Factors Controlling Stability and Reactivity of Dimeric Pd(II

Dec 21, 2015 - As a result, the identity of the active catalyst in various widely utilized cross-coupling and C–H functionalization reactions with P...
6 downloads 0 Views 2MB Size
Research Article pubs.acs.org/acscatalysis

Factors Controlling Stability and Reactivity of Dimeric Pd(II) Complexes in C−H Functionalization Catalysis Brandon E. Haines,† John F. Berry,‡ Jin-Quan Yu,§ and Djamaladdin G. Musaev*,† †

Cherry L. Emerson Center for Scientific Computation, Emory University, 1515 Dickey Drive, Atlanta, Georgia 30322, United States Department of Chemistry, University of Wisconsin − Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States § Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, United States ‡

S Supporting Information *

ABSTRACT: We have systematically explored the factors impacting the stability and reactivity of the [Pd(O2CR)2]2 and [Pd(OAc)(DG-Ar)]2 dimers in Pd(OAc)2-catalyzed and directing group (DG)-assisted C−H functionalization using density functional theory (DFT) calculations. It was shown that the palladium acetate dimer stability (vs the monomer) predominantly arises from interactions between the paddlewheel ligands and the Pd centers. R-substitution in Pd(O2CR)2 leading to an increase (or decrease) in the electron density of the ligand orbitals polarizes the Pd−ligand interaction and weakens (or strengthens) the stability of the [Pd(O2CR)2]2 dimer relative to the monomer. The nature of the substrate/ligand-Pd interaction is shown to be another factor contributing to the nuclearity and reactivity of the Pd-acetates in directing group-assisted C−H functionalization. For the [Pd(OAc)(DG-Ar)]2 dimer, we have found that the major interactions contributing to the stability of the dimer are the bridging acetate interactions and interactions between the substrates, such as π−π stacking. Our data reveal that, starting from the dinuclear [Pd(OAc)2]2 complex, pathways involving either monomerization of [Pd(OAc)2]2 or C−H activation by the dimer can compete with each other, and in general, dinuclear complexes require a higher C−H activation barrier than mononuclear complexes. Even if C−H activation of the substrate is initiated by the Pd(OAc)2 monomer, involvement of the dimeric [Pd(OAc)(DG-Ar)]2 complex in C−H functionalization, especially for electrophiles with small C−X bond formation barriers, cannot be excluded. KEYWORDS: C−H functionalization, palladium dimer, directing group, active catalyst, density functional theory

I. INTRODUCTION Palladium continues to be one of the most widely utilized transition metals in catalysis.1 Its versatile reactivity is mostly attributed to its ability to (a) form complexes in the (0), (I), (II), (III), and (IV) oxidation states and, consequently, can be easily involved in one- and two-electron redox processes,2 (b) act as a redox-neutral transition metal center in critical catalytic processes,3 (c) facilitate substrate binding and release through ligand exchange rates that are significantly faster than those of other second- and third-row transition metals,4 and (d) form catalytically active high-order aggregates (including nanoscale Pd clusters) with specific sizes.5 These complex physicochemical properties of Pd not only make it catalytically rich but also make the identification and characterization of the true catalytically active species extremely difficult in many important Pd-catalyzed transformations. As a result, the identity of the active catalyst in various widely utilized cross-coupling and C− H functionalization reactions with Pd(0) and Pd(II) precatalysts still remains the subject of active debate.6 Growing evidence indicates that the widely accepted view on mono-Pd complexes being an active species needs to be re-evaluated,7 © 2015 American Chemical Society

particularly because a catalytic mixture can contain multiple high-order Pd species, including dinuclear Pd complexes.8 The presence of such high-order Pd species in a catalytic mixture could play several critical roles, including being (a) the restingstate of the catalyst, i.e., a source of catalytic active species,9 and/or (b) a direct participant in the important steps of the reaction. Limited experiments show that the stability and reactivity of these high order Pd species is a function of several factors, such as solvent, substrates, auxiliary ligands, bases, and cocatalysts (including light).10 For example, Schoenebeck and co-workers have demonstrated that the [(PtBu3)PdBr]2 dimer reacts with aryl iodides as a stable dinuclear species, whereas it reacts with aryl bromides and chlorides as a mononuclear Pd(0) complex.11 They show that the reactivity of the [(PtBu3)PdBr]2 dimer complex depends on the nature of the aryl halide substrate as well as the precise reaction conditions, including additives and Received: October 30, 2015 Revised: December 10, 2015 Published: December 21, 2015 829

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

Figure 1. Schematic presentation of the dimerization, cyclopalladation, and oxidation processes investigated in this paper. core potential for Pd, Cs, and I.18 Because of the importance of dispersion for noncovalent interactions,19 the reported energies of every structure were calculated (at their M06/BS1 optimized geometries) with empirical dispersion-corrected DFT20 at the B3LYP-D3/[6-311+G(d,p)] + SDD (Pd, Cs, I)] level of theory (B3LYP-D3/BS2) with the corresponding SDD effective core potentials for Pd, Cs, and I.21 All presented geometries and energies incorporate solvent effects calculated at the self-consistent reaction field polarizable continuum model (IEF-PCM)22 level with N,Ndimethylformamide (DMF) as solvent. Frequency calculations were performed for all structures, at the M06/BS1 level of theory, to confirm identity of the located minima and transition states and to calculate enthalpy and entropy corrections to their energies under standard conditions (1 atm and 298.15 K). The translational component of the entropy correction was further corrected using the free volume theory method of Whitesides in DMF.23 NBO analysis was performed for selected structures using the NBO program (version 3.1), as implemented in G09.24 Donor− acceptor interactions are analyzed with the second-order perturbation of the natural bond orbitals. This analysis provides a quantitative measure (E(2)i→j) of the interaction between an occupied donor NBO (i) and an empty acceptor NBO (j).

solvent. In studies of nondirected arene C−H acetoxylation, Cook and Sanford have shown that the optimal ratio of catalyst to ligand [1:1, Pd(OAc)2:pyridine] produces a dimeric Pd(II)− Pd(II) species that easily converts to the coordinatively unsaturated monomeric Pd(II) active catalyst.12 Recently, Ritter13 and Sanford14 each have reported C−H functionalization reactions in which dinuclear Pd(III) intermediates are proposed in place of more commonly proposed mononuclear Pd(IV) intermediates.15 In our recent study on the mechanism of Pd(II)-catalyzed C−H iodination with I2, we found that the barrier for the redox-neutral electrophilic cleavage pathway is lowered in the dinuclear Pd(II)−Pd(II) complex relative to the analogous process in the mononuclear Pd(II) complex. Thus, metal−metal synergy can benefit either redox or redox-neutral reactivity of dinuclear Pd(II) complexes in C−H functionalization.16 The aforementioned examples demonstrate the complexity of determining the precise nature of the active Pd species in a catalytic mixture. A detailed understanding of these aspects requires more comprehensive investigation. In particular, we set out to systematically explore the factors impacting the stability and reactivity of the [Pd(O2CR)2]2 and [Pd(OAc)(DG-Ar)]2 dimers in Pd(OAc)2-catalyzed and directing group (DG)assisted C−H functionalization. For the [Pd(O2CR)2]2 dimer, we determine the effect of substitution of the paddlewheel bridging ligand on the stability of the dimer and the strength of metal−metal interactions. In the case of the [Pd(OAc)(DGAr)]2 dimer, we elucidate the roles of commonly employed directing groups and palladacycle ring size on the stability of the dimer complex and palladium acetate’s reactivity toward C−H activation (see Figure 1). These studies are expected to greatly advance our knowledge of the role of dinuclear Pd complexes in C−H functionalization reactions and will contribute to our ability to design more efficient reactions with mild experimental conditions and develop analogous processes with earth-abundant transition metal catalysts.

III. RESULTS AND DISCUSSION III.A.1. Structure and Bonding in the [Pd(OAc)2]2 Dimer. As established in the literature, the most stable form of Pd-acetate in the absence of coordinating ligands is the trimeric complex [Pd(OAc)2]3, of which the crystal structure has been carefully analyzed,25 along with structures of its derivatives, such as [Pd(TFA)2]3.26 It was shown that each of the three Pd atoms of [Pd(OAc)2]3 is in a square planar environment. The calculated structure of [Pd(OAc)2]3, in this paper, is in good agreement with the reported crystal structure despite expected structural differences due to solvation and crystal packing (see Figure 2).25 Furthermore, our computation echoes experimental data25 and shows that the trimeric complex [Pd(OAc)2]3 is stable relative to the dissociation limits of [Pd(OAc)2]2 + Pd(OAc)2 (ΔG/ΔH = 34.9/44.6 kcal/ mol) and 3x Pd(OAc)2 (ΔG/ΔH = 59.3/81.7 kcal/mol), respectively (see the Supporting Information (SI)). Thus, the trimer/dimer equilibrium 2/3 [Pd(OAc)2]3 ⇄ [Pd(OAc)2]2 is unfavorable by 15.1/17.4 kcal/mol, a significant amount of energy but small enough to be overcome by effects of ligation or solvation. For example, substitution of the acetate ligands for

II. COMPUTATIONAL DETAILS Calculations were performed with the Gaussian 09 (G09) program.17 All reported structures were fully optimized (without geometry constraint) at the M06/[6-31G(d,p) + Lanl2dz (Pd, Cs, and I)] level of theory (M06/BS1) with the corresponding Hay−Wadt effective 830

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

isomers of the [Pd(OAc)2]2 dimer with four (1), two (2), and zero (6) bridging acetate ligands in more detail (Figure 4). The calculated dimerization energies (relative to the Pd(OAc)2 monomer) are found to be ΔGdimer/ΔHdimer = −24.3/−37.1, −15.8/−27.8, and 3.0/−10.0 kcal/mol for isomers 1, 2, and 6, respectively. This finding indicates that each bridging ligand increases the stability of the dimer, but the correlation between the number of bridging ligands and dimer stability is not linear. Comparison of the calculated free energies for structures 1 and 6 clearly shows that the bridging ligands account for all of the free energy of dimerization. However, the excess enthalpy of dimerization for structure 6 indicates an energetically favorable interaction between the metal centers (ΔH = −10.0 kcal/mol). This conclusion is also supported by total energy scans of the Pd−Pd distance in 1 and 6, which show welldefined minima for both structures (see SI). Previously, Berry and co-workers33 and Bercaw, Gray, and co-workers34 thoroughly analyzed the Pd−Pd interaction in various dinuclear Pd(II) complexes. As concluded in these papers, the frontier orbital picture given in Figure 5a for the idealized Pd(II)−Pd(II) interaction shows that there should not be formal bonding between the Pd(II) atoms. However, mixing of the filled Pd(II)−Pd(II) dσ* antibonding orbital with the unoccupied Pd s and pz orbitals creates a hybridized orbital with less antibonding character (Figure 5b) and introduces some degree of interaction between the metal centers. The degree of Pd dσ*/s/p mixing, and therefore the degree of Pd− Pd bonding, will depend critically on the Pd−Pd distance as well as the electronic nature of the bridging ligands. The influence of the bridging ligands occurs due to σ overlap of the ligand lone pairs with the Pd dz2 orbitals. We may therefore anticipate a stronger ligand-Pd interaction with more electronrich equatorial ligands, which will push the Pd2 dσ* orbital higher in energy, facilitating mixing with the Pd s/p orbitals and thus leading to a stronger Pd−Pd interaction. The residual Pd−Pd bonding interactions can be assessed computationally by the calculated Wiberg bond index (WBI), which correlates roughly with bond orders that one would expect from a molecular orbital analysis.35 The performed NBO analysis and the calculated Wiberg bond indexes (WBI) are in agreement with the bonding picture presented above.14b,15b,34,36 Indeed, as shown in Table 1, the most important Pd−Pd orbital interaction (measured by the second order perturbation energy, E(2)i→j) is the interaction of the Pd2 σ*(dz2) with Pd2 σ(pz), which is made possible by the orbital hybridization described above (see Figure 5 and SI). As shown in Figure 4, the value of this orbital interaction energy increases dramatically with the number of bridging acetate ligands (6 < 2 < 1), indicating its strong dependence on the Pd−Pd distance. Consistent with this conclusion, the calculated WBI values are small but nonzero and increase in the same order as the Pd−Pd distance (DPd−Pd). In summary, the palladium acetate dimer stability predominantly arises from the interactions between the paddlewheel ligands and Pd center. However, a Pd−Pd interaction (induced by the shorter Pd−Pd distances and Pd−ligand orbital interactions) also contributes to the stability of the dimeric [Pd(OAc)2]2 structure, but to a lesser extent. III.A.2. Effect of Ligand Substitution (R) in the [Pd(O2CR)2]2 Dimer. Next, we examine the impact of the electronic properties of the carboxylate ligands on the stability of the [Pd(O2CR)2]2 dimer and its Pd−Pd interaction. Let us first provide a detailed orbital analysis of the Pd−ligand and Pd−Pd interactions in the [Pd(O2CR)2]2 dimer.

Figure 2. A comparison of the calculated and experimental geometric parameters for the trimeric Pd-acetate complex [Pd(OAc)2]3.

S,S-donor dithiocarboxylates27 or N,N-donor formamidinates28 results in the preparation of stable dimeric compounds. Thus, the precise nature of the active palladium acetate species in a catalytic mixture is a function of several factors, including the solvent, additives, substrate, and the relative concentration of Pd. For example, using Raman spectroscopy, Kragten and coworkers have concluded that, in acetic acid solution and at high KOAc or ethylene concentration, a palladium acetate dimer can exist.29 Despite the fact that the palladium acetate dimer [Pd(OAc) 2 ] 2 , 30 as well as its thioacetate analogue [Pd(SAc)2]2,27d,31 have been the subject of several previous studies, the factors controlling their precise structural motifs require special attention. In general, [Pd(OAc)2]2 may have several different isomeric forms, and five of them are computationally addressed in this paper. As shown in Figure 3, the isomer with four bridging acetate ligands, 1, is the

Figure 3. Conformational analysis of the dimeric Pd-acetate complex [Pd(OAc)2]2 in which the complex with four paddlewheel bridging ligands is the most stable. Energies are reported as ΔG/ΔH in kcal/ mol.

thermodynamically most stable dimer structure. Isomer 2 with two bridging ligands and a possible Pd−Pd interaction lies only ΔG = 8.6 kcal/mol higher in energy. All other calculated isomers (including the type of planar isomer (5) examined by Lledos and co-workers32) are much higher in energy and will not be discussed further (see SI for structures and energies). The stability of the palladium acetate dimer (relative to monomers) can be attributed to the direct metal−metal Pd(II)−Pd(II) interaction and the bridging interactions of the paddlewheel acetate ligands. To evaluate roles of each of these factors in the stability of the [Pd(OAc)2]2 dimer, we analyze them separately. For this purpose, we examine the 831

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

Figure 4. Impact of removing the bridging acetate ligands on the geometries and orbital interactions of the Pd acetate complexes 1, 2, and 6, which have four, two, and zero bridging acetate ligands, respectively.

shield the electrostatic repulsive interaction between the Pd cations, which have a formal +2 charge. On the basis of this analysis, it is natural to expect that an increase in the electron density on the ligands will decrease the electrostatic repulsion between the Pd atoms and, consequently, shorten the Pd−Pd distance. On the other hand, a decrease in the electron density on the ligand will increase the positive charge on the Pd centers, increase electrostatic repulsion between them, and consequently, elongate the Pd−Pd distance. The calculated results for the impact of different R groups with varying electronic properties on the Pd−Pd interaction in [Pd(O2CR)2]2 dimers are consistent with the aforementioned orbital analyses (Table 1). Further examination of the data reveals an inverse relationship between the strength of the Pd− Pd interaction and the stability of the dimer. Indeed, as shown in Table 1, substitution of R = H to an electron-donating group (EDG), such as R = Ph, CH3, OMe, and NMe2, decreases the overall stability of the [Pd(O2CR)2]2 dimer (relative to separated monomers) via the order R = H > Ph > Me > OMe > NMe2. In contrast, substitution of R = H to an electronwithdrawing group (EWG), such as R = CF3 and CN, only slightly changes the overall stability of the dimer. In contrast, EDG substitution reduces the calculated Pd−Pd distances, whereas EWG substitution elongates it by 0.03−0.04 Å. Thus, in general, the factors affecting the electron-rich and -poor nature of the bridging O-centers determine the stability of the [Pd(O2CR)2]2 dimer. For ligands with varying σ-donation, the trends are straightforward: the R ligands reducing the negative charge of the bridging O-centers increase the stability of the dimer complexes. However, the trends are complicated for the R-groups with π donating or withdrawing nature, partly because they change the π-electron density of the bridging O-centers and impact the Pd−ligand interaction via the δ*(dxy)-π(CO) channel. One should mention that similar conclusions were

Figure 5. (a) Idealized Pd−Pd molecular orbitals ignoring any d/s/p mixing, highlighting the shape of the unhybridized Pd−Pd dz2 σ* orbital. (b) Shapes of the Pd−Pd dz2 σ and σ* orbitals taking into account hybridization of the latter with Pd−Pd s/p orbitals; note, the Pd−Pd dz2 σ orbital is too low in energy to engage in this hybridization and remains mainly unhybridized.

To begin, the Pd−ligand interaction is a result of the σ donation into the empty Pd dx2−y2 orbitals. There are additional filled−filled interactions of σ and π symmetry involving the Pd dz2 orbital and Pd2 δ/δ* dxy orbitals, respectively. It is natural to expect that increasing the electron density of the ligand orbitals will increase the covalency with the Pd orbitals. On the other hand, decreasing the electron density of the ligand orbitals will lead to more polarization of the Pd−ligand interaction. The polarization of these interactions is manifested computationally in the calculated charge of the Pd centers. Therefore, changing the electron density of the ligand orbitals can serve to partially

Table 1. Effect of Changing R to Electron-Donating and -Withdrawing Groups on the Electronic Structure and Stability of the Dinuclear d8-d8 Pd(II)−Pd(II) Complex R

ΔGdimer/ΔHdimer (kcal/mol)

DPd−Pd (Å)

WBI Pd−Pd

E(2)i→j (kcal/mol) dzz → pz

NBO charge Pd (dimer)

NBO charge O (free ligand)

NMe2 OMe CH3 Ph H CF3 CN

−14.7/−25.4 −19.7/−33.3 −24.3/−37.1 −28.2/−38.2 −33.4/−46.9 −35.2/−51.2 −32.6/−46.5

2.61 2.64 2.62 2.62 2.65 2.68 2.69

0.164 0.155 0.166 0.165 0.172 0.160 0.154

123.7 99.8 109.7 96.2 94.6 89.3 97.4

0.76 0.80 0.75 0.77 0.76 0.81 0.84

−0.85 −0.82 −0.84 −0.81 −0.82 −0.77 −0.73

832

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

Figure 6. Substrates with commonly utilized directing groups (highlighted in red) in C−H functionalization and differing sizes that were employed in this computational study. Here and below, “Xn” in notation of substrates indicates a ring-size of the substrate by placing the coordinating atom of the DG in a 1,5 (5n) or 1,6 (6n) relationship to the activated hydrogen atom.

made by Berry and co-workers in an X-ray study of the [Pd2(DAniF)4] and [Pd2(DAniF)4]PF6 complexes, where DAniF = di-p-anisylformamidinate.33 The authors have shown that one-electron oxidation of [Pd2(DAniF)4] decreases the Pd−Pd bond distance by 0.052 Å. In contrast, an X-ray study of [Pd2(DTolF)4], where DTolF = di-p-tolylformamidinate, showed that its one-electron oxidation results in a 0.015 Å increase of the Pd−Pd bond distance. Notably, the Pd−Pd distance in the less electron-rich Pd2(DTolF)4 molecule is ∼0.03 Å shorter than that in the more electron-rich Pd2(DAniF)4 compound. These findings appear to be in conflict with experimental data that indicate the monomeric form of Pd(TFA)2 (R = CF3) is present in coordinating organic solvents, such as EtOAc.12a,37 At this stage, we have no firm explanation for this discrepancy, although it could relate to the lability of the TFA ligand in the monomer Pd(O2CR)2 complex (e.g., relative to acetate) and the impact of coordinating solvents to the stability of the dimer. III.B. The Palladium Acetate Catalyst in C−H Functionalization. Above, we have analyzed the factors impacting the stability of the [Pd(RCO 2 ) 2 ] 2 dimers relative to the corresponding monomers. In general, we have shown that (a) the stability of the [Pd(RCO2)2]2 dimer predominantly arises from the interactions between the paddlewheel ligands and Pd centers. The Pd−Pd interaction also contributes to the stability of the dimeric [Pd(OAc)2]2 structure, but to a lesser extent; (b) R-substitution leading to increase (or decrease) in the electron density on the ligand decreases (increases) the electrostatic repulsion between the Pd centers and therefore strengthens (weakens) the Pd−Pd interaction, whereas R-substitution leading to increase (or decrease) in the electron density on the ligand greatly impacts on the stability of dimer through the bridging Pd−O interactions. These findings are consistent with and complement numerous previous studies. Another factor contributing to the nuclearity and reactivity of the Pd-acetates in a catalytic mixture is the nature of the substrate/ligand−Pd interaction. Indeed, as mentioned in the Introduction, Cook and Sanford12a have shown that, during acetoxylation of benzene, the Pd(OAc)2/pyridine catalyst generated from either 1:2 or 1:1 combination of Pd(OAc)2

and pyridine rests as either a monomer (pyridine)Pd(OAc)2 or dimer [(pyridine)Pd(OAc)2]2, respectively. Recent kinetic studies concluding that C−H activation occurs on the Pd(II) dimer with a benzylamine DG38 and on the Pd(II) monomer with 2-phenylpyridine echo these conclusions.39 These and other experimental findings make it clear that an understanding of the effect of the Pd−substrate interaction on the nuclearity of the catalytically active species in Pd-acetate-catalyzed chemical transformations is imperative for designing better catalysts. Here, we illustrate this issue for directing group (DG) assisted C−H bond activation by Pd(II) systems because the directed C−H functionalization catalyzed by Pd(II)-acetates is one of the actively used synthetic strategies for the C−C and C−heteroatom bond formation. For this reason, we evaluate the impact of the Pd−DG interaction on the nuclearity and reactivity of the Pd-acetate in the C(sp2)−H functionalization, namely, for its important steps: substrate coordination, the C(sp2)−H bond cleavage (i.e., carbopalladation), and Pd−C bond formation. We start our investigations with the C−H activation in substrates B-5n [here and below, 5n indicates the coordinating atom of the substrate DG being in a 1,5 position (see Figures 1 and 6 and section III.B.2 below for more details)] and G, as representative substrates by the dimeric and monomeric forms of Pd-acetate. III.B.1. C−H Activation. It is natural to assume that C−H activation initiated from the [Pd(OAc)2]2 dimer would require coordination of a substrate to the Pd center through its DG. To accommodate the substrate, one of the bridging acetate ligands of [Pd(OAc)2]2 should dissociate. This process could occur either in a concerted or stepwise manner (Figure 7). The barrier for concerted displacement of one of the bridging acetates is calculated to be 3.7 and 6.7 kcal/mol for substrates B-5n and G, respectively. On the other hand, the stepwise process that would be initiated by dissociation of one arm of a bridging acetate ligand requires 13.4 kcal/mol of free energy. Comparison of these energies allows us to conclude that the concerted ligand exchange pathway is more likely. In either case, the energy required for exchange of one of the bridging acetate ligands by the DG is much less than the calculated dimerization energy for [Pd(OAc)2]2. Therefore, one may 833

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

dissociates and abstracts the aryl proton (Figure 8).14b,41 This process may occur twice on the (tetra)acetate-bridged Pd2 complex (labeled below as step CMD1 and CMD2) and form the “clam shell” structure that is commonly proposed as an intermediate arising from C−H carbopalladation.42 The calculated C−H activation barriers by the Pd(II) dimers, given in Table 3, are comparable with the obtained dimerization energy. Interestingly, the calculated C−H activation barrier is slightly lower for substrate G than for substrate B-5n (Table 3). However, the small difference in the magnitude of the C−H activation barriers for substrates G and B-5n indicate that the effect of the DG on this step is relatively minor. Intriguingly, the nuclearity of the active catalyst has a much larger impact on the C−H activation barrier than the nature of the DG. Indeed, as shown in Table 3, the dinuclear complexes require a higher C−H activation barrier than the mononuclear complexes. Comparison of these C−H activation barriers with the calculated [Pd(OAc)2]2 dimerization energy (24.3/37.1 kcal/mol) shows that starting from the dinuclear [Pd(OAc)2]2 complex, the monomerization and the first C−H activation (CMD1 = 23.2/21.0 and 21.0/18.9 kcal/mol for substrates B5n and G, respectively) processes will compete with each other with a slight preference for the latter. The second C−H activation is a relatively easy process and requires only CMD2 = 15.0/16.3 and 11.9/12.8 kcal/mol energy barriers for substrates B-5n and G, respectively. On the other hand, if the higher energy mononuclear complex is present in the reaction mixture in significant quantities, then the lower C−H activation barrier step might effectively compete with the dimerization process: the calculated C−H activation barriers by monomers are 13.8/ 12.1 and 8.7/8.9 kcal/mol for substrates B-5n and G, respectively. For both substrates, regardless of the nuclearity of the active catalyst, the C−H activation is an exergonic process. Furthermore, the exergonicity of this step is almost two times larger for substrate G than for substrate B-5n. Therefore, the aforementioned experimental findings for the substrates with benzylamine (not studied here) 38 and pyridine39 DGs can be explained by the differences in the preparation of the reaction mixture and the impact of the reaction conditions as opposed to the nature of the DG. III.B.2. Dimerization of the Carbopalladated Product Complex. After investigating the key mechanistic details of the C−H activation in substrates B-5n and G by Pd(OAc)2 dimer and monomer, we are set for a careful examination of the role of the DG in the dimerization of Pd(OAc)(DG-Ar). For this reason, we studied dimer complexes [Pd(OAc)(DG-Ar)]2 with two (9) and zero (9a) bridging ligands (see Figure 9) for the substrates (A−F) with many commonly used DGs (shown in red in Figure 6). On the basis of previous knowledge,16a,40 here, we modeled anionic DGs (A−D) with a coordinated counterion, Cs+ or K+. Previously, the coordination modes of these counterions have been extensively examined both computationally and experimentally in the literature.43 In addition, we varied the ring-size of the substrate by placing the coordinating atom of the DG in a 1,5 (5n) or 1,6 (6n) relationship to the activated hydrogen atom. C−H activation in these substrates leads to the formation of 5- and 6-membered carbopalladated product complexes, respectively. For comparison, we also examined two substrates for which experimental evidence of a dinuclear [Pd(OAc)(DG-Ar)]2 intermediate in C−H functionalization has been reported by Ritter9 (G) and Sanford10 (H).

Figure 7. Concerted and stepwise pathways for DG substitution on the [Pd(OAc)2]2 dimer complex.

safely conclude that substrate coordination to the Pd center and displacement of one of acetate ligands is a facile process and will occur before the dissociation of dimer to monomers for almost all substrates studied here. Overall, substrate coordination to the Pd center of the [Pd(OAc)2]2 dimer and displacement of one of the acetate ligands to form complex 8 is −17.9 and −5.1 kcal/mol exergonic for substrates B-5n and G, respectively. On the basis of the calculated Pd−Pd distance (DPd−Pd) and WBI (see Tables 1 and 2), complex 8 has a comparable metal−metal Table 2. Effect of the DG on the Pd−Pd Interaction upon Substrate Coordination to the [Pd(OAc)2]2 Dimer substrate

DPd−Pd (Å)

WBI (Pd− Pd)

E(2)i→j (kcal/mol)

NBO charge Pd1/ Pd2

B-5n G

2.79 2.77

0.15 0.15

76.2 89.1

0.66/0.73 0.64/0.74

interaction to the [Pd(OAc)2]2 dimer for both substrates despite losing a bridging ligand and the steric bulkiness of the C−H substrates. However, examination of the E(2)i→j data shows that the DG also has an impact on the orbital interactions between the Pd atoms, which is less for substrate B-5n than for substrate G. Thus, substrate G produces a stronger Pd−Pd orbital interaction than substrate B-5n. This is likely because the substrates have different types of p orbitals that are aligned for interactions with the Pd−Pd σ* orbital above and below the coordination planes of the Pd atoms (see SI). This is in contrast to the bridging O2CR ligands examined in the previous section, indicating that the orientation of the DGs in this study produce a different effect on the dimer structure than the bridging O2CR ligands. After substrate coordination, the next step of the C−H activation reaction is Pd−C bond formation that occurs through an “external-acetate assisted” concerted deprotonation-metalation (CMD) pathway.40 At the transition state associated with this process, an adjacent bridging acetate ligand 834

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

Figure 8. Transition state structures for carbopalladation through a CMD pathway for substrates B-5n (left) and G (right) initiated from the dinuclear palladium acetate active species.

attributed to the existence of strong attractive ligand−ligand interactions, such as π−π stacking.36 The Pd−Pd interaction seems to be insignificant based on the long calculated metal− metal distance relative to that in the [Pd(OAc)2]2 dimer. Here, one should mention that the calculated Pd−Pd bond distances, 3.07, 2.90, and 2.89 Å for substrates A-5n,43d E-5n,34 and G15h are in good agreement with their available experimental values of 3.00, 2.86, and 2.84 Å, respectively. Thus, on the basis of the above presented findings, it is conceivable that even if C−H activation of substrates would proceed by Pd(OAc)2 monomer, the C−H functionalization may go via dimeric [Pd(OAc)(DG-Ar)]2 complexes, especially for systems with strongly interacting substrates and for electrophiles with small C−X bond formation barriers. In examining the effect of the substrate size, we observe that 5-membered palladacycles generally form more stable dimeric complexes than 6-membered palladacycles with exceptions being A and C. These differences are rooted in the geometric feature of the monomeric 5- and 6-membered palladacycles as shown in Figure 8: 5-membered palladacycles are planar and 6membered palladacycles assume a nonplanar, boatlike conformation. The effect of these structural differences on the dimer geometries is that the 5-membered palladacycles stack more efficiently, and consequently, the DG and substrates can interact more closely.44 The extent of ligand interaction is measured by the distance between the atoms in the ipso position (relative to Pd) in adjacent DGs and substrates (Dipso, Figure 10). As shown in Table 4, Dipso is always larger in the 6membered substrate. The lone exception is D, which also happens to be the smallest DG. However, there is poor correlation between the calculated DPd−Pd and Dipso values and the free energy/enthalpy of dimerization (ΔGdimer/ΔHdimer) across the examined substrates. In fact, many substrates have the same calculated DPd−Pd to within 0.02 Å, (C-5n−H), even with fluctuations in the dimerization energy of up to 12 kcal/ mol. We therefore examined the impact of the bridging ligands and the factors impacting the metal−metal interaction. Similar to that observed for the [Pd2(OAc)2]2 dimer, the bridging acetate ligands play a crucial role in stability of the [Pd(OAc)(DG-Ar)]2 dimer (see SI). In fact, in the [Pd(OAc)(DG-Ar)]2 dimer, the stabilization of the dimer (i.e., free energy of dimerization) can be entirely attributed to the

Table 3. Calculated C−H Activation Barriers and Reaction Energies for Carbopalladation by the Dinuclear (m = 2) and Mononuclear (m = 1) Pd(II)-Acetate Active Species with Substrates B-5n and Ga CMD1 reaction complex B-5n dimer B-5n monomer G dimer G monomer

ΔG‡CH/ ΔH‡CH

CMD2 ΔGmCH/ ΔHmCH

ΔG‡CH/ ΔH‡CH

(kcal/mol)

(kcal/mol)

(kcal/mol)

ΔG2CH/ ΔΔH2CH (kcal/mol)

23.2/21.0 13.8/12.1

−4.7/−6.5 −4.4/−3.4

15.0/16.3

−10.9/−8.5

18.1/15.8 8.7/8.9

−6.0/−5.7 −9.0/−7.8

11.9/12.8

−16.9/−14.2

a

The calculated monomerization energy for the [Pd(OAc)2]2 dimer is 24.3/37.1 kcal/mol (see Table 1).

Figure 9. Definition of the dimerization energy and energy associated with the bridging ligands for the dimeric carbopalladation product [Pd(OAc)(DG-Ar)]2 (9) and its (9a) derivative.

As shown in Table 4, [Pd(OAc)(DG-Ar)]2 (9) dimer formation is favored in both enthalpy and free energy, indicating that the dimer is likely to be present in the reaction mixture in DMF (used in our calculations). Furthermore, the identity of the DG and ring-size of the substrate both have a significant impact on the stability of the corresponding dimer complexes. Dimer stability (vs monomer) is calculated in the order G > E ∼ H > F > D > B > C > A. For substrate G, the [Pd(OAc)(DG-Ar)] dimerization energy (−25.4 kcal/mol) surpasses that of the [Pd(OAc)2]2 dimer (which is −24.3 kcal/ mol), whereas the other substrates (E, F, H) approach it. Taking into account that the dimeric [Pd(OAc)(DG-Ar)]2 complexes have two fewer bridging acetate ligands, the large calculated dimerization energy for substrates E−H can be 835

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

Table 4. Important Structural Parameters, Dimerization Energies (ΔGdimer/ΔHdimer), and NBO Analysis of the Pd(II)−Pd(II) Interaction for the Examined Substrates structure

DPd−Pd calc. (Å)

DPd−Pd exp. (Å)

Dipso calc. (Å)

ΔGdimer (kcal/mol)

ΔΔHdimer (kcal/mol)

WBI Pd−Pd

E(2)i→j (kcal/mol)

A-5n B-5n C-5n D-5n E-5n F-5n G H A-6n B-6n C-6n D-6n E-6n F-6n

3.07 2.99 2.88 2.90 2.90 2.90 2.89 2.90 3.10 3.09 3.01 2.95 2.98 2.96

3.00

4.40 3.99 3.60 3.68 3.59 3.56 3.57 3.53 4.79 4.56 4.35 3.68 3.92 3.71

−9.9 −16.0 −10.2 −18.4 −22.7 −20.8 −25.4 −22.8 −13.8 −11.6 −12.8 −16.8 −22.4 −18.5

−34.1 −27.5 −28.7 −32.3 −35.7 −36.0 −39.1 −36.7 −32.8 −29.3 −28.0 −33.6 −35.7 −35.1

0.12 0.14 0.15 0.14 0.14 0.14 0.14 0.15 0.12 0.13 0.13 0.14 0.14 0.14

21.0 32.4 38.3 35.2 48.4 42.8 41.4 51.0 21.7 29.3 29.1 35.4 41.9 36.9

2.86 2.84

the Pd−Pd distance and stability of the [Pd(OAc)(DG-Ar)]2 dimer. For this purpose, we again turned our attention to the calculated DPd−Pd bond distance, Pd−Pd WBI, and E(2)i→j in each dimer (see Table 4).45 These data show that the shortest Pd−Pd distance does not always correspond to the most stable dimer. The very narrow range of WBI values (0.12−0.15) indicates that the Pd−Pd interactions in these compounds are essentially the same. The strength of the Pd−Pd interactions therefore are unrelated to the stability of the dimer complexes. Direct ligand−ligand interactions appear to have more importance in the Pd(II)−Pd(II) dimeric complexes formed from palladacycles with C−H functionalization substrates. It is worth noting that the most stable dimer complex studied forms from substrate G, which has been implicated by Ritter as playing an important role in C−H functionalization.13 III.B.3. Role of Axial Ligands in the Stability of the Carbopalladated [Pd(OAc)(DG-Ar)]2 Dimer. Another factor that contributes to the thermodynamic stability and Pd−Pd bonding in the dimeric structures, such as [Pd(OAc)2]2 and [Pd(OAc)(DG-Ar)]2, is the Pd−Lax (Lax is an axial ligand, see Figure 1) interaction. On the basis of the molecular orbital picture presented above (see Figure 5), it is natural to expect that Pd−Lax interaction with an electron-withdrawing Lax will reduce electron population/density of the antibonding Pd−Pd

Figure 10. A comparison of the dimers formed by the substrates E-5n and E-6n. The planar monomers in E-5n form more efficient stacking interactions than those of E-6n, as measured by Dipso.

bridging acetate ligand interactions for substrates A−D; however, for substrates E−H, the contribution of the bridging acetate ligand interactions becomes 80−90%, which suggests the existence of stronger interactions between the substrates and possibly some metal−metal interaction. On the other hand, the ΔHbridge values (i.e., enthalpy of removing the bridging ligands, 9 → 9a) are consistently 40−60% of the overall ΔHdimer values (%ΔH). This again indicates favorable interactions between the two metal centers as well as between the substrates. We therefore further examined the impact of the nature of Pd-(DG-Ar) and substrate−substrate interactions on

Figure 11. Schematic presentation of the oxidation processes examined for the [Pd(OAc)(DG-Ar)]2 (9) complex, which produces [I2− Pd(OAc)(DG-Ar)]2 (10) and [I−Pd(OAc)(DG-Ar)]2 (11). The dimerization processes to form complexes 10 and 11 were also examined. 836

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis orbitals and partially oxidize the Pd centers. This will lead to additional stabilization of the dimer structures and shortening of their Pd−Pd distances. These phenomena have the same origin as one- or twoelectron oxidation of Pd(II) dimers leading to the experimentally reported Pd−Pd motif with a three-electron/twocenter bond33 (like in complex [Pd2(DAniF)4]+) or a twoelectron/two-center bond between the palladium centers,11−14,28i respectively. Because the most important component of the Pd−Pd bonding in these dimers is expected to be the σ [Pd(dz2)-Pd(dz2)] interaction (see Figure 5), one may expect that the axial ligands with strong σ-withdrawing character will show more partial oxidation of the Pd−Pd motif. As we16a and others46 have shown recently, molecular iodine, which interacts with the Pd dz2 orbital as an electron acceptor, can be one such ligand. Therefore, as a next step, we analyze the impact of coordination of molecular iodine and iodine atom (Lax = I2 and I, respectively) to the Pd centers of the [Pd(OAc)(DG-Ar)]2 dimers on the stability of these dimers (Figure 11). Comparison of the calculated DPd−Pd distance for dimer complexes [Pd(OAc)(DG-Ar)]2 (9), [I2−Pd(OAc)(DG-Ar)]2 (10), and [I−Pd(OAc)(DG-Ar)]2 (11), given in Tables 4 and 5, shows that coordination of two I2 molecules to the Pd

Table 6. Thermodynamic Energies for Dimerization and Oxidation of the Studied Pd(II)−Pd(II) Dimeric Structures with I2 (See Figure 11 for a Definition of the Reported Energies)

DPd−Pd calc. 10 (Å)

Pd−Pd WBI 10

DPd−Pd calc. 11 (Å)

Dipso calc. 11 (Å)

Pd−Pd WBI 11

A-5n B-5n C-5n D-5n E-5n F-5n G H A-6n B-6n C-6n D-6n E-6n F-6n

3.05 2.95 2.87 2.85 2.85 2.85 2.85 2.86 3.10 3.07 3.11 2.93 2.99 2.86

0.135 0.128 0.147 0.144 0.146 0.143 0.145 0.146 0.115 0.121 0.102 0.129 0.119 0.139

2.85 2.70 2.72 2.69 2.72 2.72 2.72 2.72 2.90 2.88 2.69 2.72 2.78 2.76

3.78 3.28 3.25 3.20 3.27 3.14 3.24 3.25 3.99 3.75 3.65 3.14 3.41 3.26

0.324 0.292 0.349 0.358 0.340 0.339 0.339 0.340 0.314 0.321 0.324 0.355 0.333 0.336

ΔGdimer/ ΔHdimer 10 (kcal/mol)

[ΔG2OX/ ΔH2OX]I (kcal/mol)

ΔGdimer/ ΔHdimer 11 (kcal/mol)

A-5n B-5n C-5n D-5n E-5n F-5n G H A-6n B-6n C-6n D-6n E-6n F-6n

−33.5/−46.5 −39.7/−55.9 −35.8/−49.7 −33.5/−45.9 −26.4/−42.2 −26.8/−42.1 −26.8/−41.6 −28.3/−43.4 −27.3/−43.8 −32.9/−49.6 −23.5/−41.7 −32.5/−46.8 −24.6/−39.5 −30.1/−43.5

−17.2/−35.6 −16.6/−34.5 −19.7/−34.4 −21.7/−37.9 −24.4/−41.1 −27.7/−41.1 −27.5/−44.5 −26.3/−42.1 −9.7/−29.8 −9.5/−27.2 −8.1/−27.0 −18.5/−36.6 −21.2/−36.9 −24.2/−38.9

16.4/22.9 9.6/15.2 14.6/17.7 8.0/13.4 6.8/12.1 6.8/12.2 6.1/11.8 6.2/11.9 16.3/22.7 18.1/23.6 15.9/23.6 9.2/15.4 12.4/16.3 12.3/16.6

−14.9/−33.0 −23.2/−41.9 −20.7/−37.6 −25.3/−42.4 −29.7/−46.8 −30.3/−47.6 −35.0/−50.6 −31.5/−48.1 −7.4/−27.8 −6.2/−25.6 −2.4/−22.5 −19.7/−38.5 −20.9/−37.5 −23.3/−40.3

oxidize, which is consistent with the experimental studies on the oxidation of the dimeric structures of substrates G and H.13,14 In addition, less favorable two-electron oxidation is observed with 6-membered substrates, which can be attributed to the reduction of stacking ability of the substrates at shorter Pd−Pd distances as measured by Dipso in Table 5. Dimerization energies to form the Pd(III)−Pd(III) complex 11 from separated Pd(III) monomers47 are slightly more favorable than those reported for the Pd(II)−Pd(II) dimers in this paper. This is again generally consistent with the increasing strength of the Pd−Pd interaction by the order 11 > 10 > 9. These results show that while oxidation of the [Pd(OAc)(DG-Ar)]2 complex by I2 is unlikely, strongly interacting substrates and smaller palladacycles better facilitate general oxidation of the Pd(II)− Pd(II) motif.

Table 5. Structural parameters for the complexes [I2− Pd(OAc)(DG-Ar)]2 (10) and [I−Pd(OAc)(DG-Ar)]2 (11) structure

structure

[ΔG2OX/ ΔH2OX]I2 (kcal/mol)



CONCLUSIONS Above, we have systematically explored the factors impacting the stability and reactivity of the [Pd(O2CR)2]2 and [Pd(OAc)(DG-Ar)]2 dimers in Pd(OAc)2-catalyzed and directing group (DG) -assisted C−H functionalization. We have shown that (1) the palladium acetate dimer is more stable than separated monomers predominantly due to the interactions between the paddlewheel ligands and the Pd center and that a Pd−Pd interaction contributes to the stability of the dimeric [Pd(OAc)2]2 structure to a lesser extent. (2) R-substitution in Pd(O2CR)2 leads to an increase (or decrease) in the electron density of the ligand orbitals, polarizing the Pd−ligand interaction and weakening (or strengthening) the stability of the [Pd(O2CR)2]2 dimer relative to the monomers. For ligands with varying σ-donation, the trends are straightforward. The R ligands reducing the negative charge of the bridging oxygen ligands increase the stability of the dimer complexes. However, the trends are complicated for the R-group with π-donating or -withdrawing nature. Another factor contributing to the nuclearity and reactivity of the Pd-acetates in directing group assisted C−H functionalization is the nature of substrate/ligand-Pd interaction. It was shown that (3) similar to the [Pd2(OAc)2]2 dimer, for the [Pd(OAc)(DG-Ar)]2 dimers, the major interactions contribu-

centers only slightly (∼0.01−0.05 Å) reduces DPd−Pd, which is close to the 0.05 Å reduction in DPd−Pd observed upon 1electron oxidation of the [Pd2(DAniF)4] dimer.33 Upon going from [I2−Pd(OAc)(DG-Ar)]2 to [I−Pd(OAc)(DG-Ar)]2, DPd−Pd changes dramatically by ∼0.15−0.30 Å, which signifies the formation of the Pd(III)−Pd(III) bonding motif.28i Indeed, the calculated DPd−Pd values for 11 are consistent with that reported experimentally (2.57 Å) for dimer [Cl−Pd(OAc)(DG-Ar)]2 with substrate G.15h The calculated trends in the Pd−Pd WBIs for complexes 10 and 11 closely correlate with the above-presented trends in DPd−Pd and the change in oxidation state of the Pd−Pd motif (see Table 5). As previously observed,16a the coordination of two I2 molecules to complex 9 to form 10 is favorable for all studied substrates [DG-Ar] (Table 6). The trends for DGs and size of substrates follow those of stability (relative to two monomers) of complex 9 (Table 4). Taking the oxidation of 10 further, i.e., 10 → 11 transformation, is highly unfavorable. As shown in Table 6, dimeric structures with substrates E−H are easier to 837

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

(9) (a) Hruszkewycz, D. P.; Guard, L. M.; Balcells, D.; Feldman, N.; Hazari, N.; Tilset, M. Organometallics 2015, 34, 381−394. (b) Melvin, P. R.; Nova, A.; Balcells, D.; Dai, W.; Hazari, N.; Hruszkewycz, D. P.; Shah, H. P.; Tudge, M. T. ACS Catal. 2015, 5, 3680−3688. (c) Oldenhof, S.; Lutz, M.; de Bruin, B.; van der Vlugt, J. I.; Reek, J. N. H. Organometallics 2014, 33, 7293−7298. (10) (a) Ananikov, V. P.; Orlov, N. V.; Zalesskiy, S. S.; Beletskaya, I. P.; Khrustalev, V. N.; Morokuma, K.; Musaev, D. G. J. Am. Chem. Soc. 2012, 134, 6637−6649. (b) Kashin, A. S.; Ananikov, V. P. J. Org. Chem. 2013, 78, 11117−11125. (c) Leyva-Perez, A.; Oliver-Meseguer, J.; Rubio-Marques, P.; Corma, A. Angew. Chem., Int. Ed. 2013, 52, 11554−11559. (11) (a) Proutiere, F.; Aufiero, M.; Schoenebeck, F. J. Am. Chem. Soc. 2012, 134, 606−612. (b) Aufiero, M.; Proutiere, F.; Schoenebeck, F. Angew. Chem., Int. Ed. 2012, 51, 7226−7230. (c) Bonney, K. J.; Proutiere, F.; Schoenebeck, F. Chem. Sci. 2013, 4, 4434−4439. (d) Aufiero, M.; Scattolin, T.; Proutiere, F.; Schoenebeck, F. Organometallics 2015, 34, 5191−5195. (12) (a) Cook, A. K.; Sanford, M. S. J. Am. Chem. Soc. 2015, 137, 3109−3118. (b) Emmert, M. H.; Cook, A. K.; Xie, Y. J.; Sanford, M. Angew. Chem., Int. Ed. 2011, 50, 9409−9412. (13) (a) Powers, D. C.; Ritter, T. Nat. Chem. 2009, 1, 302−309. (b) Powers, D. C.; Geibel, M. A. L.; Klein, J. E. M. N.; Ritter, T. J. Am. Chem. Soc. 2009, 131, 17050−17051. (14) (a) Deprez, N. R.; Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 11234−11241. (b) Canty, A. J.; Ariafard, A.; Sanford, M. S.; Yates, B. F. Organometallics 2013, 32, 544−555. (15) (a) Powers, D. C.; Ritter, T. Acc. Chem. Res. 2012, 45, 840−850. (b) Powers, D. C.; Ritter, T. Organometallics 2013, 32, 2042−2045. (c) Powers, D. C.; Xiao, D. Y.; Geibel, M. A. L.; Ritter, T. J. Am. Chem. Soc. 2010, 132, 14530−14536. (d) Ariafard, A.; Hyland, C. J. T.; Canty, A. J.; Sharma, M.; Brookes, N. J.; Yates, B. F. Inorg. Chem. 2010, 49, 11249−11253. (e) Ariafard, A.; Hyland, C. J. T.; Canty, A. J.; Sharma, M.; Yates, B. F. Inorg. Chem. 2011, 50, 6449−6457. (f) Kornecki, K. P.; Berry, J. F.; Powers, D. C.; Ritter, T. Prog. Inorg. Chem. 2014, 58, 225−302. (g) Nielsen, M. C.; Lyngvi, E.; Schoenebeck, F. J. Am. Chem. Soc. 2013, 135, 1978−1985. (h) Powers, D. C.; Benitez, D.; Tkatchouk, E.; Goddard, W. A.; Ritter, T. J. Am. Chem. Soc. 2010, 132, 14092−14103. (i) Powers, D. C.; Lee, E.; Ariafard, A.; Sanford, M. S.; Yates, B. F.; Canty, A. J.; Ritter, T. J. Am. Chem. Soc. 2012, 134, 12002−12009. (16) (a) Haines, B. E.; Xu, H. Y.; Verma, P.; Wang, X.; Yu, J. Q.; Musaev, D. G. J. Am. Chem. Soc. 2015, 137, 9022−9031. (b) Wang, X. C.; Hu, Y.; Bonacorsi, S.; Hong, Y.; Burrell, R.; Yu, J. Q. J. Am. Chem. Soc. 2013, 135, 10326−10329. (17) 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.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, M. J.; Klene, M.; Knox, J. E.; 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.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc., Gaussian 09, revision D.01; Wallingford, CT, 2009. (18) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270−283. (b) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299−310. (c) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284−298. (19) (a) Grimme, S.; Djukic, J. P. Inorg. Chem. 2010, 49, 2911−2919. (b) Grimme, S.; Djukic, J. P. Inorg. Chem. 2011, 50, 2619−2628. (c) Schwabe, T.; Grimme, S.; Djukic, J. P. J. Am. Chem. Soc. 2009, 131, 14156−14157.

ting to the stability of the dimer are the bridging acetate interactions and π−π stacking between the substrate ligands and that a Pd−Pd interaction (which is stronger for the substrates with a strong Pd-(DG-Ar) interaction) is a less important contributor. (4) Starting from the dinuclear [Pd(OAc)2]2 complex, the monomerization of [Pd(OAc)2]2 and C−H activation by the dimer complex compete with each other; in general, the dinuclear complexes require a higher C− H activation barrier than the mononuclear complexes. (5) Even if C−H activation of the substrates was initiated by the Pd(OAc)2 monomer, the C−H functionalization process may proceed via the dimeric [Pd(OAc)(DG-Ar)] 2 complex, especially for systems with strongly interacting substrates and for electrophiles with small C−X bond formation barriers.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.5b02447. Energetic analysis of the Pd-acetate trimer, Pd−Pd distance energy scans, energetic analysis of the isomeric forms of the Pd−Pd dimers and the effect of the bridging ligands, analysis of the orbital interactions in the Pd(II) dimer, and energies and Cartesian coordinates for the reported structures (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the National Science Foundation under the CCI Center for Selective C−H Functionalization (CHE-1205646). The authors gratefully acknowledge NSF MRI-R2 grant (CHE-0958205) and the use of the resources of the Cherry Emerson Center for Scientific Computation.



REFERENCES

(1) (a) Bonney, K. J.; Schoenebeck, F. Chem. Soc. Rev. 2014, 43, 6609−6638. (b) Musaev, D. G.; Figg, T. M.; Kaledin, A. L. Chem. Soc. Rev. 2014, 43, 5009−5031. (c) Sperger, T.; Sanhueza, I. A.; Kalvet, I.; Schoenebeck, F. Chem. Rev. 2015, 115, 9532−9586. (2) Mirica, L. M.; Khusnutdinova, J. R. Coord. Chem. Rev. 2013, 257, 299−314. (3) Sole, D.; Fernandez, I. Acc. Chem. Res. 2014, 47, 168−179. (4) Helm, L.; Merbach, A. E. Coord. Chem. Rev. 1999, 187, 151−181. (5) (a) Molnar, A. Chem. Rev. 2011, 111, 2251−2320. (b) Baumann, C. G.; De Ornellas, S.; Reeds, J. P.; Storr, T. E.; Williams, T. J.; Fairlamb, I. J. S. Tetrahedron 2014, 70, 6174−6187. (c) Fihri, A.; Bouhrara, M.; Nekoueishahraki, B.; Basset, J. M.; Polshettiwar, V. Chem. Soc. Rev. 2011, 40, 5181−5203. (d) Francesco, I. N.; FontaineVive, F.; Antoniotti, S. ChemCatChem 2014, 6, 2784−2791. (e) Moiseev, I. I.; Vargaftik, M. N. New J. Chem. 1998, 22, 1217−1227. (6) (a) Beletskaya, I. P.; Cheprakov, A. V. J. Organomet. Chem. 2004, 689, 4055−4082. (b) Phan, N. T. S.; Van Der Sluys, M.; Jones, C. W. Adv. Synth. Catal. 2006, 348, 609−679. (c) Yin, L. X.; Liebscher, J. Chem. Rev. 2007, 107, 133−173. (7) Paton, R. S.; Brown, J. M. Angew. Chem., Int. Ed. 2012, 51, 10448−10450. (8) Adrio, L. A.; Nguyen, B. N.; Guilera, G.; Livingston, A. G.; Hii, K. K. Catal. Sci. Technol. 2012, 2, 316−323. 838

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839

Research Article

ACS Catalysis

(42) The “clam shell” Pd(II)−Pd(II) dimer complex consists of two stacked, carbopalladated substrates (DG-Ar) bound to each square planar Pd(II) atom with two bridging ligands that are bidentately coordinated across them. The stacked substrates can assume cis and trans orientations. In cis structures, both DGs are in the same relative coordination site, whereas in the trans structures, each DG occupies opposite coordination sites. In most cases, the trans orientation of the substrates is favored over the cis orientation; thus, structures with the cis orientation will not be discussed further (see ref 32 and SI). (43) (a) Anand, M.; Sunoj, R. B.; Schaefer, H. F. J. Am. Chem. Soc. 2014, 136, 5535−5538. (b) Engle, K. M.; Mei, T. S.; Wasa, M.; Yu, J. Q. Acc. Chem. Res. 2012, 45, 788−802. (c) Figg, T. M.; Wasa, M.; Yu, J. Q.; Musaev, D. G. J. Am. Chem. Soc. 2013, 135, 14206−14214. (d) Zhang, X. G.; Dai, H. X.; Wasa, M.; Yu, J. Q. J. Am. Chem. Soc. 2012, 134, 11948−11951. (44) (a) Dupont, J.; Consorti, C. S.; Spencer, J. Chem. Rev. 2005, 105, 2527−2571. (b) Lentijo, S.; Miguel, J. A.; Espinet, P. Organometallics 2011, 30, 1059−1066. (c) Mawo, R. Y.; Johnson, D. M.; Wood, J. L.; Smoliakova, I. P. J. Organomet. Chem. 2008, 693, 33− 45. (45) Kameo, H.; Kawamoto, T.; Bourissou, D.; Sakaki, S.; Nakazawa, H. Organometallics 2015, 34, 1440−1448. (46) Rogachev, A. Y.; Hoffmann, R. J. Am. Chem. Soc. 2013, 135, 3262−3275. (47) It is likely that the oxidized Pd−Pd motif is best described as Pd(III)−Pd(III) as opposed to Pd(II)−Pd(IV) when Lax = I. See ref 16b.

(20) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132, 154104−154119. (21) (a) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990, 77, 123−141. (b) Igelmann, G.; Stoll, H.; Preuss, H. Mol. Phys. 1988, 65, 1321−1328. (c) Vonszentpaly, L.; Fuentealba, P.; Preuss, H.; Stoll, H. Chem. Phys. Lett. 1982, 93, 555− 559. (22) (a) Cances, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032−3041. (b) Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 106, 5151−5158. (c) Scalmani, G.; Frisch, M. J. J. Chem. Phys. 2010, 132, 114110−114124. (23) Mammen, M.; Shakhnovich, E. I.; Deutch, J. M.; Whitesides, G. M. J. Org. Chem. 1998, 63, 3821−3830. (24) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO, version 3.1; University of Wisconsin: Madison, WI, 1996. (25) Bakhmutov, V. I.; Berry, J. F.; Cotton, F. A.; Ibragimov, S.; Murillo, C. A. Dalton Trans. 2005, 1989−1992. (26) Batsanov, A. S.; Timko, G. A.; Struchkov, Y. T.; Gerbeleu, N. V.; Indrichian, K. M.; Popovich, G. A. Koord. Khim 1989, 15, 688−693. (27) (a) Bonamico, M.; Dessy, G.; Fares, V. J. Chem. Soc., Dalton Trans. 1977, No. 23, 2315−2319. (b) Givaja, G.; Castillo, O.; Mateo, E.; Gallego, A.; Gomez-Garcia, C. J.; Calzolari, A.; di Felice, R.; Zamora, F. Chem. - Eur. J. 2012, 18, 15476−15484. (c) Kobayashi, A.; Kojima, T.; Ikeda, R.; Kitagawa, H. Inorg. Chem. 2006, 45, 322−327. (d) Piovesana, O.; Bellitto, C.; Flamini, A.; Zanazzi, P. F. Inorg. Chem. 1979, 18, 2258−2265. (28) (a) Cotton, F. A.; Matusz, M.; Poli, R. Inorg. Chem. 1987, 26, 1472−1474. (b) Cotton, F. A.; Matusz, M.; Poli, R.; Feng, X. J. J. Am. Chem. Soc. 1988, 110, 1144−1154. (c) Abul-Haj, M.; Quiros, M.; Salas, J. M. Polyhedron 2004, 23, 2373−2379. (d) Berry, J. F.; Cotton, F. A.; Ibragimov, S. A.; Murillo, C. A.; Wang, X. P. Inorg. Chem. 2005, 44, 6129−6137. (e) Visnjevac, A.; Luic, M.; Kobetic, R.; Gembarovski, D.; Zinic, B. Polyhedron 2009, 28, 1057−1064. (f) Chuang, G. J.; Wang, W. K.; Lee, E.; Ritter, T. J. Am. Chem. Soc. 2011, 133, 1760− 1762. (g) Chiarella, G. M.; Cotton, F. A.; Ibragimov, S. A.; Murillo, C. A.; Wilkinson, C. C.; Young, M. D. Polyhedron 2013, 58, 7−12. (h) Yao, C. L.; He, L. P.; Korp, J. D.; Bear, J. L. Inorg. Chem. 1988, 27, 4389−4395. (i) Cotton, F. A.; Gu, J. D.; Murillo, C. A.; Timmons, D. J. J. Am. Chem. Soc. 1998, 120, 13280−13281. (29) Kragten, D. D.; van Santen, R. A.; Crawford, M. K.; Provine, W. D.; Lerou, J. J. Inorg. Chem. 1999, 38, 331−339. (30) (a) Kragten, D. D.; van Santen, R. A.; Lerou, J. J. J. Phys. Chem. A 1999, 103, 80−88. (b) Kragten, D. D.; van Santen, R. A.; Neurock, M.; Lerou, J. J. J. Phys. Chem. A 1999, 103, 2756−2765. (31) (a) Bellitto, C.; Flamini, A.; Piovesana, O.; Zanazzi, P. F. Inorg. Chem. 1980, 19, 3632−3636. (b) Ciullo, G.; Piovesana, O. Inorg. Chem. 1980, 19, 2871−2875. (c) Piovesana, O.; Sestili, L.; Bellitto, C.; Flamini, A.; Tomassini, M.; Zanazzi, P. F.; Zanzari, A. R. J. Am. Chem. Soc. 1977, 99, 5190−5192. (32) Aullon, G.; Ujaque, G.; Lledos, A.; Alvarez, S.; Alemany, P. Inorg. Chem. 1998, 37, 804−813. (33) Berry, J. F.; Bill, E.; Bothe, E.; Cotton, F. A.; Dalal, N. S.; Ibragimov, S. A.; Kaur, N.; Liu, C. Y.; Murillo, C. A.; Nellutla, S.; North, J. M.; Villagrán, D. J. Am. Chem. Soc. 2007, 129, 1393−1401. (34) Bercaw, J. E.; Durrell, A. C.; Gray, H. B.; Green, J. C.; Hazari, N.; Labinger, J. A.; Winkler, J. R. Inorg. Chem. 2010, 49, 1801−1810. (35) Wiberg, K. B. Tetrahedron 1968, 24, 1083−1096. (36) Weber, M.; Klein, J. E. M. N.; Miehlich, B.; Frey, W.; Peters, R. Organometallics 2013, 32, 5810−5817. (37) Stephenson, T. A.; Morehouse, S. M.; Powell, A. R.; Heffer, J. P.; Wilkinson, G. J. Chem. Soc. 1965, No. 0, 3632−3640. (38) Font, H.; Font-Bardia, M.; Gomez, K.; Gonzalez, G.; Granell, J.; Macho, I.; Martinez, M. Dalton Trans. 2014, 43, 13525−13536. (39) Fabre, I.; von Wolff, N.; Le Duc, G.; Flegeau, E. F.; Bruneau, C.; Dixneuf, P. H.; Jutand, A. Chem. - Eur. J. 2013, 19, 7595−7604. (40) Haines, B. E.; Musaev, D. G. ACS Catal. 2015, 5, 830−840. (41) Sanhueza, I. A.; Wagner, A. M.; Sanford, M. S.; Schoenebeck, F. Chem. Sci. 2013, 4, 2767−2775. 839

DOI: 10.1021/acscatal.5b02447 ACS Catal. 2016, 6, 829−839