Exploring the Intricacies of Weak Interactions in Metal–Metal Bonds

Article pubs.acs.org/IC

Exploring the Intricacies of Weak Interactions in Metal−Metal Bonds Using an Unsymmetrical Carbonyl Precursor and a Triple-Bonded W26+ Paddlewheel Gina M. Chiarella,† Chao Feng,§ Carlos A. Murillo,*,†,‡ and Qinliang Zhao*,§ †

Department of Chemistry, Texas A&M University, P.O. Box 3012, College Station, Texas 77842-3012, United States Department of Chemistry, University of Texas at El Paso, El Paso, Texas 79968, United States § Department of Chemistry, University of the Pacific, 3601 Pacific Avenue, Stockton, California 95211, United States ‡

S Supporting Information *

ABSTRACT: Stepwise reaction of W(CO)6 with tetramethylated bicyclic guanidinate ligands, characterized by a central C(N)3 unit joining two fused six-membered rings with CH2CMe2CH2 units spanning two of the nitrogen atoms, allowed isolation of W2(μ-CO)2(μ-TMhpp)2(η2-TMhpp)2, 1, a precursor of W2(TMhpp)4Cl2 (J. Am. Chem. Soc. 2013, 135, 17889; TMhpp = [(CH2CMe2CH2)2(C(N)3)]). Subsequent heating of 1 followed by reaction with TlPF6 generates [W2(TMhpp)4](PF6)2, 2. Compound 1 has an edge-sharing bioctahedral (ESBO) arrangement with a W2(μ-CO)24+ core having semibridging carbonyl groups, while 2 has a paddlewheel structure with a W26+ core spanned by four tetramethyl-substituted bicyclic guanidinate ligands. This compound also has hexafluorophosphate anions along the metal−metal bond that are nestled within methylene groups with the aid of a network of weak C−H···F interactions that prevent a close approach of the fluorine atoms to the dimetal unit. Theoretical computations were carried out on ditungsten model complexes supported by three ligand sets: bicyclic guanidinate, guanidinate, and formamidinate. The computations show that the π-accepting ability of the carbonyl groups significantly lowers the energy of the σ* orbital, and thus, the energy falls below that of the δ orbital. This information along with the diamagnetism of both 1 and 2as shown by the sharp signals in the 1H NMR spectra that support a lack of unpaired electrons (S = 0)is consistent with the electronic configuration of σ2π2σ*2δ2 (π2δ2) and thus a formal bond order of 2 for 1 and σ2π4 for the triple-bonded W26+ core in 2. A comparison of the W−W bond lengths in 2, its chloro precursor W2(TMhpp)4Cl2, and the corresponding analogue W2(hpp)4Cl2 shows a substantial effect from the axially coordinated ligand, distal lone pair in determining the length of the metal−metal bond for these paddlewheel species. The importance of the ligands in tuning the energy level of the metal−metal bonds that may lead to dramatic changes in physical properties is also discussed. It is noteworthy that bicyclic guanidinates with the strongest π-donating ability push upward the energy level of the δ orbital, thus allowing the compounds to be easily oxidized.

■

INTRODUCTION There was a time when guanidinate ligands, that is, those having a C(N)3 backbone, were rarely used in coordination chemistry. However, in the past few years, their relevance has increased because of their demonstrated ability to stabilize mononuclear, dinuclear, and polynuclear species.1,2 Moreover, it is not unusual to find the parent compounds as well as some metal-containing species being employed in catalytic processes.3 For species with dimetal units, the use of the commercially available bicyclic Hhpp (the neutral bicyclic guanidine compound 1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2a]pyrimidine, the anion shown in Scheme 1) has been preponderant. This ligand has been utilized for the syntheses of compounds with rare and often unprecedented oxidation states such as those in Nb24+,4 Cr22+,5 Mo26+,6 W26+,7 Re28+,8 Ir26+,9 and Pd26+ units.10 The bicyclic guanidinates have also been employed for preparation of a group of intriguing © XXXX American Chemical Society

compounds having fractional formal bond orders that yield paramagnetic species such as those having Cr23+,5 Cr25+,11 Mo25+,6 W25+,12 Re27+,13 Os27+,14 and Rh25+ cores.15 Many of these compounds have been studied by EPR spectroscopy.16 Perhaps one of the most striking properties is their extraordinary capacity for destabilizing quadruple-bonded M24+ species (M = group 6 transition element), which allows facile oxidation that in turn generates triple-bonded M26+ compounds. Because of this destabilization, such quadruplebonded compounds act as very strong reducing agents.7b,17 The chief reason why these M24+ precursors are so easily oxidized lies in the destabilization of the electrons in δ bonds of the dimetal unit that strongly interact with the π electrons of the central C(N)3 unit. Initially, there were complications isolating Received: July 19, 2016

A

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 1. Structures of the Bicyclic Guanidinate, Guanidinate, and Formamidinate Ligands

Scheme 2. Synthetic Route to Complexes 1 and 2

pure compounds because some of the byproducts were insoluble and difficult to remove when species such as M2(hpp)4 (M = Mo, W) were used as reductants. However, the design of synthetic routes to alkyl-substituted ligands such as TMhpp (the anion of 3,3,9,9-tetramethyl-1,5,7triazabicyclo[4.4.0]dec-4-ene, Scheme 1) and TEhpp (the anion of 3,3,9,9-tetraethyl-1,5,7-triazabicyclo[4.4.0]dec-4-ene, Scheme 1) has eliminated such concerns,17g,18 and most dimetal species with these ligands have significant solubility even in nonpolar organic solvents such as hexanes.19 These reactive yet thermally stable quadruple-bonded compounds not only have ionization energies that are less than that of cesium but also have electrode potentials that make them stronger reducing agents than the well-known and highly used decamethylcobaltocene.19 Here we report the syntheses and structural characterization of two precursors of W2(TMhpp)4, namely, the edge-sharing bioctahedral (ESBO) compound W2(μ-CO)2(μ-TMhpp)2(η2TMhpp)2 (1) that has semibridging carbonyl groups and the paddlewheel species [W2(TMhpp)4](PF6)2 (2). The latter provides relevant insight into the effect on metal−metal bond lengths of lone electron pairs from the axial ligands. With the aid of DFT computations on the ditungsten complexes supported by bicyclic guanidinate, guanidinate, and formamidinate ligands, some of the intricacies of metal−metal bonding are analyzed.

■

2W(CO)6 + 4HTMhpp → W2(μ‐CO)2 (μ‐TMhpp)2 (η2‐TMhpp)2 + 2H 2 + 10CO (1) 2

W2(μ‐CO)2 (μ‐TMhpp)2 (η ‐TMhpp)2 + 2o‐C6H4Cl 2 → W2(μ‐TMhpp)4 Cl 2 + (ClC6H4)2 (2,6‐dichlorobiphenyl) (2)

W2(TMhpp)4 Cl 2 + 2TlPF6 → [W2(TMhpp)4 ](PF6)2 + 2TlCl (3)

Structure. The structure of W2(μ-CO)2(μ-TMhpp)2(η2TMhpp)2 (1) in Figure 1 shows a W24+ core spanned by two semibridging carbonyl and two bicyclic TMhpp ligands. In addition, each W atom is bound to a chelating TMhpp anion. Despite the differences in binding mode of the two types of guanidinate ligands, the carbon atom in the NC(N)N unit retains sp2 hybridization. Accordingly, the sum of the three independent N−C−N angles is 360°, as shown by the data in Table 1. The structure of 1 is reminiscent of that of the carbonyl analogues having hpp or formamidinate ligands.7b,22 A comparison between bond lengths and angles of some carbonyl analogues is provided in Table 2. The W−CCO distances of 1.960(5) and 2.287(5) Å indicate that the core is rather skewed. Although not unprecedented, the formal divalent oxidation state of the metal atoms in 1 is higher than those found in most transition metal carbonyls.23 The bonding in the now ubiquitous transition metal carbonyls has been customarily described as a combination of two synergistic processes with one involving the formation of a σ bond (carbon → metal) derived from the interaction of the lone electron pair of the carbon atom in carbon monoxide with an unoccupied d orbitals in the metal unit. The second process is the back-bonding attained from the interaction of electrons in a filled metal d orbital with an empty π* MO of the carbonyl group.24 The stretching frequencies (νCO) have often been used to evaluate the strength of the carbonyl binding. In general, for d-block metals, the νCO frequencies are lower than those in free CO but only by 100−200 cm−1, while for bridging CO groups in such metal carbonyls, the νCO frequencies are ∼100 cm−1, below those for terminal CO groups. The νCO of 1725 cm−1 in 1 is comparable to that in the hpp analogue for which the corresponding frequency is 1708 cm−1,7b as well as that of

RESULTS AND DISCUSSION

Syntheses. Reaction of tungsten hexacarbonyl with neutral HTMhpp in o-dichlorobenzene at temperatures slightly below the normal boiling point produces the ESBO intermediate W2(μ-CO)2(μ-TMhpp)2(η2-TMhpp)2, 1. When this compound is heated at a temperature slightly above the normal boiling point of o-dichlorobenzene, 1 is further oxidized to W2(TMhpp)4Cl2.19 To reach a temperature that produces a high yield of the dichloro compound, a system equipped with a bubbler with a column of mercury of ∼7.5 cm is used.20 This arrangement increases the internal pressure and thus the boiling temperature of the solvent. Addition of TlPF6 to a solution of W2(TMhpp)4Cl2 allows formation of 2. The process is summarized in eqs 1−3 and pictorially in Scheme 2:21 B

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

(s), 2.67 (s), 2.34 (s), 1.15 (s), and 0.85 (s) ppm having a ratio of 1:1:1:1:3:3.26 One set of signals having a 1:1:3 ratio corresponds to the bridging, while the other one corresponds to the chelating ligands. Thus, the solution structure is consistent with that in the solid state.27 The W−W distance of 2.4472(8) Å suggests that there is substantial metal−metal bonding. In general, ESBO compounds such as those having M2(μ-O)2n+ cores have a maximum metal−metal bond order of 3, as shown in Scheme 3. If the z direction is perpendicular to the M2(μ-O)2 entity, then these four atoms lie in the xy plane.28 In this environment the dxy and dz2 orbitals would be used in metal-to-ligand bonding and thus are unavailable for metal−metal bonding. Therefore, for such ESBO species, the maximum possible bond order would be 3 when there are six metal-based electrons. However, it is important to note that the relative ordering of MOs in ESBO species is quite dependent on the nature of the ligands and the metal atom.29 Indeed, DFT calculations on the hpp analogue have shown that several of the MOs generally expected to have higher energy are found much to be lower. For example, the σ* orbitalcomposed of the out-of-phase combination of the metal-based dx2−y2 orbitalswould be expected to be the MO with the highest energy but is instead the HOMO−1. The reason for this stabilization is the backbonding interaction with one of the π* orbitals of the μ-CO ligands. The HOMO (bg) shows a bonding tungsten−tungsten interaction of δ symmetry caused by the in-phase combination of the dyz orbitals. Furthermore, there is back-bonding to the second set of π* orbitals of the μ-CO ligands. This situation is in accord with previous finding showing that semibridging carbonyl ligands are capable of acting as π acceptors from filled distal metal d-orbitals.30 The HOMO−3 is the ag (σ), while the HOMO−2 is the au (π) MO that is derived from the in-phase combination of the dxz orbitals. This is summarized in Scheme 3 where an idealized C2h symmetry was chosen for calculations on the model W2(μ-CO)2(μ-NHC(H)NH)2(η2-NHC(H)NH)2. By analogy, the electronic configuration for 1 can be described as σ2π2σ*2δ2δ*0π*0 (π2δ2). This is consistent with a formal bond order of 2 between metal atoms (vide infra.) As shown in Figure 1, the cation in 2 represents one of a small but growing number of W26+ complexes that have been structurally characterized.19,31 The overall structures of 2 and the dichloro precursor of 2, W2(TMhpp)4Cl2 in Figure 2, are those of a paddlewheel with four bridging alkyl-substituted bicyclic guanidinates. Similar to the precursor of 2, W2(TMhpp)4Cl2, that has two very weakly interacting ligands in axial positions (2.9781[4] Å),19 2 has no strong σ-bonding contacts along the W−W axis.19 Nevertheless, the PF6 anions are located along this axis but have W···F distances of 2.786(4) and 2.810(4) Å, which are considerably longer than those expected for a typical W−F covalent bond. Each PF6 anion is nestled within methylene groups with the aid of a series of weak C−H···F interactions and located at the far side of the dimetal axis. The eight C−H···F contacts have distances varying from 2.40 to 2.58 Å. The existence of such contacts prevents the anions from getting closer to the positively charged dimetal unit. In W2(TMhpp)4Cl2, the four Cl···H(methylene) distances from each of the Cl atoms vary from 2.56(2) to 2.62(2) Å. Similar to the fluorine atoms in 2, the chlorine atoms are nestled within the guanidinate ligands and located along the W−W bond. The W···Cl separations of 2.9781[4] Å are long when compared to related covalent bonds. For comparison, the

Figure 1. Displacement ellipsoid plots at the 50% probability level of 1 (top) showing the semibridging carbonyl groups and the cation in 2· 3CH2Cl2 (bottom). The two halves of molecule 1 are related by an inversion center.

1745 cm−1 for an isostructural formamidinate compound with the formula W2(μ-CO)2(μ-DAniF)2(η2-DAniF)2,22 where DAniF = N,N′-di-p-anisylformamidinate (Scheme 1). This value is considerably smaller than that of 2143 cm−1 in unbound carbon monoxide23 and is in the lower end of the range of 1700−1860 cm−1 commonly found for bridging carbonyls.23 This low stretching frequency is consistent with substantial back-bonding despite the high formal oxidation state of the tungsten atoms in 1. The C23−O1 distance of 1.184(6) Å is considerably longer than that in CO itself (1.128 Å)23 but very close to that of 1.184−1.188 Å calculated for the optimized model compound Ru2(BF)2(CO)6 that has unsymmetrical carbonyls as well as the isoelectronic fluoroborylene ligands.25 This distance is also close to that of 1.2194 Å calculated for the parent hpp analogue.7b It should be noted that most terminal C−O bond lengths in metal carbonyl molecules are ∼1.15 Å, which also agrees with the calculated values in the Ru dinuclear model.25 The 1H NMR spectrum from a benzene-d6 solution of 1 in Figure S10 shows two sets of guanidinate signals having the same relative ratio of 1:1. These signals appear at 3.36 (s), 2.81 C

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Table 1. Selected Bond Lengths (Å) and Angles (deg) for W2(μ-CO)2(μ-TMhpp)2(η2-TMhpp)2, (1) and [W2(TMhpp)4](PF6)2 (2·3CH2Cl2) W2(μ-CO)2(μ-TMhpp)2(η2-TMhpp)2, 1

[W2(TMhpp)4](PF6)2, 2 in 2·3CH2Cl2 Bond Lengths

W1−W1i W1−C23 W1−C23i C23−O1 W1−N1 W1−N2i W1−N4 W1−N5

2.4472(8) 1.960(5) 2.287(5) 1.184(6) 2.139(4) 2.128(4) 2.190(4) 2.173(4)

W1−W2

W1−N2 W1−N4 W1−N7 W1−N10 W2−N1 W2−N5 W2−N8 W2−N11

2.2046(5) − − − 2.099(7) 2.097(7) 2.094(6) 2.075(7) 2.095(6) 2.073(7) 2.124(7) 2.102(6)

Angles N1−C1−N2bridge N4−C12−N5chelate W1−C23−W1i W1−C23−O1 W1−C23i−O1i N1−C1−N3 N2−C1−N3 N4−C12−N6 N5−C12−N6

120.1(4) 112.8(5) 69.9(2) 166.0(4) 124.1(4) 120.2(5) 119.6(4) 124.6(5) 122.6(5)

N1−C1−N2bridge N10−C34−N11bridge N7−C23−N8bridge N4−C12−N5bridge N1−C1−N3 N2−C1−N3

118.0(7) 118.2(7) 118.7(7) 116.7(7) 120.5(7) 121.6(8)

Table 2. Structural Comparisons of Analogous Carbonyl-Bridged ESBO Compounds with Bond Lengths in Å and Angles in deg W−W W−N(range) W−Nbridge W−Nchelate W−CCO W−CCO C−O W−C−O W−C−O C−W−W reference

1

W2(μ-CO)2(μ-hpp)2(η2-hpp)2

W2(μ-CO)2(μ- DAniF)2(η2-DAniF)2

2.4472(8) 2.139(4)−2.190(4) 2.1335[6] 2.1815[6] 1.960(5) 2.287(5) 1.184(6) 166.0(4) 124.1(4) 61.4(1) this work

2.4675(3) 2.128(3)−2.181(3) 2.133[4] 2.167[5] 1.944(4) 2.294(4) 1.188(5) 167.0(3) 122.4(3) 61.3(1) 7b

2.476(1) 2.103(9)−2.202(8) 2.12[13] 2.2065[12] 1.99(1) 2.28(1) 1.16(1) 163(1) 126.2(9) 60.2(3) 22

related covalent W−Cl distance in W2Cl84− is 2.429[5] Å,32 and it is 2.332(8) Å in W2(NEt2)4Cl2.33 In analysis of Table 3, it is striking that the W−W bond length of 2.2046(5) Å in 2 is ∼0.05 Å shorter than that in the precursor with axially coordinated chloride atoms. This value is over 60 times greater than the standard deviations and thus highly significant. The interaction of the chloride anion with the ditungsten unit in W2(TMhpp)4Cl2 has been previously examined by means of DFT computations.7b,34 Despite the long W···Cl separations that range from 2.85 to 3.06 Å, which are far longer than those expected for a W−Cl covalent bond, the computations on the hpp analogue indicate that they exert a significant effect on the metal−metal distance. This is attributed to the interaction of the σ* virtual orbital of the [W2(hpp)4]2+ fragment with the filled antisymmetric combination of the Cl pσ orbitals that gains a large Mulliken occupation (0.249, a qualitative indicator of its significant involvement in weakening the metal−metal bonding) in the W2(hpp)4Cl2 molecule.7b The question that arises is why a similar effect is not observed when hexafluorophosphate anions are used. To answer this query, it is

important to note that in the PF6 anion the P−F distances are short. In 2, these distances vary from 1.582(6) to 1.639(6) Å. This suggests that there is significant π bonding between the F and P atoms that lessens the ability of the lone pairs in the F atoms to interact with the [W2(TMhpp)4]2+ fragment, and consequently, the metal−metal bond is not destabilized. Therefore, the W−W length is shorter in 2 than that in the chloro analogue and almost the same as that in [W2(hpp)4](TFPB)2 (2.1920(3) Å, TFPB is the anion of tetrakis[3,5bis(trifluoromethyl)phenyl]borate), where the anions are very far from the dimetal unit.7b A similar effect is also observed for the Mo26+ compounds Mo2(hpp)4Cl(BF4) and [Mo2(hpp)4](BF4)2 for which the triple-bonded Mo−Mo distances are 2.1722(9) and 2.142(2) Å, respectively.6 Importantly, the Mo··· Cl distance in the first of those compounds of 2.983(2) Å is long. Additionally, a recent report describing the first Ru25+ compound devoid of axial interactions also showed the importance of σ and π interactions in determining the length of the metal−metal bond.35 Moreover, it is important to note that the weak hydrogen-bonding interactions in 2 can be ruled D

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 3. Ordering of the MOs in W2(μ-O)2(μNHC(H)NH)2(η2-NHC(H)NH)2 (Top) and W2(μ-CO)2(μNHC(H)NH)2(η2-NHC(H)NH)2 (Bottom)a

a

The z axis is perpendicular to the W2(μ-O)24+ or W2(μ-CO)24+ unit that lies on the xy plane. The brackets join the corresponding bonding and antibonding MOs, and the center shows the metal atom orbitals that give rise to the MOs. For example, overlap from two of the tungsten dyz orbitals results in the formation of δ and δ* orbitals. In this configuration the dz2 and the dxy orbitals are used for metal-toligand bonding and are unavailable for metal-to-metal bonding. For source data, see refs 29c and 22.

out as the source of the difference in W−W bond lengths in 2 and W2(TMhpp)4Cl2 because the chlorine atoms in the latter show similar contacts as those observed in 2, as shown in Figure 2. Similar to 1, the solution 1H NMR spectrum of 2 shown in Figure S12 is consistent with the solid-state structure showing only one set of guanidinate signals in a ratio of 1:1:3 (16:16:48). The diamagnetism supports the existence of a triple bond and a σ2π4 electronic configuration for 2. DFT Studies. Guanidinate and especially formamidinate ligands have been widely used in the studies of bimetallic complexes. Importantly, guanidinates have provided a window to yet unobserved electronic properties.1−3 Computational studies conducted for the complexes W2(μ-CO)2(μ-formamidinate)2(η2-formamidinate)2,22 W2(hpp)4,17a W2(hpp)42+,7b and W2(hpp)4Cl27b as well as W2(μ-O)2(TMhpp)2Cl429c offered some preliminary understanding. To provide a deeper grasp on these metal−metal species, DFT calculations were carried out for the series of model compounds I−IX depicted in Scheme 4. For comparison, both B3LYP and BH&HLYP functionals were used in calculations for all nine models. As shown in Tables 4 and 5, computations using functional BH&HLYP provided optimized geometry more closely resembling the experimental data, which is in accord with recent observations from calculations carried out on octahaloditechnetate anions.36 Because of this, subsequent data analyses were done using computations with the BH&HLYP functional. Molecular orbital analyses on model I (Figure 3) mimicking 1 indicate that the HOMO is a metal-based δ orbital. The metal-based σ and π orbitals are also occupied. Importantly, the energy of the σ* orbital significantly drops to values below those for the HOMO δ orbital. This is attributed to the

Figure 2. Upper image: Structure of 2 including the cation and anions. The PF6 anions nestled by methylene groups are positioned along the metal−metal axial. Lower image: View of the weak hydrogen bonding interactions from the methylene groups to the chlorine atoms in the precursor of 2, W2(TMhpp)4Cl2. Interstitial and disordered atoms and most hydrogen atoms have been removed for clarity.

interaction between the σ* orbital of the W2 unit with the CO π* orbital shown in Figure 4. Thus, the electronic configuration in 1 can be ascribed to σ2π2σ*2δ2 (π2δ2). This results in a formal bond order of 2. Comparison between the metal-based frontier orbitals for the three carbonyl-bridged ditungsten models I−III (Figure 5) shows consistency in the energy levels of the metal-based orbitals. It is clear that the relative energy for each occupied metal-based orbital decreases from that of the hpp model I to that of the guanidinate model II to that of the formamidinate model III. This trend correlates with the ligand electronic donor ability that decreases from hpp to noncyclic guanidinate to formamidinate. Similar trends were observed for the remaining model complexes (IV−VI in Figure 6 and VII−IX in Figure 7). Even though there are small differences in the calculated energy values for the different ligands, a σ2π2σ*2δ2 (π2δ2) electronic configuration for models II and III remains unaffected. Wiberg bond indices are useful to discriminate the slight differences in bond strength between bonds having the same formal bond order.37 From natural bond orbital (NBO) analysis, the Wiberg bond index for the double-bonded π2δ2 W2 unit in I is 1.25 (Figure 8). For comparison, the corresponding Wiberg bond index of 1.61 calculated for the ESBO compound W2(μ-O2)(TMhpp)2Cl4 is much larger. This dioxo compound has a short W−W bond length of 2.3318(8) Å E

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Table 3. Comparison of Selected Bond Lengths in Å of Triple-Bonded W26+/Bicyclic Guanidinate Compounds W2(hpp)4Cl2 W2(TMhpp)4Cl2 W2(TEhpp)4Cl2 [W2(hpp)4](PF6)2b [W2(hpp)4](TFPB)2c [W2(TMhpp)4](PF6)2, 2

W−W

W−N

W−Lax

reference

2.250(2) 2.2483[2]a 2.2575(5) 2.2083(7) 2.1920(3) 2.2046(5)

2.08(1) 2.096[16] 2.104[9] 2.091(4) 2.076[9] 2.095[10]

3.064(9) 2.9781[4] 2.8495[2] 2.774(5) nod 2.786(4); 2.810(4)

12 19 19 7b 7b this work

a When there is more than one multiple bond of the same type, the average is given and the standard deviations are in square brackets. bThere is considerable crystallographic disorder. cTFPB = the anion of tetrakis[3,5-bis(trifluoromethyl)phenyl]borate. dThe separation from the cation to the anion is long and falls well outside of the normal bonding range.

Scheme 4. Models I−IX Used for DFT Calculations

Table 4. Selected Bond Length (Å) for Models I−III from Experiments and Theoretical Calculations Using Functionals BH&HLYP and B3LYPa I

II III

ref 22 a

exp BH&HLYP B3LYP BH&HLYP B3LYP exp BH&HLYP B3LYP B3LYP

W−W

W−Nbridge

W−Nchelate

W−Cshort

W−Clong

C−O

2.4675(3) 2.546 2.493 2.598 2.513 2.476(1) 2.494 2.493 2.4898

2.133[4] 2.16 2.172 2.162 2.172 2.12[1] 2.15 2.16 −

2.167[5] 2.213 2.231 2.206 2.222 2.21[1] 2.221 2.238 −

1.944(4) 1.93 1.965 1.921 1.961 1.99(1) 1.954 1.978 1.966

2.294(4) 2.431 2.37 2.461 2.377 2.28(1) 2.399 2.372 2.337

1.188(5) 1.171 1.187 1.174 1.188 1.16(1) 1.164 1.181 1.219

When there is more than one multiple bond of the same type, the average is given and the standard deviations are in square brackets.

F

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Table 5. Selected W−W and W−Naverage Bond Lengths (Å) for Models IV−IX from Experiments and Theoretical Calculations Using Functionals BH&HLYP and B3LYPa IV

ref 7b V VI VII

ref 17a VIII IX

exp BH&HLYP B3LYP B3LYP BH&HLYP B3LYP BH&HLYP B3LYP exp BH&HLYP B3LYP B3LYP BH&HLYP B3LYP exp BH&HLYP B3LYP

W−W

W−N

2.1920(3) 2.193 2.217 2.1990 2.194 2.22 2.222 2.247 2.1617(4) 2.149 2.183 2.180 2.158 2.192 2.1956(5) 2.175 2.208

2.076[9] 2.108 2.123 2.0971 2.112 2.126 2.127 2.145 2.128[5] 2.169 2.175 2.200 2.16 2.166 2.131[3] 2.16 2.168

Figure 4. Depiction of the interaction between the W−W σ* bond and the CO π* orbital in 1 (left). The π-accepting ability of the carbonyl groups significantly lowers the energy of the σ* orbital. (right).

and a σ2π2 electronic configuration.29c The reason for the smaller Wiberg value in I is apparent upon comparison of the electronic configurations. Even though in both species the formal bond order is two, both compounds have two electrons in a π orbital; in I, the δ orbital is occupied, while in W2(μO2)(TMhpp)2Cl4, the electrons occupy the σ orbital. Thus, the shorter W−W bond length in W2(μ-O2)(TMhpp)2Cl4 than that in 1 is in agreement with the known fact that a σ bond is generally stronger than a δ bond. Finally, the Wiberg bond index of 1.08 for the noncyclic guanidinate II (Figure S2) resembles that in I, but that of 1.54 for the formamidinate III (Figure S3) is larger than that for I. For comparison, they are smaller than that of 1.61 for the dioxo-bridged W2(μO2)(TMhpp)2Cl4.

a

When there is more than one multiple bond of the same type, the average is given and the standard deviations are in square brackets.

■

CONCLUDING REMARKS Two new compounds (1 and 2·3CH2Cl2) are reported; the former is an ESBO species with two unsymmetrically bridging carbonyl groups, and the latter has a triple-bonded W26+ core in which each dimetal unit is spanned by four bridging bicyclic guanidinate ligands as well as two hexafluorophosphate anions located along the metal−metal axis and nestled within methylene groups with the aid of a series of weak C−H···F interactions. The closest fluorine atoms are positioned along the W−W bond axis at long W···F separations of 2.786(4) and 2.810(4) Å. Compound 2 has a formal bond order of 3 and a σ2π4 electronic configuration with S = 0. Importantly, the average W···F distances of 2.798[6] Å are similar to the W···Cl distances in the chloro analogue (2.85−3.06 Å). Because the W−W distance is much shorter in the PF6 species, it is clear that the electron pairs in the F atoms are not as available as those in the chlorine atoms in W2(hpp)4Cl2 to interact with the dimetal unit because the fluorine atoms in the PF6 anions are engaged in pF → dP bonding. In contrast to those in the chloro analogues, the electron pairs are free to interact with the ditungsten core. Perhaps the most important lesson is that interactions of lone electron pairs with dimetal units can have important effects even at long distances. Therefore, this should serve to caution against neglecting such long interactions as being unimportant in determining metal-to-metal distances and possibly in other types of compounds.

■

EXPERIMENTAL SECTION

General Syntheses. Syntheses were performed using oven-dried glassware under a nitrogen atmosphere, unless otherwise noted. The ligand precursor HTMhpp was synthesized as reported. 18 W2(TMhpp)4Cl2 was prepared by heating solutions of 1 following literature procedures.19 W(CO)6 and TlPF6 were obtained from commercial sources. Hexanes, o-dichlorobenzene, toluene, and dichloromethane were acquired from Aldrich; solvents were purified using a Glass Contour solvent system. Infrared spectra were recorded with a PerkinElmer 16PC FT IR spectrophotometer with KBr pellets.

Figure 3. Selected MO orbitals with a 0.05 contour surface for model complex I, derived from calculations using the functional BH&HLYP. Front view (left) and top view (right). Positive and negative values are represented as red and yellow surfaces, respectively. Values on the left are given as the percent of electron density divided among the two W ions and the bridged carbonyl groups, respectively.

G

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 5. Selected frontier orbital assignment and energies for models I−III, obtained from DFT calculations using the functional BH&HLYP.

Figure 6. Selected frontier orbital assignment and energies for models IV−VI obtained from DFT calculations using the functional BH&HLYP. 1 H NMR spectra were recorded on a Unity Plus 300 NMR spectrometer, using residual solvent peaks to reference chemical shifts (δ). Elemental analyses were performed by Robertson Microlit Laboratories, Inc. (Madison, NJ, USA) on crystalline material that had been previously placed under vacuum. Synthesis of W2(μ-CO)2(μ-TMhpp)2(η2-TMhpp)2, 1. A mixture of 0.240 g (0.682 mmol) of W(CO)6 and 0.300 g (1.53 mmol) of HTMhpp was placed in an oven-dried 100 mL Schlenk flask equipped with a stir bar and filled with nitrogen. After addition of 15 mL of dried and oxygen-free o-dichlorobenzene, the flask was fitted with a previously oven-dried water-cooled coldfinger. The pale yellow reaction mixture was heated under nitrogen to 180 °C for 6−8 h. During the reflux period, the color of the reaction mixture changed progressively to deep yellow, orange, and red and then to green. The solvent was removed under vacuum at 70 °C. The greenish solid was extracted with 50 mL of toluene, and the mixture was filtered under nitrogen using an oven-dried fritted-glass packed with Celite. The solvent from the green solution was removed under vacuum and the solid washed with cold hexanes (to remove a small amount of stopcock

grease). The solid was dissolved in 10 mL of dichloromethane, and the solution was carefully layered with 40 mL of hexanes using a 70 mL Schlenk tube equipped with a glass cap wrapped with Parafilm. After 1 week, dark green, block-shaped crystals suitable for X-ray diffraction were obtained. Yield: 0.368 g (90%). Anal. Calcd for W2C46H80N12O2: C, 46.01; H, 6.71; N, 13.99%. Found: C, 45.67; H, 6.99; N, 13.58%. 1H NMR in C6D6 ppm: 3.36 (s), 2.81 (s), 2.67 (s), 2.34 (s), 1.15 (s), and 0.85 (s). The integration was 1:1:1:1:3:3, respectively. IR (KBr plates, cm−1): 1725 (ν(μ‑CO)). Synthesis of [W 2 (TMhpp) 4 ](PF 6 ) 2 , 2. A mixture of W2(TMhpp)4Cl2 (0.100 g, 0.082 mmol) and TlPF6 (0.057 g, 0.164 mmol) was placed in an oven-dried 100 mL Schlenk flask equipped with a stir bar and filled with nitrogen. After addition of 15 mL of dichloromethane, the solution was stirred overnight. The reaction mixture was then filtered, and the volume of the solution was reduced to about half. The remaining solution was layered with hexanes. After 10 days, green block-shaped crystals suitable for X-ray diffraction formed. They were separated by filtration. Yield after placing the material under vacuum: 0.114 g (97%). Anal. Calcd for H

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 7. Selected frontier orbital assignment and energies for models VII−IX, obtained from DFT calculations using the functional BH&HLYP. Pbca with eight molecules per unit cell. The asymmetric unit in 2· 3CH2Cl2 was composed of the cation plus two hexafluorophosphate anions. There was some crystallographic disorder. In 1, the methyl and methylene groups attached to the quaternary carbon atoms of the chelated TMhpp ligands were disordered, a common occurrence in these types of compounds because of the envelope-type structure.42 This disorder was modeled using an additional free variable, resulting in two site occupancy factors, with a major component of 87.3% and a minor component of 12.7%. Similar treatment for 2·3CH2Cl2 showed three disordered groups of atoms. A ring of one of the four TMhpp ligands resulted in a major component of 66.85% and minor component of 33.15%; four fluorine atoms of one PF6 anion had a major component of 58.58% and a minor component of 41.4%. Finally, one of the interstitial methylene chlorides had a major component of 78.24% and a minor component of 21.46%. Selected bond lengths and angles are shown in Table 1. Data collection and refinement parameters are summarized in Table 6. Computational Details. Density functional theory (DFT)43 calculations using isolated molecules, as if they were in the gas phase, were performed with the hybrid Becke three-parameter exchange functional and the Lee−Yang−Parr nonlocal correlation functionals B3LYP44 and BH&HLYP45 implemented in the Gaussian

Figure 8. Selected Wiberg bond indices obtained from NBO analysis using the functional BH&HLYP. Each numerical atom-centered orbital is based on model I.

Table 6. Crystallographic Data

W2C44H80F12N12P2: C, 36.83; H, 5.62; N, 11.71%. Found: C, 37.12; H, 5.30; N, 11.39%. 1H NMR in CDCl3 ppm: 3.16 (s), 3.04 (s), and 1.07 (s). The integration was 1:1:3, respectively. X-ray Structure Determination. Crystals of 1 and 2·3CH2Cl2 were coated with Paratone oil and mounted on a nylon cryoloop affixed to a goniometer head. Data were collected at 213 K for 1 and at 110 K for 2·3CH2Cl2 on a Bruker APEX-II 1000 CCD area detector system using omega scans of 0.3 deg/frame, with exposures of 30 s/ frame. Cell parameters were determined using the SMART software suite.38 Data reduction and integration were performed with the software SAINT.39 Absorption corrections were applied using the program SADABS.40 The positions of the metal atoms were found via direct methods using the program SHELXTL.41 Subsequent cycles of least-squares refinement followed by difference Fourier syntheses revealed the positions of the remaining non-hydrogen atoms. Hydrogen atoms were added in idealized positions. All hydrogen atoms were included in the calculation of the structure factors. All nonhydrogen atoms were refined with anisotropic displacement parameters. Refinement for 1 was carried out in the monoclinic space group P21/c possessing two molecules per unit cell, while refinement of 2·3CH2Cl2 was done in the orthorhombic space group

compound

1

2·3CH2Cl2

chemical formula fw space group a (Å) b (Å) c (Å) β (deg) V (Å3) Z dcalcd (g·cm−3) μ (mm−1) T (K) R1a (wR2b)

W2C46H80N12O2 1200.92 P21/c 9.773(4) 18.025(8) 14.571(6) 98.688(6) 2537(2) 2 1.572 4.578 213(2) 0.0320 (0.0795)

W2C47H86Cl6F12N12P2 1689.62 Pbca 20.097(3) 19.113(3) 34.639(6) 90 13305(4) 8 1.687 3.820 110(2) 0.0499 (0.1247)

R1 = [∑w(Fo − Fc)2/∑wFo2]1/2. bwR2 = [∑[w(Fo2 − Fc2)2]/ ∑w(Fo2)2]1/2, w = 1/[σ2(Fo2) + (aP)2 + bP], where P = [max(Fo2,0) + 2(Fc2)]/3. a

I

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry 03 (revision E.01) program suite.46 All model compounds were computed using SDD for W with an effective core potential and 6311G** for N, O, C, and H.47 Calculations were carried out without symmetry constraints using the nine models depicted in Scheme 4. Geometry optimizations of the model compound W2(μ-CO)2(μhpp)2(η2-hpp)2 (I), [W2(hpp)4]2+ (IV), and W2(hpp)4 (VII) were done using the parameters modified from the crystal structures of 1, 2, and W2(hpp)417a as starting points, respectively. Models III and IX were derived by using hydrogen atoms in place of the p-tolyl groups from the crystal structures of W2(μ-CO)2(μ-DAniF)2(η2-DAniF)2 and W2(μ-DAniF)4, respectively.22 The geometries of models II, V, VI, and VIII used for the optimizations were directly modified from the models of I, IV, IX, and VII, respectively. Selected molecular orbitals for the model of 1 are in Figure 3, while a molecular orbital interaction diagram is in Figure 4. The general agreement between the calculated, published, and experimental geometric data, shown in Tables 2−5, suggests that such a simplification is reasonable and appropriate. Selected frontier orbital assignment and energies for all nine models I−IX are shown in Figures 5−7. Plots of molecular orbitals were generated from Gaussian cube files using Visual Molecular Dynamics (VMD 1.8.6)48 and rendered using POV-Ray v3.649 for Windows. NBO analysis was performed by using the method built in the Gaussian 03 program.46,50 NBO orbitals were generated from Gaussian fchk files using Multiwfn v3.2.51 Selected bonding and antibonding orbitals in the W2(μ-CO)24+ core for the geometry-optimized model 1 from NBO analysis are shown in Figure S1. Wiberg bond indices for model I from NBO analysis are shown in Figure 8, while those for models II−IX are in Figures S2−S9. All computations were carried out on a Dirac SGI Altix 350 32-processor computer located at the University of the Pacific.

■

Chem. 2004, 43, 7564−7566. (g) Coles, M. P.; Hitchcock, P. B. Eur. J. Inorg. Chem. 2004, 2004, 2662−2672. (h) Khalaf, M. S.; Coles, M. P.; Hitchcock, P. B. Dalton Trans. 2008, 4288−4295. (i) Coles, M. P. Chem. Commun. 2009, 3659−3676. (j) Irwin, M. D.; Abdou, H. E.; Mohamed, A. A.; Fackler, J. P., Jr. Chem. Commun. 2003, 2882−2883. (k) Feil, F.; Harder, S. Eur. J. Inorg. Chem. 2005, 2005, 4438−4443. (l) Wilder, C. B.; Reitfort, L. L.; Abboud, K. A.; McElwee-White, L. Inorg. Chem. 2006, 45, 263−268. (m) Rische, D.; Baunemann, A.; Winter, M.; Fischer, R. A. Inorg. Chem. 2006, 45, 269−277. (n) Edelmann, F. T. Chem. Soc. Rev. 2009, 38, 2253−2268. (o) Chiarella, G. M.; Melgarejo, D. Y.; Rozanski, A.; Hempte, P.; Perez, L. M.; Reber, C.; Fackler, J. P., Jr. Chem. Commun. 2010, 46, 136−138. (p) Lee, R.; Yang, Y. Y.; Tan, G. K.; Tan, C.-H.; Huang, K.W. Dalton Trans. 2010, 39, 723−725. (q) Heitmann, D.; Jones, C.; Mills, D. P.; Stasch, A. Dalton Trans. 2010, 39, 1877−1882. (r) Dong, H.; Meng, Q.; Chen, B.-Z.; Wu, Y.-B. J. Organomet. Chem. 2012, 717, 108−115. (s) Jones, C. Coord. Chem. Rev. 2010, 254, 1273−1289. (2) There is also a recent topical issue of Australian Journal of Chemistry edited by Martyn Coles dedicated to the chemistry of compounds derived from the guanidine functionality; see: Tan, C. H.; Coles, M. Aust. J. Chem. 2014, 67, 963−1137. (3) (a) Deutsch, J.; Eckelt, R.; Köckritz, A.; Martin, A. Tetrahedron 2009, 65, 10365−10369. (b) Collins, S. Coord. Chem. Rev. 2011, 255, 118−138. (4) Cotton, F. A.; Matonic, J. H.; Murillo, C. A. J. Am. Chem. Soc. 1997, 119, 7889−7890. (5) For examples, see: (a) Harisomayajula, N. V. S.; Nair, A. K.; Tsai, Y.-C. Chem. Commun. 2014, 50, 3391−3412. (b) Noor, A.; Glatz, G.; Müller, R.; Kaupp, M.; Demeshko, S.; Kempe, R. Z. Anorg. Allg. Chem. 2009, 635, 1149−1152. (c) Noor, A.; Bauer, T.; Todorova, T. K.; Weber, L.; Gagliardi, L.; Kempe, R. Chem. - Eur. J. 2013, 19, 9825− 9832. (6) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Timmons, D. J.; Wilkinson, C. C. J. Am. Chem. Soc. 2002, 124, 9249−9256. (7) (a) Clérac, R.; Cotton, F. A.; Daniels, L. M.; Donahue, J. P.; Murillo, C. A.; Timmons, D. J. Inorg. Chem. 2000, 39, 2581−2584. (b) Cotton, F. A.; Donahue, J. P.; Gruhn, N. E.; Lichtenberger, D. L.; Murillo, C. A.; Timmons, D. J.; Van Dorn, L. O.; Villagrán, D.; Wang, X. Inorg. Chem. 2006, 45, 201−213. (8) Chiarella, G. M.; Cotton, F. A.; Murillo, C. A. Chem. Commun. 2011, 47, 8940−8942. (9) Cotton, F. A.; Murillo, C. A.; Timmons, D. J. Chem. Commun. 1999, 1427−1428. (10) Cotton, F. A.; Gu, J.; Murillo, C. A.; Timmons, D. J. J. Am. Chem. Soc. 1998, 120, 13280−13281. (11) Cotton, F. A.; Dalal, N. S.; Hillard, E. A.; Huang, P.; Murillo, C. A.; Ramsey, C. M. Inorg. Chem. 2003, 42, 1388−1390. (12) Cotton, F. A.; Huang, P.; Murillo, C. A.; Timmons, D. J. Inorg. Chem. Commun. 2002, 5, 501−504. (13) (a) Berry, J. F.; Cotton, F. A.; Huang, P.; Murillo, C. A. Dalton Trans. 2003, 1218−1219. (b) Cotton, F. A.; Dalal, N. S.; Huang, P.; Ibragimov, S. A.; Murillo, C. A.; Piccoli, P. M. B.; Ramsey, C. M.; Schultz, A. J.; Wang, X.; Zhao, Q. Inorg. Chem. 2007, 46, 1718−1726. (14) For example, see: (a) Cotton, F. A.; Dalal, N. S.; Huang, P.; Murillo, C. A.; Stowe, A. C.; Wang, X. Inorg. Chem. 2003, 42, 670− 672. (b) Cotton, F. A.; Chiarella, G. M.; Dalal, N. S.; Murillo, C. A.; Wang, Z.; Young, M. D. Inorg. Chem. 2010, 49, 319−324. (15) Berry, J. F.; Cotton, F. A.; Huang, P.; Murillo, C. A.; Wang, X. Dalton Trans. 2005, 3713−3715. (16) Dalal, N. S.; Murillo, C. A. Dalton Trans. 2014, 43, 8565−8576. (17) (a) Cotton, F. A.; Gruhn, N. E.; Gu, J.; Huang, P.; Lichtenberger, D. L.; Murillo, C. A.; Van Dorn, L. O.; Wilkinson, C. C. Science 2002, 298, 1971−1974. (b) Cotton, F. A.; Donahue, J. P.; Lichtenberger, D. L.; Murillo, C. A.; Villagrán, D. J. Am. Chem. Soc. 2005, 127, 10808−10809. (c) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Timmons, D. J.; Wilkinson, C. C. J. Am. Chem. Soc. 2002, 124, 9249−9256. (d) Cotton, F. A.; Durivage, J. C.; Gruhn, N. E.; Lichtenberger, D. L.; Murillo, C. A.; Van Dorn, L. O.; Wilkinson, C. C. J. Phys. Chem. B 2006, 110, 19793−19798. (e) Cotton, F. A.; Murillo,

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b01728. X-ray crystallographic data for 1 and 2·3CH2Cl2 in standard CIF format (CIF) Depiction of selected NBO orbitals in the W2(μ-CO)24+ core of model I, selected Wiberg bond indices for models II−IX, 1H NMR spectra of 1 and 2, and the IR spectrum of 1 (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (C.A.M.). *E-mail: qzhao@pacific.edu (Q.Z.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Robert A. Welch Foundation and Texas A&M University. C.A.M. also thanks the National Science Foundation (IR/D support). C.F. and Q.Z. are grateful to the Department of Chemistry at University of the Pacific for financial support.

■

REFERENCES

(1) See, for example: (a) Foley, S. R.; Yap, G. P. A.; Richeson, D. S. Polyhedron 2002, 21, 619−627. (b) Soria, D. B.; Grundy, J.; Coles, M. P.; Hitchcock, P. B. J. Organomet. Chem. 2005, 690, 2278−2284. (c) Coles, M. P.; Hitchcock, P. B. Organometallics 2003, 22, 5201− 5211. (d) Coles, M. P.; Hitchcock, P. B. Dalton Trans. 2001, 1169− 1171. (e) Coles, M. P.; Hitchcock, P. B. Inorg. Chim. Acta 2004, 357, 4330−4334. (f) Oakley, S. H.; Coles, M. P.; Hitchcock, P. B. Inorg. J

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry C. A.; Wang, X.; Wilkinson, C. C. Dalton Trans. 2006, 4623−4631. (f) Cotton, F. A.; Murillo, C. A.; Wang, X.; Wilkinson, C. C. Dalton Trans. 2007, 3943−3951. (g) Murillo, C. A. Aust. J. Chem. 2014, 67, 972−979. (h) Falvello, L. R.; Foxman, B. R.; Murillo, C. A. Inorg. Chem. 2014, 53, 9441−9456. (18) Chiarella, G. M.; Cotton, F. A.; Ibragimov, S. A.; Murillo, C. A.; Wilkinson, C. C.; Young, M. D. Polyhedron 2013, 58, 7−12. (19) Chiarella, G. M.; Cotton, F. A.; Durivage, J. C.; Lichtenberger, D. L.; Murillo, C. A. J. Am. Chem. Soc. 2013, 135, 17889−17896. (20) Alternatively, a sealed Parr bomb can be used to replace the mercury bubbler. (21) The compound 2,6-dichlorobiphenyl was not isolated from this reaction, but it was characterized during the preparation of W2(hpp)4Cl2. See ref 7b. In that reaction, an additional intermediate, W(CO)4(Hhpp)2, was also obtained and fully characterized. (22) Cotton, F. A.; Donahue, J. P.; Hall, M. B.; Murillo, C. A.; Villagrán, D. Inorg. Chem. 2004, 43, 6954−6964. (23) For example, see: Advanced Inorganic Chemistry, 6th ed.; Cotton, F. A., Wilkinson, G., Murillo, C. A., Bochmann, M. Eds.; John Wiley & Sons, Inc.: New York, 1999; pp 638−639. (24) For excellent background and historical perspective of carbonyl binding to species with metal atoms, see: (a) Li, H.; Feng, H.; Sun, W.; King, R. B.; Schaefer, H. F., III Inorg. Chem. 2013, 52, 6893−6904 and references therein. (b) Calderazzo, F. In Encyclopedia of Inorganic Chemistry, 2nd ed.; King, R. B., Ed.; Wiley: New York, 2005; pp 764− 768. (25) Xu, L.; Li, Q.-S.; King, R. B. Polyhedron 2012, 38, 44−49. (26) Note that for each TMhpp there are two independent CH2 groups, for a total of 4 hydrogen atoms each and 12 hydrogen atoms for the methyl groups. This gives a relative ratio of 1:1:3 (thus an overall ratio of 1:1:1:1:3:3). Because there are no adjacent carbon atoms with hydrogen atoms bound to them, each signal is a singlet. (27) Strictly speaking, because of the semibridging carbonyls, the methylene groups in the hpp groups are not equivalent. However, one should not expect a very large difference in their spectral position. Another alternative is that the CO ligands are fluxional and thus the NMR spectra show average signals. (28) See p 1034 in ref 23. (29) (a) Poli, R.; Torralba, R. C. Inorg. Chim. Acta 1993, 212, 123− 134. (b) Cotton, F. A. Polyhedron 1987, 6, 667−677. (c) Chiarella, G. M.; Cotton, F. A.; Murillo, C. A.; Zhao, Q. Inorg. Chem. 2014, 53, 2288−2295. (30) For example, see: (a) Williamson, R. L.; Low, A. A.; Hall, M. B.; Guest, M. F. In Metal-Ligand Interactions: from Atoms, to Clusters, to Surfaces; Salahub, D. R., Russo, N., Eds; Kluwer Academic Publishers: The Netherlands, 1992; pp 155−173. (b) Sargent, A. L.; Hall, M. H. Polyhedron 1990, 9, 1799−1808. (c) Breed, L. W.; Wiley, J. C. J. Organomet. Chem. 1975, 102, 29−39. (d) Miller, T. F., III; Strout, D. L.; Hall, M. H. Organometallics 1998, 17, 4164−4168. (31) Cotton, F. A., Murillo, C. A., Walton, R. A., Eds. Multiple Bonds between Metal Atoms, 3rd Ed.; Springer Science and Business Media, Inc.: New York, 2005. (32) Cotton, F. A.; Mott, G. N.; Schrock, R. R.; Sturgeoff, L. G. J. Am. Chem. Soc. 1982, 104, 6781−6782. (33) Chisholm, M. H.; Cotton, F. A.; Extine, M.; Stults, B. R. J. Am. Chem. Soc. 1976, 98, 4477−4485. (34) Chisholm, M. H.; Gallucci, J.; Hadad, C. M.; Huffman, J. C.; Wilson, P. J. J. Am. Chem. Soc. 2003, 125, 16040−16049. (35) Chiarella, G. M.; Cotton, F. A.; Murillo, C. A.; Ventura, K.; Villagrán, D.; Wang, X. J. Am. Chem. Soc. 2014, 136, 9580−9589. (36) Pastor, M. B.; Zhao, Q. Inorg. Chim. Acta 2015, 424, 308−315. (37) (a) Hu, P.; Luo, Q.; Li, Q.-S.; Xie, Y.; King, R. B.; Schaefer, H. F., III Inorg. Chem. 2014, 53, 5300−5310. (b) Wang, H.; Sun, Z.; King, R. B.; Schaefer, H. F., III; Xie, Y. Inorg. Chem. 2011, 50, 9256−9265. (c) Sun, Z.; Schaefer, H. F., III; Xie, Y.; Liu, Y.; Zhong, R. Mol. Phys. 2013, 111, 2523−2535. (d) Wang, H.; Wang, H.; King, R. B.; Schaefer, H. F., III J. Comput. Chem. 2016, 37, 250−260. (e) Mednikov, E. G.; Ivanov, S. A.; Dahl, L. F. J. Organomet. Chem. 2015, 792, 229−235.

(f) Harper, L.; Shoaf, A. L.; Bayse, C. A. ChemPhysChem 2015, 16, 3886−3892. (38) SMART for Windows NT, version 5.618; Bruker Advanced X-ray Solutions, Inc.: Madison, WI, 2001. (39) SAINT. Data Reduction Software, version 6.36A; Bruker Advanced X-ray Solutions, Inc.: Madison, WI, 2001. (40) SADABS. Area Detector Absorption and other Corrections Software, version 2.05; Bruker Advanced X-ray Solutions, Inc.: Madison, WI, 2001. (41) Sheldrick, G. M. SHELXTL, version 6.12; Bruker Advanced Xray Solutions, Inc.: Madison, WI, 2002. (42) Cotton, F. A.; Murillo, C. A.; Wang, X.; Wilkinson, C. C. Inorg. Chim. Acta 2003, 351, 191−200. (43) (a) Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864−B871. (b) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and Molecules; Oxford University Press: Oxford, U.K., 1989. (44) (a) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 1372−1377. (c) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (45) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (b) Freeman, F.; Fang, C.; Hoang, D. V.; Trinh, K. M. Int. J. Quantum Chem. 2005, 105, 416−428. (46) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision E.01; Gaussian, Inc.: Wallingford, CT, 2004. (47) Woon, D. E.; Dunning, T. H., Jr. J. Chem. Phys. 1993, 98, 1358− 1371. (48) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33−38. (49) Persistence of Vision (TM) Raytracer. Persistence of Vision Pty. Ltd.: Williamstown, Victoria, Australia http://www.povray.org./ (2004). (50) For examples of the appropriateness of using NBO on systems containing metal−metal bonds, see: (a) Fang, W. J.; He, Q.; Tan, Z. F.; Liu, C. Y.; Lu, X.; Murillo, C. A. Chem. - Eur. J. 2011, 17, 10288− 10296. (b) Tanabe, M.; Omine, S.; Ishikawa, N.; Osakada, K.; Hayashi, Y.; Kawauchi, S. Angew. Chem., Int. Ed. 2015, 54, 2679−2721. (51) Lu, T.; Chen, F. J. Comput. Chem. 2012, 33, 580−592.

K

DOI: 10.1021/acs.inorgchem.6b01728 Inorg. Chem. XXXX, XXX, XXX−XXX