules is quite covalent, so oxidation state assignments may not be meaningful. Because the oxidized complex 4 could not be structurally characterized, its geometry and electronic structure were explored computationally. Because the one-electron-reduction events in the CVs of 2 and 3 were tentatively assigned as ligand-based redox events, a calculation was also performed on the hypothetical one-electron-reduction product [[PP]Co(PMe3)2]− (5). The crystallographic coordinates of 3 were used as a starting point for the geometry optimizations of 4 and 5. The optimized geometries of both 4 and 5 feature distorted tetrahedral geometries about the Co centers and elongated Co−PNHP distances (2.12 and 2.15 Å) compared to those calculated for the neutral precursor 3 (2.00 Å; Table 3). The geometry about the PNHP atom in 4 remains trigonal planar (sum of angles about the PNHP atom = 359°). However, the optimized geometry of 5 reveals a trigonal-pyramidal geometry about the PNHP atom, with the angles about the PNHP atom summing to 337°.

Figure 6. Calculated unpaired spin-density surface of complexes 4 and 5, with Mulliken spin densities for the Co and PNHP atoms labeled.

suggests a stereochemically active lone pair and an NHPcentered reduction from NHP+ phosphenium to NHP− phosphido. Consistent with this hypothesis, the N−P distances are predicted to elongate by ∼0.1 Å upon reduction (Table S2). Furthermore, the sum of the computed natural charges on the bidentate ligand is predicted to decrease by 1 upon reduction, while the natural charge on Co increases slightly (Table 4). It Table 4. Computed Natural Charges on the Co Center and [PP] Ligands in Complexes 3−5 complex

Co

Σ[PP]

[PP]Co(PMe3)2 (3) [[PP]Co(PMe3)2]+ (4) [[PP]Co(PMe3)2]− (5)

−1.56 −0.67 −1.33

−1.15 −0.99 −2.11

a Σ[PP] = sum of natural charges on all atoms of the [PP] ligand fragment.

can therefore be concluded that one-electron reduction is, indeed, NHP-based but that the net result is an NHP− phosphido cobalt(0) complex with an unpaired electron

Table 2. Computed Co and PNHP Natural Charges, Co Natural Electron Configurations, and Compositions of the Co−PNHP NBOs for Complexes 2 and 3 Co−PNHP NBO composition Co natural electron configuration [PP]Co(CO)2 (2)

4s0.433d8.684p0.93

[PP]Co(PMe3)2 (3)

4s0.423d8.764p0.92

Co 30.1% 45.4% 25.8% 48.4%

(24% s, 62% p, 14% d) (1% s, 58% p, 41% d) (20% s, 70% p, 10% d) (9% s, 41% p, 50% d) E

P 69.9% 54.6% 74.2% 51.6%

NBO occupancy

(56% s, 44% p) (8% s, 92% p) (53% s, 47% p) (15% s, 85% p)

1.82 1.66 1.91 1.78

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

were purchased from Cambridge Isotope Laboratories, Inc., degassed via repeated freeze−pump−thaw cycles, and dried over 3 Å molecular sieves. The ligand [PP]Cl (1)6b and Na[Co(CO)4]11,12 were synthesized using literature procedures. All other chemicals were purchased from Aldrich, Strem, or Alfa Aesar and used without further purification. NMR spectra were recorded at ambient temperature on a Varian Inova or 400MR 400 MHz instrument. Chemical shifts are reported in δ (ppm). For 1H and 13C NMR spectra, the solvent resonance was used as an internal reference. For 31P NMR spectra, 85% H3PO4 was referenced as an external standard (0 ppm). For the paramagnetic molecule, the 1H NMR data are reported with the chemical shift, followed by the peak width at half-height (in Hz) and a rough estimate of the peak integration in parentheses. IR spectra were recorded on a Varian 640-IR spectrometer controlled by Resolutions Pro software. UV−vis spectra were recorded on a Cary 50 UV−vis spectrophotometer using Cary WinUV software. Elemental analyses were performed at Complete Analysis Laboratory Inc., Parsippany, NJ. Solution magnetic moments (μeff) were measured using Evans’ method.13 [PP]Co(CO)2 (2). Na[Co(CO)4] (85 mg, 0.17 mmol) was dissolved in 2 mL of THF, and to this light-yellow solution was added a solution of 1 (33 mg, 0.17 mmol) in 2 mL of THF. The light-yellow solution immediately started to turn red and effervesce as CO(g) was evolved. The mixture was allowed to stir at room temperature for 12 h. The resulting dark-red mixture was filtered through Celite, and the dark-red filtrate was concentrated to 1 mL and layered with 1 mL of Et2O for 24 h, affording 2 as red crystals suitable for X-ray crystallography. Once collected, the crystals were washed once with 1 mL of Et2O and further dried in vacuo. Crystalline yield: 65 mg, 66%. 1H NMR (400 MHz, THF-d8): δ 7.41−7.33 (overlap m, 5H, P−Ar−H, Ar−H), 7.31−7.28 (overlap m, 6H, P−Ar−H), 7.12 (dd, 1H, J = 7.9 and 4.9 Hz, Ar−H), 6.93 (dd, 1H, J = 7.5 and 7.3 Hz, Ar−H), 6.89 (s, 2H, Mes−H), 6.85 (m, 1H, Ar−H), 3.97 (m, 2H, CH2), 3.49 (m, 2H, CH2), 2.23 (s, 3H, p-CH3), 2.08 (s, 6H, o-CH3). 31P{1H} NMR (161.8 MHz, THF-d8): δ 241.8 (br s, 1P), 41.1 (s, 1P). 13C{1H} NMR (100.5 MHz, THF-d8): δ 213.4 (br, CO), 147.3 (m), 139.6 (d, 1JCP = 9.7 Hz), 139.2 (d, 1JCP = 9.7 Hz), 138.2 (s), 137.4 (d, 3JCP = 2.9 Hz), 134.6 (d, 1 JCP = 10.6 Hz), 133.5 (s), 133.4 (s), 133.1 (d, 3JCP = 1.9 Hz), 130.8 (s), 129.9 (s), 129.6 (s), 128.6 (s), 128.5 (s), 123.2 (d, 1JCP = 7.8 Hz), 123.0 (d, 2JCP = 4.8 Hz), 122.8 (d, 1JCP = 7.7 Hz), 117.6 (m), 49.5 (d, 3 JCP = 4.8 Hz), 48.7 (d, 3JCP = 3.0 Hz), 20.9 (s), 17.9 (s). IR (KBr solution cell, THF): 1967, 1915 cm−1. UV−vis [THF; λ, nm (ε, M−1 cm−1)]: 445 (1.75 × 104). Anal. Calcd for C31H29N2P2O2Co: C, 63.92; H, 5.02; N, 4.81. Found: C, 63.25; H, 5.86; N, 4.21. [PP]Co(PMe3)2 (3). Solid 1 (49 mg, 0.096 mmol) was dissolved in THF (4 mL), added to a suspension of CoCl2 (13 mg, 0.096 mmol) in THF (6 mL), and stirred for 1 h at room temperature. Once CoCl2 was fully dissolved, PMe3 (20 μL, 0.19 mmol) was added, and the mixture was stirred for 2 min. The resulting green solution was then added to a suspension of KC8 (40 mg, 0.29 mmol) in THF (2 mL). The reaction mixture was allowed to stir at room temperature for 3 h. The resulting brown mixture was filtered through Celite, and the volatiles were removed from the filtrate in vacuo, affording analytically pure product as a red-brown solid. Crystals suitable for X-ray crystallography were grown from a concentrated Et2O solution of 3. Yield: 65 mg, 98%. 1H NMR (400 MHz, C6D6): δ 7.73 (dd, 4H, J = 8.3 and 8.3 Hz, P−Ar−H), 7.48 (dd, 1H, J = 7.9 and 7.9 Hz, Ar−H), 7.21 (dd, 1H, J = 7.5 and 7.5 Hz, Ar−H), 7.14 (m, 4H, P−Ar−H), 7.06 (m, 2H, P−Ar−H), 6.85−6.79 (overlap m, 4H, Ar−H, Mes−H), 3.38 (m, 2H, CH2), 2.94 (m, 2H, CH2), 2.19 (s, 6H, o-CH3), 2.14 (s, 3H, p-CH3), 1.09 (d, 2JHP = 5.4 Hz, 18H, P(CH3)3). 31P{1H} NMR (161.8 MHz, C6D6): δ 207.0 (br, 1P), 41.6 (s, 1P), 3.8 (s, 2P). 13 C{1H} NMR (100.5 MHz, C6D6): δ 148.8 (m), 144.8 (m), 138.1 (d, 1 JCP = 11.6 Hz), 137.8 (s), 136.1 (s), 133.4 (d, 1JCP = 14.5 Hz), 131.6 (d, 3JCP = 1.75 Hz), 129.2(s), 128.6 (s), 128.2 (s), 127.4 (d, 2JCP = 8.5 Hz), 127.3 (s), 120.7 (d, 2JCP = 4.83 Hz), 116.3 (d, 3JCP = 2.8 Hz), 47.4 (s), 28.7 (m, PMe3), 21.1 (s), 19.0 (s). UV−vis [THF; λ, nm (ε, M−1 cm−1)]: 420 (3.99 × 104). Anal. Calcd for C35H47N2P4Co: C, 61.95; H, 6.98; N, 4.13. Found: C, 61.77; H, 7.06; N, 4.24.

localized on the Co center rather than a phosphinyl radical. The reduction of the ligand to an NHP− phosphido also leads to a decrease in the Co−P bond order to a single bond, with a Co− P distance more in line with the previously reported [PPP]Co(CO)2 complex G that also features a pyramidal geometry about P.4a Mulliken population analysis of the oxidized compound 4 reveals that the unpaired spin density resides on both the Co and PNHP atoms with little electron delocalization onto the other atoms in the molecule [Mulliken spin densities: 1.58 (Co), −0.42 (PNHP); Figure 6]. The computed distribution of the unpaired electron onto the 59Co and only one 31P center is consistent with the hyperfine coupling pattern observed in the EPR spectrum of 4 (vide supra). The oxidation of 3 to 4 is best described as cobalt-based, as demonstrated by the increase in the computed natural charge on cobalt by ∼1 upon oxidation, while the sum of natural charges on the NHP ligand atoms changes very little. However, the increase in the Co−P bond distance and distribution of the unpaired spin onto both cobalt and phosphorus suggests a more complicated electronic state, resulting from strong spin polarization, antiferromagnetic exchange between a cobalt-centered triplet and a phosphorusbased radical, or some combination of the two. More extensive computational and spectroscopic studies will be required to further understand the electronic structure of complex 4.



CONCLUSION In summary, we have isolated and fully characterized two new NHP+ transition-metal complexes, including the first structurally characterized examples of a Co(NHP+) compound without stabilizing carbonyl ligands. Both [PP]Co(CO)2 (2) and [PP]Co(PMe3)2 (3) exhibit a planar geometry about the PNHP atom and a short Co−P distance indicative of CoP double bonds. These compounds therefore lie in sharp contrast to the previously reported [PPP]Co(CO)2 complex, which has a pyramidal NHP geometry and a Co−P distance that is >0.2 Å longer.4a Cyclic voltammetry of the monometallic Co−I complexes 2 and 3 revealed reductions at nearly identical potentials and oxidation processes whose potentials and reversibility were highly dependent on the L-type donor ligand (CO or PMe3) coordinated to the Co center. Therefore, the reductive events were assigned as NHP+-based, and the oxidative events were deduced to be metal-centered. The low oxidation potential of 3 permitted its oxidation with a mild oxidant to generate the S = 1/2 cation 4. Rather than exclusive oxidation of the Co center from Co−I to Co0, EPR and computational studies of 4 suggest delocalization of the unpaired electron between the Co and PNHP atoms. DFT calculations also support that the reduction of complex 3 is largely NHP+-centered. Thus, the experimental and computational studies reported herein support the redox noninnocence of NHP+ ligands and suggest further applications in cooperative metal−ligand redox processes.



EXPERIMENTAL SECTION

General Considerations. Unless specified otherwise, all manipulations were performed under an inert atmosphere using standard Schlenk or glovebox techniques. Glassware was oven-dried before use. n-Pentane, tetrahydrofuran (THF), toluene, diethyl ether (Et2O), benzene, and dichloromethane were degassed and dried by sparging with ultrahigh-purity argon gas followed by passage through a series of drying columns using a Seca Solvent System by Glass Contour. All solvents were stored over 3 Å molecular sieves. Deuterated solvents F

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry [[PP]Co(PMe3)2][BPh4] (4). Complex 3 (52 mg, 0.077 mmol) was dissolved in THF (2 mL) and added to a suspension of [Fc][BPh4] (39 mg, 0.077 mmol) in THF (2 mL). The reaction mixture was stirred for 2 h at room temperature, and then the volatiles were removed in vacuo. The residue was washed with 10 mL of benzene and dried in vacuo, resulting in a brown paramagnetic solid. Yield: 48 mg, 63%. 1H NMR (400 MHz, THF-d8): δ 21.7 (2770 Hz, 18H), 13.6 (410 Hz, 4H), 12.6 (279 Hz, 1H), 8.8 (180 Hz, 1H), 7.4 (s, BPh4, 8H), 6.9 (s, BPh4, 8H), 6.7 (s, BPh4, 4H), 5.8 (215 Hz, 2H)), 4.2 (151 Hz, 1H), 3.1 (237 Hz, 6H), 1.5 (57 Hz, 3H), −5.3 (1190 Hz, two signals overlapping, 6H), −11.1 (738 Hz, 2H), −13.7 (826 Hz, 2H). (Note: One expected resonance corresponding to 1H could not be discerned due to either broadening or accidental overlap with another resonance or solvent signal.) μeff (THF-d8): 1.7 μB. UV−vis [THF; λ, nm (ε, M−1 cm−1)]: 570 (709 × 103). Multiple attempts to obtain satisfactory elemental analysis data were unsuccessful due to the thermal and air/moisture sensitivity of complex 4. X-ray Crystallography. All operations were performed on a Bruker-Nonius Kappa Apex2 diffractometer, using graphite-monochromated Mo Kα radiation. All diffractometer manipulations, including data collection, integration, scaling, and absorption corrections, were carried out using the Bruker Apex2 software.14 Preliminary cell constants were obtained from three sets of 12 frames. Crystallographic parameters are summarized in Table S1, and data collection, solution, and refinement details are included on pp S8−S12 of the Supporting Information. Electrochemistry. Cyclic voltammetry measurements were carried out in a glovebox under a dinitrogen atmosphere in a onecompartment cell using a CH Instruments electrochemical analyzer. A glassy carbon electrode and platinum wire were used as the working and auxiliary electrodes, respectively. A silver wire was used as a pseudoreference electrode, and potentials are reported relative to an internal ferrocene/ferrocenium reference. Solutions of electrolyte (0.2 M [nBu4N][ClO4] in THF) and analyte (2 mM) were also prepared in the glovebox. EPR Spectroscopy. X-band EPR spectra were obtained on a Bruker ElexSys E500 EPR spectrometer (fitted with a cryostat for measurements at 3 K). The EPR sample was a noncrystalline powder sample that was dissolved in THF, and the spectrum was measured at 3 K. The spectrum was referenced to diphenylpicrylhydrazyl (DPPH; g = 2.0037) and modeled using EasySpin for MATLAB. The EPR spectrum of complex 4 was simulated as a rhombic signal with g values of 2.227, 2.113, and 2.028 using a combination of Gaussian (2.32 mT) and Lorentzian (2.18 mT) peak-to-peak line widths. The hyperfine coupling in the spectrum was fit using hyperfine coupling to the 59Co nucleus (I = 7/2, Ax = 110.1 MHz, Ay = 115.4 MHz, Az = 162.9 MHz) and one 31P nucleus (I = 1/2, Ax = 111.7 MHz, Ay = 245.4 MHz, Az = 175.0 MHz). Computational Details. All calculations were performed using Gaussian0915 for the Linux operating system. DFT calculations were carried out using the B3LYP hybrid functional.16 A mixed basis set was employed, using the LANL2DZ(p,d) double-ζ basis set with effective core potentials for the P and Co atoms17 and Gaussian09’s internal LANL2DZ basis set (equivalent to D95V18) for C, O, N, and H atoms. Using crystallographically determined geometries as the starting point, the geometries were optimized to a minimum, followed by analytical frequency calculations to confirm that no imaginary frequencies were present. Single-point NBO calculations were subsequently performed on the optimized geometries of 2−5 using NBO 3.1,19 as implemented in Gaussian09.





XYZ coordinates of DFT-optimized geometries computed for 2−5 (PDF) X-ray crystallographic data in CIF format for 2 and 3 (CIF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Christine M. Thomas: 0000-0001-5009-0479 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This material is based on work supported by the National Science Foundation under Award CHE-1148987. Kathryn Gramigna (Brandeis University), Bing Wu (Brandeis University), and Matthew J. T. Wilding (Harvard University) are acknowledged for their assistance with EPR data collection. The authors are also grateful for access to the Brandeis University high-performance computing cluster.



REFERENCES

(1) (a) Bansal, R. K.; Gudat, D., Recent Developments in the Chemistry of N-Heterocyclic Phosphines. In Phosphorus Heterocycles II; Springer: Berlin, 2009; Vol. 21, pp 63−102. (b) Gudat, D. Cationic Low Coordinated Phosphorus Compounds As Ligands: Recent Developments. Coord. Chem. Rev. 1997, 163, 71−106. (c) Gudat, D. Diazaphospholenes: N-Heterocyclic Phosphines between Molecules and Lewis Pairs. Acc. Chem. Res. 2010, 43, 1307−1316. (d) Cowley, A. H.; Kemp, R. A. Synthesis And Reaction Chemistry Of Stable TwoCoordinate Phosphorus Cations (Phosphenium Ions). Chem. Rev. 1985, 85, 367−382. (e) Rosenberg, L. Metal Complexes Of Planar PR2 Ligands: Examining The Carbene Analogy. Coord. Chem. Rev. 2012, 256, 606−626. (2) (a) Hardman, N. J.; Abrams, M. B.; Pribisko, M. A.; Gilbert, T. M.; Martin, R. L.; Kubas, G. J.; Baker, R. T. Molecular and Electronic Structure of Platinum Bis(N-arylamino)phosphenium Complexes including [Pt(phosphane)(phosphenium)(N-heterocyclic carbene)]. Angew. Chem., Int. Ed. 2004, 43, 1955−1958. (b) Nakazawa, H.; Miyoshi, Y.; Katayama, T.; Mizuta, T.; Miyoshi, K.; Tsuchida, N.; Ono, A.; Takano, K. Syntheses, Structures, and DFT Calculations of Phosphenium Phosphite Complexes of Molybdenum: Preference of Nonbridging Form to Bridging Form of a Donor Group. Organometallics 2006, 25, 5913−5921. (c) Price, J. T.; Lui, M.; Jones, N. D.; Ragogna, P. J. Group 15 Pnictenium Cations Supported by a Conjugated Bithiophene Backbone. Inorg. Chem. 2011, 50, 12810− 12817. (d) Gudat, D.; Haghverdi, A.; Nieger, M. Complexes With Phosphorus Analogues Of Imidazoyl Carbenes: Unprecedented Formation Of Phosphenium Complexes By Coordination Induced P—Cl Bond Heterolysis. J. Organomet. Chem. 2001, 617−618, 383− 394. (e) Caputo, C. A.; Brazeau, A. L.; Hynes, Z.; Price, J. T.; Tuononen, H. M.; Jones, N. D. A Cation-Captured Palladium(0) Anion: Synthesis, Structure, and Bonding of [PdBr(PPh3)2]− Ligated by an N-Heterocyclic Phosphenium Cation. Organometallics 2009, 28, 5261−5265. (f) Caputo, C. A.; Jennings, M. C.; Tuononen, H. M.; Jones, N. D. Phospha-Fischer Carbenes: Synthesis, Structure, Bonding, and Reactions of Pd(0)- and Pt(0)-Phosphenium Complexes. Organometallics 2009, 28, 990−1000. (g) Spinney, H. A.; Yap, G. P. A.; Korobkov, I.; DiLabio, G.; Richeson, D. S. Construction of a Stable N-Heterocyclic Phosphenium Cation with an Electron-Rich Frame-

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b02374. Spectroscopic data for complexes 2−4, crystallographic data collection and refinement details for 2 and 3, and G

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry work and Its Complexation to Rhodium. Organometallics 2006, 25, 3541−3543. (h) Nakazawa, H.; Yamaguchi, Y.; Mizuta, T.; Miyoshi, K. Cationic Phosphenium Complexes of Group 6 Transition Metals: Reactivity, Isomerization, and X-ray Structures. Organometallics 1995, 14, 4173−4182. (i) Nickolaus, J.; Bender, J.; Nieger, M.; Gudat, D. Sterically Controlled Synthesis and Nucleophilic Substitution Reactions of Di- and Trimeric N-Heterocyclic Phosphenium Metal(0) Halides. Eur. J. Inorg. Chem. 2014, 2014, 3030−3036. (j) Förster, D.; Nickolaus, J.; Nieger, M.; Benkõ, Z.; Ehlers, A. W.; Gudat, D. DonorFree Phosphenium-Metal(0)-Halides with Unsymmetrically Bridging Phosphenium Ligands. Inorg. Chem. 2013, 52, 7699−7708. (k) Hutchins, L. D.; Paine, R. T.; Campana, C. F. Structure And Bonding In A Phosphenium Ion-Metal Complex, CH3NCH2CH2N(CH3)PMo(η5C5H5)(CO)2. An Example Of A Molybdenum-Phosphorus Multiple Bond. J. Am. Chem. Soc. 1980, 102, 4521−4523. (l) Hutchins, L. D.; Reisacher, H. U.; Wood, G. L.; Duesler, E. N.; Paine, R. T. Synthesis And Structure Of Metallophosphenium Complexes Derived From Related Cyclic And Acyclic Aminohalophosphines. J. Organomet. Chem. 1987, 335, 229−237. (3) (a) Nakazawa, H.; Yamaguchi, Y.; Kawamura, K.; Miyoshi, K. Comparison of the Reactivity of Cationic Phosphenium Complexes of Iron Containing a Group 14 Element Ligand. Organometallics 1997, 16, 4626−4635. (b) Stadelmann, B.; Bender, J.; Forster, D.; Frey, W.; Nieger, M.; Gudat, D. An Anionic Phosphenium Complex As An Ambident Nucleophile. Dalton Trans. 2015, 44, 6023−6031. (c) Light, R. W.; Paine, R. T. Interaction Of The Dicoordinate Phosphorus Cation 1,3-Dimethyl-1,3,2-Diazaphospholidide With Transition Metal Nucleophiles. J. Am. Chem. Soc. 1978, 100, 2230−2231. (d) Hutchins, L. D.; Duesler, E. N.; Paine, R. T. Structure and bonding in a phosphenium ion-iron complex, Fe[η5-(CH3)5C5](CO)2[PN(CH3)CH2CH2NCH3]. A demonstration of phosphenium ion acceptor properties. Organometallics 1982, 1, 1254−1256. (4) (a) Pan, B.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C. M. Coordination of an N-Heterocyclic Phosphenium Containing Pincer Ligand to a Co(CO)2 Fragment Allows Oxidation To Form an Unusual N-Heterocyclic Phosphinito Species. Organometallics 2011, 30, 5560−5563. (b) Burck, S.; Daniels, J.; Gans-Eichler, T.; Gudat, D.; Nättinen, K.; Nieger, M. N-Heterocyclic Phosphenium, Arsenium, and Stibenium Ions as Ligands in Transition Metal Complexes: A Comparative Experimental and Computational Study. Z. Anorg. Allg. Chem. 2005, 631, 1403−1412. (c) Hutchins, L. D.; Light, R. W.; Paine, R. T. Synthesis, structure, and bonding of the bis(phosphenium) iondicobalt carbonyl complex Co2(CO)5(μ.-PN(CH3)CH2CH2NCH3)2. Inorg. Chem. 1982, 21, 266−272. (5) Snow, S. S.; Jiang, D. X.; Parry, R. W. Formation Of A Nickel Carbonyl Cation Containing A Cyclophosphenium Ligand By Hydride Abstraction. Inorg. Chem. 1987, 26, 1629−1631. (6) (a) Evers-McGregor, D. A.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C. M. N-heterocyclic phosphenium and phosphido nickel complexes supported by a pincer ligand framework. Dalton Trans. 2016, 45, 1918−1929. (b) Bezpalko, M. W.; Foxman, B. M.; Thomas, C. M. Use of a Bidentate Ligand Featuring an N-Heterocyclic Phosphenium Cation (NHP+) to Systematically Explore the Bonding of NHP+ Ligands with Nickel. Inorg. Chem. 2015, 54, 8717−8726. (7) Pan, B.; Xu, Z.; Bezpalko, M. W.; Foxman, B. M.; Thomas, C. M. N-Heterocyclic Phosphenium Ligands as Sterically and ElectronicallyTunable Isolobal Analogues of Nitrosyls. Inorg. Chem. 2012, 51, 4170−4179. (8) Day, G. S.; Pan, B.; Kellenberger, D. L.; Foxman, B. M.; Thomas, C. M. Guilty as charged: non-innocent behavior by a pincer ligand featuring a central cationic phosphenium donor. Chem. Commun. 2011, 47, 3634−3636. (9) (a) Wong, W.-K.; Hou, A.; Guo, J.; He, H.; Zhang, L.; Wong, W.Y.; Li, K.-F.; Cheah, K.-W.; Xue, F.; Mak, T. C. W. Synthesis, Structure And Near-Infrared Luminescence Of Neutral 3d-4f Bi-Metallic Monoporphyrinate Complexes. J. Chem. Soc., Dalton Trans. 2001, 3092−3098. (b) Takemura, Y.; Nishida, T.; Kure, B.; Nakajima, T.; Iida, M.; Tanase, T. Stepwise Construction of Au4Ag2Cu2 Coinage Rings Supported by Linear Tetraphosphine Ligands. Chem. - Eur. J.

2011, 17, 10528−10532. (c) Van der Sluys, L. S.; Miller, M. M.; Kubas, G. J.; Caulton, K. G. Deprotonation Of Uncharged TransitionMetal Dihydrogen Complexes With Copper Alkoxides, Characterization Of The Heterometallic Complexes (PR3)xMHyCu(PR3) (M = Iron, Tungsten), And X-Ray Crystal Structure of (PEtPh2)3FeH3Cu(PEtPh2). J. Am. Chem. Soc. 1991, 113, 2513−2520. (d) Koshevoy, I. O.; Karttunen, A. J.; Shakirova, J. R.; Melnikov, A. S.; Haukka, M.; Tunik, S. P.; Pakkanen, T. A. Halide-Directed Assembly of Multicomponent Systems: Highly Ordered AuI−AgI Molecular Aggregates. Angew. Chem., Int. Ed. 2010, 49, 8864−8866. (10) (a) Melenkivitz, R.; Mindiola, D. J.; Hillhouse, G. L. Monomeric Phosphido and Phosphinidene Complexes of Nickel. J. Am. Chem. Soc. 2002, 124, 3846−3847. (b) Termaten, A. T.; Aktas, H.; Schakel, M.; Ehlers, A. W.; Lutz, M.; Spek, A. L.; Lammertsma, K. Terminal Phosphinidene Complexes CpR(L)MPAr of the Group 9 Transition Metals Cobalt, Rhodium, and Iridium. Synthesis, Structures, and Properties. Organometallics 2003, 22, 1827−1834. (11) Banerjee, S.; Karunananda, M. K.; Bagherzadeh, S.; Jayarathne, U.; Parmelee, S. R.; Waldhart, G. W.; Mankad, N. P. Synthesis and Characterization of Heterobimetallic Complexes with Direct Cu−M Bonds (M = Cr, Mn, Co, Mo, Ru, W) Supported by N-Heterocyclic Carbene Ligands: A Toolkit for Catalytic Reaction Discovery. Inorg. Chem. 2014, 53, 11307−11315. (12) An anonymous reviewer suggested that a more efficient synthetic route to Na[Co(CO)4] uses anhydrous NaOH in place of Na/Hg to reduce Co2(CO)8. However, the authors found that, in our hands, Na[Co(CO)4] was generated more cleanly using the Na/Hg method. (13) (a) Sur, S. K. Measurement Of Magnetic-Susceptibility And Magnetic-Moment Of Paramagnetic Molecules In Solution By HighField Fourier-Transform Nmr-Spectroscopy. J. Magn. Reson. 1989, 82, 169−173. (b) Evans, D. F. The Determination Of The Paramagnetic Susceptibility Of Substances In Solution By Nuclear Magnetic Resonance. J. Chem. Soc. 1959, 2003−2005. (14) Apex 2, Version 2 User Manual, M86-E01078; Bruker Analytical X-ray Systems: Madison, WI, 2006. (15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.2; Gaussian, Inc.: Wallingford, CT, 2009 (see the Supporting Information for the full reference). (16) (a) Becke, A. D. Density-functional thermochemistr. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (b) 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. (17) (a) Hay, P. J.; Wadt, W. R. Ab Initio Effective Core Potentials for Molecular Calculations - Potentials for K to Au Including the Outermost Core Orbitals. J. Chem. Phys. 1985, 82, 299−310. (b) Hay, P. J.; Wadt, W. R. Ab Initio Effective Core Potentials for Molecular Calculations - Potentials for the Transition-Metal Atoms Sc To Hg. J. Chem. Phys. 1985, 82, 270−283. (c) Hay, P. J.; Wadt, W. R. Ab initio effective core potentials for molecular calculations. Potentials for main group elements Na to Bi. J. Chem. Phys. 1985, 82, 284−298. (d) Roy, L. E.; Hay, P. J.; Martin, R. L. Revised Basis Sets for the LANL Effective Core Potentials. J. Chem. Theory Comput. 2008, 4, 1029− 1031. (18) Dunning, T. H.; Hay, P. J. Modern Theoretical Chemistry; Plenum: New York, 1976; Vol. 3. (19) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO, version 3.1.

H

DOI: 10.1021/acs.inorgchem.6b02374 Inorg. Chem. XXXX, XXX, XXX−XXX