Mechanism of AlcoholâWater Dehydrogenative Coupling into

Article pubs.acs.org/IC

Mechanism of Alcohol−Water Dehydrogenative Coupling into Carboxylic Acid Using Milstein’s Catalyst: A Detailed Investigation of the Outer-Sphere PES in the Reaction of Aldehydes with an Octahedral Ruthenium Hydroxide Faraj Hasanayn,* Lara M. Al-Assi, Rasha N. Moussawi, and Boushra Srour Omar Department of Chemistry, The American University of Beirut, Beirut, Lebanon S Supporting Information *

ABSTRACT: In aqueous basic media, the square-pyramidal complex [Ru(PNN)(CO)(H)] (1-Ru, where PNN is a dearomatized bipyridyl-CH-PtBu2 pincer ligand) catalyzes the transformation of alcohols and water into carboxylates and H2. A previous theoretical investigation reported the following mechanism for the reaction: (i) metal-catalyzed dehydrogenation of the alcohol into an aldehyde, (ii) metal−ligand cooperation (MLC) addition of water to 1-Ru to give an octahedral ruthenium hydroxide (2-Ru−OH), (iii) concerted MLC hydration of the aldehyde by 2-Ru−OH to give separated 1-Ru and a gem-diol, and (iv) concerted MLC dehydrogenation of the gem-diol by 1-Ru into an octahedral ruthenium dihydride (2-Ru−H) and a carboxylic acid. We calculate the outersphere PES in the reaction between the aldehyde and 2-Ru−OH to start with a localized coupling step yielding an ion-pair minimum (7-ip-OH) in which the hydroxyl group of an α-hydroxyl-alkoxide (gem-diolate) is coordinated to the metal of a cationic square-pyramidal complex. From 7-ip-OH, we identify a route to carboxylic acid that circumvents ligand deprotonation involving (i) 1,1-rearrangement of the gem-diolate within the contact ion pair through an α-OH/O− slippage TS into the octahedral 2-Ru−OCH(OH)R and (ii) a second 1,1-rearrangement through an α-O−/H slippage TS that gives a new ion-pair minimum in which the α-hydrogen of the anion is coordinated to the metal, followed by a localized hydride-transfer TS that gives a carboxylic acid and the octahedral hydride complex (2-Ru−H). The net transformation from 2-Ru−OH and the aldehyde to the carboxylic acid and 2-Ru−H can be viewed as a H/OH metathesis in which a hydride and a hydroxide are exchanged between the acyl group of the aldehyde and the metal center of 2-Ru−OH. The MLC mechanism gives the same metathesis products through the intermediacy of a gem-diol. When the SMD solvent continuum model is applied during geometry optimization with water as the solvent, the Gibbs free energy profile of the slippage pathway is predicted to be much lower than that predicted for MLC. The possibility of dissociation of the ion pair 7-ip-OH into free ions and reassociation is also briefly addressed. Some calculations are also performed to address why no esters are observed in the given system.

■

INTRODUCTION Dearomatized square-pyramidal complexes related to 1-Ru and 1′-Ru were used by Milstein and co-workers to catalyze the acceptorless dehydrogenative coupling of primary alcohols and amines into esters and carboxamides, respectively (eqs 1 and 2).1−3 The same catalysts were also successfully used for the hydrogenation of carboxylic and carbonic acid derivatives, such as the reverse of eqs 1,4,5 and for the dehydrogenative transesterification and amidation of esters using alcohols and amines, respectively.6,7 The given transformations are of great current interest in green chemistry.8,9 Other classes of catalysts and novel applications in this area have been the subject of a number of recent reviews.10,11 As noted by Beller and co-workers, the activities of most of the reported catalysts are still not high enough for industrial applications, and there is a need for more mechanistic understanding in this area.12 © XXXX American Chemical Society

Complexes related to 1-Ru and 1′-Ru undergo a number of characteristic metal−ligand cooperation (MLC) reactions such as the heterolytic addition of H2, primary alcohols, and some anilines that produce octahedral d6 complexes, as illustrated in eqs 3 and 4.1a,13,14 In these two reactions the metal receives a hydride or an alkoxide from the substrate, and the carbon linker of the phosphine arm of the PNN ligand (a dearomatized bipyridyl-CH-PtBu2 pincer ligand) is protonated, so there is no change in the formal oxidation state of the metal. Given the prevalence of these reactions, Milstein and coworkers have systematically postulated a role for MLC in catalysis.15 For example, ester hydrogenation by 2′-Ru−H was proposed to start with hydrogenation of the carbonyl group of the ester to give a free hemiacetal (eq 5). In turn, theoretical Received: March 28, 2016

A

DOI: 10.1021/acs.inorgchem.6b00766 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Scheme 1. Outline of the MLC Mechanism Reported by Li and Hall25

for eq 6.21 Subsequently, Li and Hall calculated a lower-energy mechanism based on an outer-sphere MLC mode that can be divided into the three stages, as outlined in Scheme 1.25 The preparatory stage in Scheme 1 has two components: (i) MLC addition of water to 1-Ru to give the octahedral ruthenium hydroxide (2-Ru−OH) and (ii) dehydrogenation of a primary alcohol into an aldehyde. Both 1-Ru and 1′-Ru are known to add water heterolytically to give octahedral hydroxide complexes.21,26 In the case of 1′-Ru, the product was isolated and confirmed by X-ray analysis to have the hydroxide and the hydride in a trans configuration.26 Thus, a role for the octahedral 2-Ru−OH complex in catalysis is plausible. Aldehyde formation as proposed in Scheme 1 is assumed to be an entry step to all of Milstein and co-workers’ alcohol dehydrogenative coupling chemistry.15 In the first report on dehydrogenative coupling using 1′-Ru, Milstein and co-workers supported the plausibility of a role for aldehydes by demonstrating that 1′-Ru also transforms aldehyde−alcohol mixtures into esters.1a We note however that Gusev and co-workers proposed mechanisms for some osmium and ruthenium hydrogenation/coupling catalysts that do not invoke free aldehydes.27,28 Significantly, aldehyde oxidation using water is equivalent to an aldehyde− water shift reaction, a transformation of immense interest in its own right. Heinekey and co-workers recently developed several iridium and rhodium catalysts for this reaction based on piano stool complexes related to 4.29,30 They proposed inner-sphere mechanisms that became the subject of a separate theoretical investigation by Ou and Cundari.31 The second stage in Scheme 1 transforms the aldehyde and 2-Ru−OH into a gem-diol and the dearomatized 1-Ru.25 This is a carbonyl hydration reaction that was reported to proceed through a concerted pericyclic transition state (TS) (TS-MLC-1). Finally, the two products from stage 2 are reacted again to

investigations of Milstein’s catalysts have focused primarily on MLC-type mechanisms.16−20 Recently, Milstein and co-workers used the bipyridyl PNN complex 1-Ru in basic oxygen-free aqueous media to catalyze the coupling of alcohols and water into carboxylates and H2 (eq 6).21 1‐Ru

RCH 2OH + OH− ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ RCOO− + 2H 2 H 2O,reflux

(6)

Equation 6 has a remarkably broad substrate scope, and it also works with methanol to give 3 equiv of H2 and a carbonate anion.22 The latter reaction is relevant to both methanol reforming into H2 and CO2 and the reverse hydrogenation of CO2, two applications of great potential to a hydrogen and CO2 economy.23 Another example of homogeneous catalysts used for these purposes that has some relevance to the present study is the amino hydride complex 3.24 The recent study by Milstein and co-workers presented preliminary calculations of an inner-sphere MLC mechanism

B


Article

Inorganic Chemistry

Scheme 3. Two Ion-Pair-Mediated Pathways to Metal/Acyl− H/OR Metathesis

dehydrogenate the gem-diol into the final carboxylic acid and the octahedral trans-dihydride complex 2-Ru−H (stage 3). This transformation was also reported to proceed concertedly through TS-MLC-2.25 At about the same time as the study by Li and Hall was published, Yang reported calculations on methanol oxidation by a derivative of the amino PNP catalyst 3 mentioned earlier.32 As summarized in Scheme 2, the mechanism proposed by Yang also starts with a preparatory stage for the formation of formaldehyde and the octahedral 3-Ru−OH. However, instead of an MLC (bifunctional) gem-diol formation step that would utilize a proton from the amino group, Yang proposed a sequence of reactions involving free ions and leading to the octahedral α-hydroxyl-alkoxide intermediate [3-Ru−OCH2(OH)]. No TSs were reported for the individual steps in this sequence. From 3-Ru−OCH2(OH), the acid was then eliminated directly through 3-TS-Hyd. Thus, Schemes 1 and 2 offer very different views for the same oxidative transformation of an aldehyde by two octahedral Ru−OH complexes. We have been interested in using electronic structure methods to study some of the reactions mediated by Milstein’s catalysts. Our calculations of the outer-sphere potential energy surface (PES) of the reaction between the dimethyl amino analogue of the octahedral trans-dihydride 2′-Ru−H and methyl acetate identified a low-energy reaction pathway that involves cleavage of the C−OR bond of the ester instead of an MLC hydrogenation of the carbonyl group into a hemiacetal.33,34 This PES starts in a localized hydride-transfer step leading to a hemiacetaloxide anion bound through the α-CH bond to the metal center of a square-pyramidal cation. As illustrated in Scheme 3, the proposed mechanism for C−OR bond cleavage proceeds by a 1,1-rearrangement taking place in an outer-sphere mode within the contact ion pair to give an aldehyde and a ruthenium alkoxide. We propose that the net reaction can be classified as a H/OR metathesis in which a hydride and an alkoxide are exchanged between a metal center and an acyl group. As a variation on the same ion-pair 1,1-rearrangement mode, we identified another pathway to metathesis proceeding by carbonyl-group insertion into the Ru−H bond of 2′-Ru−H to give an octahedral hemiacetaloxide intermediate, followed by a matching outer-sphere carbonyl-group deinsertion that transfers

the OR group to the metal and eliminates an aldehyde (Scheme 3). Similar pathways were characterized in the reaction of 2′-Ru−H with dimethyl carbonate and methyl formate.35 We also showed that ion-pair formation and rearrangement can be readily adapted to account for metal-catalyzed transesterification and ester amidation.33,34 Returning to Scheme 1, one can see that the sum of the two stages of gem-diol formation and dehydrogenation as proposed in the MLC mechanism also amounts to a metal/acyl metathesis except that the exchanging groups are a hydride and a hydroxide (eq 7). The same is, of course, true for the free-ion dissociative path in Scheme 2.

Thus, by analogy with Scheme 3, it is possible to construct ion-pair-mediated routes to eq 7 that circumvent ligand deprotonation. However, the accessibility of concerted MLC TSs as given in Scheme 1 would imply that ion-pair minima do

Scheme 2. Outline of an Aldehyde Oxidation Mechanism Involving Free Ions Proposed by Yang32

C


Article

Inorganic Chemistry

from those obtained using geometries optimized in the SMD water continuum. Because many of the DFT studies on Milstein’s catalysts, including ours, were based on gas-phase geometries, we find it relevant to describe in the present work the effects of water solvation on the geometries and energies of the transition states and minima in the different mechanisms. To facilitate the presentation, we include figures that include key geometric parameters and the “raw” electronic energies obtained with the small basis set from the two geometries (Egas and Esmd). In a separate set of figures, we present the Gibbs free energy profiles calculated with the larger basis set using the two sets of geometries (Gsmd and Gsmd/gas at 298 K and 1 atm, where we drop the naught symbol for convenience).

not exist on the outer-sphere PES or that they are high in energy. In the present work, we conduct more detailed analyses of the outer-sphere PESs of the reactions in Scheme 1. Our calculations identify a localized coupling step between the aldehyde and 2-Ru−OH leading to an ion-pair minimum of a gem-diolate coordinated to the metal through the OH group. This ion pair can deprotonate the ligand to give a gem-diol, and it can rearrange to transfer a hydride to the metal to complete a net H/OH metathesis. Based on geometries optimized in a water continuum, the calculations predict the kinetics and thermodynamics of slippage and direct carboxylic acid formation to be substantially lower than those of ligand deprotonation and formation of a gem-diol intermediate. Attempts to compare the slippage and free-ion-dissociation profiles of the ion pair are complicated by entropy considerations. We also investigate the MLC and ion-pair-slippage PESs for the transformation from an aldehyde and a ruthenium alkoxide into an ester and 2-Ru−H.

■

■

RESULTS AND DISCUSSION The objective of the present work is to elucidate the details of the outer-sphere PESs that can give a carboxylic acid from 2-Ru−OH and an aldehyde. To put the results in the context of catalysis, we first compute thermodynamic data for some reactions involved in the different mechanisms, and we include a detailed section on alcohol dehydrogenation through an outersphere ion-pair-slippage mode. We work with n-butanol (5), as was done in the study by Li and Hall.25 Thermodynamic Considerations. Scheme 4 collects thermodynamic data for reactions relevant to catalysis computed

COMPUTATIONAL AND SOLVATION METHODS

The computations were performed using Gaussian 09.36 Bulk solvent effects were included through a polarizable dielectric continuum representing water as the solvent using the SMD implementation,37 which is parametrized using an extensive set of experimental data covering both ionic and neutral solutes and several solvents. This model had been validated in many studies38,39 and has become the method of choice in computational organometallic chemistry,40 as in the aforementioned studies of Li and Hall,25 Yang,32 Ou and Cundari,31 and Gusev and co-workers,27,28 for example. In the study by Li and Hall,25 the reported results were determined by optimizing the geometries in the gas phase and then running single-point calculations in the SMD continuum. In the other studies mentioned, on the other hand, the SMD continuum was applied during geometry optimization. In the present work, we conduct two sets of full geometry optimization, one in the gas phase and one in the SMD water continuum.41 Geometry optimization, normal-mode analysis, and intrinsic reaction coordinate (IRC) calculations were all carried out using the ωB97X-D functional42 and the 6-31++G(d,p) basis set for the nonmetal elements.43 Ruthenium carried the SDD relativistic effective core potential (ECP) and associated basis set44 augmented with one f polarization function with an exponent of 1.2445 and an additional diffuse d function with an exponent of 0.015 estimated based on the diffuse function recommended for the aug-cc class of basis sets.46 Final electronic energies were obtained through single-point calculations in the SMD continuum on the two sets of geometries at the ωB97X-D, M06L,47 and M0648 levels using the larger 6-311++G(2d,2p) basis set for the nonmetal elements. To improve the description of Ru in the single-point calculations, an f polarization function with an exponent of 0.4 was added to the basis set used with the ECP in geometry optimization. A pruned (99590) grid was applied in all calculations (specified by the “grid = ultrafine” option of the “Integral” command in Gaussian). The thermal enthalpy and entropy terms used to obtain the free energies in electronic structure calculations were obtained using statistical mechanics methods on individual molecules.49 These methods cannot reproduce entropies of bimolecular reactions taking place in solution. Wang and co-workers elaborated more on this problem in studies on reactions of Milstein’s catalyst and suggested the use of a scaling factor of 0.5 as one possible improvement,17 which we also employed in our previous studies.33−35 The different studies by Li and Hall25 and Yang32 of interest to the present work, on the other hand, did not apply scaling factors. To allow for more direct comparisons with the latter studies, we give the results based on unscaled entropies, and we describe the effects of scaling them when the entropy is critical to the conclusions. The choice of density functional was found to have minor effects on the energies, so we present only the ωB97X-D results and provide the full data set as Supporting Information. In contrast, the SMD results based on the gas-phase geometries offer very different energy profiles

Scheme 4. Computed Standard-State Thermodynamic Parameters of Reactions Relevant to Catalysisa

a

SMD-optimized geometries at 298 K and 1 atm. Units of kcal/mol and eu. Entropies are not scaled.

at the ωB97X-D level using the large basis set in the SMD water continuum. Water addition to 1-Ru is computed to be slightly less exothermic than alcohol addition: ΔH = −12.7 and −15.1 kcal/mol, respectively (eqs 8 and 9, Scheme 4). The computed unscaled entropies for the two reactions are large but significantly different, at −40 and −49 eu, respectively, which makes the free energies of the two reactions similar and only slightly exoergic: approximately −0.7 kcal/mol (ΔGsmd at 298 K and 1 atm). ΔH for the MLC addition of H2 to 1-Ru is −17.9 kcal/mol. Because of unique mass and vibrational considerations,50 the computed ΔS value for H2 addition to 1-Ru (−31 eu) is much smaller than those for the other two reactions, so the corresponding ΔGsmd value is quite exoergic (−8.8 kcal/mol; eq 10 in Scheme 4). We are not aware of experimental data that allow a quantitative assessment of the given energies. D


Article

Inorganic Chemistry The net dehydrogenative coupling of n-butanol and water into butanoic acid and two hydrogen molecules is uphill by 15.6 kcal/mol (eq 11, Scheme 4). The use of stoichiometric amounts of NaOH in catalysis will obviously drive the reaction by neutralizing the acid. Note, however, that acceptorless dehydrogenative coupling is generally thermodynamically unfavorable but can still be driven in the absence of a base by running the reaction under open conditions that remove the H2 byproduct. Alcohol dehydrogenation into aldehyde and H2 is endoergic by 11.6 kcal/mol (eq 12, Scheme 4). Combining eqs 9 and 12 (Scheme 4) amounts to a thermodynamic energy input of 10.9 kcal/mol for the preparatory stage of formation of 2-Ru− OH, the aldehyde, and one H2 molecule from the separated 1-Ru, H2O, and alcohol (eq 13, Scheme 4). Alcohol Dehydrogenation. Qualitative observations indicate that alcohol addition to analogues of 1-Ru (eq 4) is fast,5b so we take 2-Ru−OCH2R as the starting point in studying one possible route to alcohol dehydrogenation. As mentioned in Scheme 4, formation of 2-Ru−OCH2R from 1-Ru and n-butanol is exoergic by 0.6 kcal/mol (eq 9, Scheme 4, using the unscaled value of ΔS = −49 eu). Scaling the entropy by 0.5 affords ΔGsmd = −7.9 kcal/mol. Octahedral d6 metal hydride complexes (including Milstein’s catalysts) are known to undergo carbonyl-group insertion for a number of substrates including ketones,51−53 amides,54 and CO2,55−58 as depicted in eq 14.

unconventional outer-sphere mode without the involvement of an empty cis-coordination site. The same route was considered in the study by Li and Hall, but the slippage TS was found to be 7.4 kcal/mol higher than the TS for hydride transfer.25 We now show that this conclusion depends on whether the water continuum is applied during geometry optimization. For this purpose, we compare in Figure 1 key structural parameters and the raw electronic energies obtained from the gas-phase and SMD optimizations (Egas and Esmd; relative to 2-Ru−OCH2R). The figure also includes Gibbs free energies (Gsmd) based on Esmd values calculated with a large basis set, as explained in the Computational and Solvation Methods section. To clarify the nature of the PES for the given rearrangement, some molecules are displayed in Figure 2 in two orthogonal projections. The alkoxide in 2-Ru−OCH2R makes a Ru−O−CH2R angle of 118° with the metal. In the specific complex used in Figure 2, the C−O bond of the alkoxide is staggered between the terminal Ru−pyridine and Ru−CO bonds,59 and the alkyl chain points upward, away from the metal (d1 = 174°; Figure 2). In this conformation, the two α-hydrogens of the alkoxide point downward, so they are at a relatively close distance from the metal (3.2 Å). The C−H-bound ion-pair minimum of interest has one of the two αCH bonds coordinated to the metal in an anagostic mode (5-ip-CH; Figures 1 and 2). To place the αCH bond in position in the given ion pair, the “intact” alkoxide of 2-Ru−CH2OR is displaced or “slipped” in the direction of the phosphine linker, and the alkyl chain is directed to the side of the complex (d1 in 5-ip-CH = 134°; Figure 2). The calculations identify an α-O−/H slippage TS connecting 2-Ru−OCH2R and 5-ip-CH having a geometry in which the terminal αCO− bond and one of the two αCH bonds of the alkoxide are simultaneously pointed toward the metal (5-TS-Slip; Figures 1 and 2). The coordinates of the imaginary frequency in 5-TS-Slip are fully consistent with a motion that alternates the coordination of the αCO− and αCH bonds of the alkoxide to the metal. Intrinsic reaction coordinate (IRC) analysis conducted in the gas phase (IRCgas) shows the slippage to be associated with a gradual increase in the Ru···OCH2R bond distance taking place against a gradual decrease in the αCH···Ru bond (r1 and r2, respectively; Figure 3). Thus, to reach the TS from 2-Ru_CH2OR, r1 lengthens from 2.2 to 3.1 Å, whereas r2 shortens from 3.1 to 2.3 Å. Interestingly, the slippage takes place with minimal change in the distance between the metal and the carbon of the alkoxide across the IRC (r3 ≈ 3.0 Å; Figure 3).

Alcohol dehydrogenation by 1-Ru can therefore be completed from 2-Ru−OCH2R by the reverse carbonyl-group deinsertion that yields an aldehyde and 2-Ru−H. In Figure 1, we elucidate an outer-sphere β-hydride elimination pathway involving (i) a preparatory 1,1-rearrangement of the alkoxide in 2-Ru−OCH2R through an α-O−/H slippage TS to point an αCH bond of the primary alkoxide toward the metal and (ii) a hydride transfer from the αCH bond to the metal.35 The given path is equivalent to β-hydride elimination taking place in an

Figure 1. Slippage PES for aldehyde deinsertion from 2-Ru−OCH2R. Energies are given in kcal/mol relative to 2-Ru−OCH2R. Selected bond distances are given in Å. E


Article

Inorganic Chemistry

Figure 2. Two orthogonal projections of the molecules depicting slippage from 2-Ru−OCH2R to 5-ip-CH (dihedral angles are from gas-phase geometries).

Figure 1), and the imaginary frequency is lowered to 57i cm−1 (without any change in the coordinates of the vibration). The SMD continuum also exerts pronounced effects on the geometry and energy of the ion-pair minimum 5-ip-CH. Specifically, in the gas phase, the distance between the α-hydrogen of the alkoxide and the metal is short (r2 = 1.85 Å), the αC−H bond is long (1.29 Å), and the energy is high (Egas = 14.5 kcal/mol). In contrast, in the SMD water continuum, r2 is long (2.41 Å), the αC−H bond is “normal” (1.12 Å), and Esmd is low (4.5 kcal/mol). Such large solvent effects are reasonable because the transformations from 2-Ru−OCH2R to 5-TS-Slip and 5-ip-CH involve large degrees of charge separation and are therefore expected to be greatly stabilized by the water continuum. The 1,1-rearrangement in Figure 1 is followed by a TS for hydride transfer from the carbon of the alkoxide to the metal (5-TS-Hyd), leading to a loose complex (6-Plx-1) between the aldehyde (6) and the octahedral trans ruthenium dihydride (2-Ru−H). To reach 5-TS-Hyd, the αCH bond giving the hydride is stretched to 1.58 Å (gas) or 1.74 Å (SMD), and the Ru···H bond being made is contracted to 1.75 Å (gas) or 1.72 Å (SMD). Clearly, the structural requirements needed to reach 5-TS-Hyd starting from 5-ip-CH are minimal in the gas phase but large in the SMD continuum. Accordingly, the electronic barrier for hydride transfer originating from 5-ip-CH is computed to be negligible in the gas phase (ΔE‡gas = 1.0 kcal/mol) but substantial in the SMD continuum (ΔE‡smd = 12.5 kcal/mol). The gas-phase and SMD Gibbs free energy profiles of the TSs and minima in Figure 1 are compared in Figure 4. The highest-energy point on the gas-phase PES is 5-TS-Slip with G‡gas = 25.5 kcal/mol, much higher than 5-TS-Hyd (14.4 kcal/mol). Even the ion-pair minimum 5-ip-CH is slightly higher than 5-TS-Hyd on the gas-phase profile (15.2 vs 14.4 kcal/mol). Including the SMD water continuum in geometry optimization has minor effects on the relative energy of 5-TS-Hyd (G‡smd = 13.9 kcal/mol) but lowers those of 5-TS-Slip and 5-ip-CH to 8.3 and 4.3 kcal/mol, respectively. The ion-pair character is expected to be lost in 6-Plx-1, so the gas and solvated energies are similar for this species, at 8.5 and 8.8 kcal/mol, respectively. Finally, the net deinsertion from 2-Ru−OCH2R to the separated aldehyde and 2-Ru−H is endoergic by 7.4 kcal/mol (gas) or 3.5 kcal/mol (SMD). If the entropies are scaled by 0.5, the energies of the two TSs on the PES are not changed

Figure 3. Energy and geometry changes (distances r in Å) along IRCgas originating from 5-TS-Slip (IRCgas in Bohr·amu1/2 and Egas in kcal/mol relative to 2-Ru−OCH2R).

The reason for going over such details at length is to convey that the structural requirements needed to reach 5-TS-Slip are overall minor. Because of the strong trans influence of the hydride ligand,60 the metal−alkoxide bond is expected to be weak in the octahedral d6 2-Ru−OCH2R complex. Furthermore, as is evident in the molecular structures in Figure 2, the bipyridyl rings in 2-Ru−OCH2R are nearly planar, so there are no major steric factors imposed by the PNN ligand that could obstruct the rearrangement. Nevertheless, the gas-phase electronic barrier needed to reach 5-TS-Slip is computed to be substantial: ΔE‡gas = 25.8 kcal/mol. However, when the geometries are optimized in the SMD water continuum, ΔE‡smd is much smaller at 9.9 kcal/mol. In the solvated 5-TS-Slip structure, the alkoxide is pulled away from the metal center by approximately 0.4 Å compared to the gas-phase geometry (r1 = 3.4 and r2 = 2.8 Å; F


Article

Inorganic Chemistry

effects is beyond the scope of the present work. In the remaining parts of the study, we focus on the Gsmd results, but we also give the Gsmd/gas values to allow for comparisons with the results from the study by Li and Hall.25 In summary, we consider a simple outer-sphere mechanism for aldehyde deinsertion from 2-Ru−OCH2R in which the alkoxide is first slipped to align the αCH bond in position to be able to transfer a hydride to the metal. The preparatory slippage has a kinetic barrier but this barrier is too small compared to the barrier for hydride transfer that “oxidizes” the alkoxide into an aldehyde. In a later section, we consider similar outer-sphere modes in the transformation from the aldehyde and 2-Ru−OH into carboxylic acid and 2-Ru−H. We first examine the details of the metal−ligand cooperation PES. PES for Gem-Diol Formation through MLC. Because of the asymmetry of the PNN ligand, the outer-sphere reaction of an aldehyde with 2-Ru−OH can proceed along several stereoisomeric pathways. In Figure 5, we follow one pathway that orients the oxygen of the carbonyl group of the aldehyde in the direction of the axial proton of the methylene linker of the phosphine arm.61 The energies in the figure are defined relative to the separated 2-Ru−OH and aldehyde (6). Molecular structures of the key minima and TSs on this PES are given in Figure 6. Additional Gibbs free energy profiles are presented and discussed in a later section. Figure 5 starts with the formation of a loose outer-sphere minimum between the aldehyde and 2-Ru−OH (6-Plx-2), followed by a TS for a localized carbonyl···OH coupling step (7-TS-OH) taking place while the hydroxide is still coordinated to the metal (Ru−OH = 2.27 Å in 7-TS-OH, compared to 2.24 Å in 6-Plx-2). The coupling center in 7-TS-OH is characterized by a C···OH bond distance of 2.07 Å and ν‡smd = 197i cm−1 for a pure C−OH stretching vibration. Consistent with the localized nature of the step, the calculations identify an ion-pair minimum of the α-hydroxyl-alkoxide coupling product (7-ip-OH) in which the Ru−OH bond distance (2.36 Å) is slightly lengthened compared to those in 6-Plx-2 and 7-TS-OH. IRCsmd originating from 7-TS-OH confirms the connectivities and reveals a rather flat PES from the coupling direction with a small barrier and slightly favorable thermodynamics (ΔE‡smd = +3.3 and ΔEsmd = −4.5 kcal/mol, relative to 6-Plx-2). A localized carbonyl···OH coupling TS and a flat IRC are also computed in the gas phase. From 7-ip-OH, gem-diol formation proceeds through 7-TSPrt-1 for a localized proton transfer from the axial methylene

Figure 4. Gas-phase and SMD Gibbs free energy profiles (kcal/mol using unscaled entropies) for aldehyde elimination from 2-Ru− OCH2R at 298 K and 1 atm. Values in brackets are from single-point SMD calculations on the gas-phase geometries.

significantly relative to that of 2-Ru−OCH2R, but the reaction thermodynamics become much more endoergic: Gsmd = 10.2 kcal/mol. Figure 4 includes results from single-point SMD calculations on the gas-phase geometries (Gsmd/gas data given in brackets). On this scale, the G‡smd/gas values for 5-TS-Slip and 5-ip-CH are 20.2 and 10.3 kcal/mol, respectively. Clearly, the SMD/gas protocol recovers only a fraction of the solvation effects obtained when the geometries are optimized in the SMD continuum. Note that including a single water molecule that is H-bonded to the alkoxide in the gas-phase geometry optimization affords G‡gas = 23.4 and 20.1 kcal/mol for 5-TS-Slip and 5-TS-Hyd, respectively. When the “hydrated” gas-phase geometries are used in single-point SMD calculations, 5-TS-Slip becomes 2.9 kcal/mol lower than 5-TS-Hyd (G‡gas/smd = 16.6 and 19.7 kcal/mol, respectively). This effect of one explicit solvent molecule parallels that of including the SMD solvent continuum during geometry optimization. Conducting further computations to refine solvation such as including dynamics

Figure 5. MLC PES for gem-diol formation from 2-Ru−OH and 6. Energies are given in kcal/mol relative to the separated reactants. Bond distances are in Å. G


Article

Inorganic Chemistry

Figure 6. Molecular structures of selected stationary points from Figure 5.

Figure 7) shows an initial flat region from the direction of 7-ip-OH corresponding to realignment of the alkoxide to bring the terminal anionic oxygen close to CHax (see Figure 6 for geometry details), followed by a steep rise in the energy near the TS region. We make two related points here: (i) The proton of the ligand does not play any role in directing the outer-sphere reaction of an aldehyde with 2-Ru−OH, and (ii) proton transfer is not coupled in any way to the initial C−OH bond-making step. In fact, the electronic barrier for proton transfer starting from 7-ip-OH is substantial: ΔE‡smd = +16.9 kcal/mol. When 7-ip-OH is optimized in the gas phase, the distance between the anionic oxygen and CHax is contracted to 2.0 Å, but the CHax bond is still not activated (1.11 Å). Thus, even in this case, the anionic oxygen has to be brought to 1.37 Å from the proton in 7-TS-Prt-1, and there is a barrier to the step, albeit a small one: ΔE‡gas = 2.7 kcal/mol. Proton transfer yields first a loose outersphere complex (7-Plx-1) between the gem-diol (7) and the dearomatized square-pyramidal 1-Ru, with one hydroxyl group coordinated to the metal at 2.4 Å and the proton of the second hydroxyl group pointed to the deprotonated carbon of the ligand at 2.2 Å (comparable SMD and gas-phase results). PES for Gem-Diol Dehydrogenation through MLC. The stationary points on the outer-sphere PES for gem-diol dehydrogenation by 1-Ru into the carboxylic acid product of interest (8) are given in Figure 8. The energies in the figure are defined relative to the separated aldehyde and 2-Ru−OH, as in Figure 5. Molecular structures of selected stationary points are given in Figure 9. Figure 8 begins with a new loose complex (7-Plx-2) between the gem-diol and 1-Ru in which the αCH bond of the gem-diol points toward the metal (at 2.73 Å) and a hydroxyl proton points toward the deprotonated carbon of the PNN ligand

linker of the PNN ligand to the terminal oxygen of the α-hydroxyl-alkoxide. In 7-TS-Prt-1, the C···H···O angle is nearly linear (173°), and the transferring proton is closer to the acceptor oxygen than to the donor carbon (1.24 vs 1.40 Å), thus implicating a “late” TS in the direction of gem-diol formation (SMD results). 7-TS-Prt-1 has ν‡smd = 1450i cm−1 for a pure motion of the proton between an oxygen of the substrate and the carbon of the ligand. In 7-ip-OH, the terminal (anionic) oxygen of the α-hydroxylalkoxide is oriented away from the Ru−PNN plane. The distance between this oxygen and the axial methylene proton of the ligand is 3.1 Å, slightly longer than in 6-Plx-2 (2.8 Å) and 7-TS-OH (3.0 Å) (SMD results). Furthermore, the CHax and CHeq bond distances in 7-ip-OH are similar: 1.10 and 1.09 Å, respectively. These values argue against any interactions that might activate the axial C−H proton of the ligand upon ion-pair formation. Accordingly, IRCsmd (originating from 7-TS-Prt-1;

Figure 7. IRCsmd plots (in Bohr·amu1/2 and kcal/mol) originating from 7-TS-OH and 7-TS-Prt-1. Electronic energies defined relative to the separated 2-Ru−OH and aldehyde.

Figure 8. MLC PES for gem-diol dehydrogenation by 1-Ru. Energies are given in kcal/mol relative to 6 and 2-Ru−OH as used in Figure 5. Bond distances are in Å. H


Article

Inorganic Chemistry

Figure 9. Molecular structures of selected stationary points from Figure 8.

(OH···C = 2.1 Å; SMD values). Thus, 7-Plx-2 is a rotamer of the same gem-diol complex produced at the end of Figure 5. From 7-Plx-2, 7-TS-Prt-2 returns a proton to the ligand to give a new ion-pair minimum of the gem-diolate anion (7-ip-CH). Again, as clarified by the molecular structures in Figures 6 and 9, these species are rotamers or linkage isomers of 7-TSPrt-1 and 7-ip-OH, respectively. Accordingly, the ν‡smd values of 7-TS-Prt-1 and 7-TS-Prt-2 are nearly identical (1450i cm−1), and their E‡smd values are comparable (4.0 vs 1.1 kcal/mol). Dehydrogenation is completed by a separate localized hydridetransfer TS (7-TS-Hyd; ν‡smd = 905i cm−1) in which the αCH bond of the gem-diolate is stretched to 1.56 Å and the Ru···H bond in the making is at 1.80 Å. Hydride transfer initially yields a loose complex (8-Plx) between the octahedral trans-dihydride 2-Ru−H and a carboxylic acid (8). IRCsmd calculations originating from 7-TS-Prt-2 and 7-TS-Hyd confirm the intermediacy of an ion-pair minimum in gem-diol dehydrogenation and reveal steep regions at both TSs (Figure 10).

Figure 11. Comparison of IRCgas plots (in Bohr·amu1/2 and kcal/mol) of stepwise and concerted CH−OH dehydrogenation reactions. Egas is defined relative to the separated 2-Ru−OH and aldehyde or the separated 9-RuN and isopropanol (eq 15).

In a seminal study by Yamakawa, Ito, and Noyori, eq 15 was calculated to proceed in the gas phase by a concerted bifunctional TS (9-TS-MLC), leading directly to a saturated amino hydride complex and a ketone.62 Consistently, our IRC gasphase calculations starting from 9-RuH-NH show the O−H and C−H bonds of the alcohol to lengthen and eventually break simultaneously along the reaction coordinate (results included in Figure 11). Significantly, recent calculations by Dub and Ikariya showed that, when the SMD solvent continuum is applied during geometry optimization, the PES of eq 15 changes from concerted to stepwise.63 Similarly, calculations on ketone hydrogenation by octahedral ruthenium amino hydride catalysts supported a stepwise reaction mode.64,65 These computational findings allowed for the formulation of ketone hydrogenation and transfer hydrogenation mechanisms that circumvent the amido/amino ligand protonation/deprotonation invoked in Noyori’s conventional bifunctional mechanism. Combined Energy Profiles for H/OH Metathesis through MLC. Two Gibbs free energy profiles for the two stages of gem-diol formation and dehydrogenation are compared in Figure 12, one based on geometries optimized in the SMD water continuum (Gsmd) and one from single-point SMD calculations on the gas-phase geometries (Gsmd/gas; values in brackets). The net transformation from the separated aldehyde and 2-Ru−OH into separated carboxylic acid and 2-Ru−H (H/OH metathesis) is computed to be exoergic by 4.1 kcal/mol (SMD result). In contrast, the separated gem-diol and 1-Ru intermediates on this scale are uphill by 9.1 kcal/mol.66,67 The TS for the coupling step between the aldehyde and 2-Ru−OH (7-TS-OH) has G‡smd = 9.9 kcal/mol. As mentioned before, this step lacks any electronic barrier relative to the separated reactants (E‡smd = −8.4 kcal/mol), so the computed G‡smd value arises primarily from the associative entropy of activation (S‡ = −45 eu; unscaled, relative to the separated reactants).

Figure 10. IRCsmd plots (in Bohr·amu1/2 and kcal/mol) originating from 7-TS-Prt-2 and 7-TS-Hyd. Electronic energies defined relative to the separated 2-Ru−OH and aldehyde.

All of the stationary points in Figure 8 could be identified in both the SMD continuum and the gas phase. It is worth noting, however that, unlike the pure proton motion characterizing ν‡ in the solvated 7-TS-Prt-2, animation of the imaginary frequency of the gas-phase 7-TS-Prt-2 reveals a small but significant motion of the αCH bond to be mixed with the principal O−H stretch motion. This might at first suggest a concerted proton-/ hydride-transfer process. However, the gas-phase IRC shows that the αCH bond does not undergo full hydride transfer past 7-TS-Prt-2 (IRCgas, Figure 11). Instead, it is lengthened from 1.12 Å in 7-Plx-2 to 1.17 in 7-TS-Prt-2 and reaches 1.32 Å in 7-ip-CH (Figure 8). Full C−H cleavage proceeds from 7-ip-CH in the gas phase by a distinct TS having ν‡gas = 270i cm−1 but no electronic barrier. To illustrate that the given gas-phase results for gem-diol dehydrogenation by 1-Ru are more representative of a stepwise than a concerted PES, we consider the dehydrogenation of isopropanol by an unsaturated piano stool ruthenium amide fragment (9-RuN; eq 15). I


Article

Inorganic Chemistry

Figure 12. ωB97X-D/SMD Gibbs free energy profiles for H/OH metathesis through MLC (at 298 K and 1 atm, in kcal/mol using unscaled entropies).

octahedral α-hydroxyl-alkoxide complex 2-Ru−OCH(OH)R, as shown in Figures 13 and 14. The reaction from the separated aldehyde and 2-Ru−OH to 2-Ru−OCH(OH)R is equivalent to a carbonyl-group insertion into a Ru−OH bond taking place in an outer-sphere mode. The study by Milstein and co-workers proposed a key role for 2-Ru−OCH(OH)R in an inner-sphere MLC mechanism.21 Although α-OH/O− slippage puts the negatively charged oxygen at the metal center in place of the neutral OH group, the step is not very favored (ΔEsmd = 2.8 kcal/mol). This is significant because it shows that 7-ip-OH, identified as a minimum in an electronically barrierless coupling step between the aldehyde and 2-Ru−OH (Figure 5), has an energy close to that of a “regular” octahedral alkoxide complex related to those observed experimentally in reactions of Milstein’s catalysts. The electronic barrier to α-OH/O− slippage is only 3.9 kcal/mol. In the equilibrium geometry of 2-Ru−OCH(OH)R, the αCH bond of the alkoxide points toward the metal at a distance of 3.1 Å, and the bulkier alkyl and hydroxyl groups both point upward (Figures 13 and 14). Obviously, this intermediate can undergo an α-O−/H slippage matching 5-TS-Slip considered in Figure 1 for aldehyde deinsertion from 2-Ru−OCH2R. Indeed, the calculations identify 7-TS-Slip-2 connecting 2-Ru−OCH(OH)R to 7-ip-CH with ΔE‡smd = 8.1 kcal/mol. Hydride transfer from 7-ip-CH gives 2-Ru−H and the carboxylic acid. As found for 2-Ru−OCH2R, the slippage TSs in Figure 13 exhibit large dependencies on the SMD continuum. First, the solvated TSs have a greater degree of ion-pair dissociation (by ∼0.4 Å in each case) than the gas-phase geometries. Second, the imaginary frequencies in the solvated 7-TS-Slip-1 and 7-TSSlip-2 (ν‡smd = 27i and 60i cm−1, respectively) are much smaller than those computed in the gas phase (ν‡gas = 147i and 212i cm−1, respectively). Finally, the energies of the solvated slippage TSs relative to the separated aldehyde and 2-Ru−OH (E‡smd = −9.0 and −7.6 kcal/mol) are greatly stabilized compared to the energies computed in the gas phase (E‡gas = −5.3 and −1.3 kcal/mol, respectively).

According to the Gsmd results, the highest-energy point on the MLC PES is 7-TS-Prt-1 (20.7 kcal/mol), slightly higher than 7-TS-Prt-2 (17.6 kcal/mol) and much higher than 7-TS-Hyd (13.5 kcal/mol). When the gas-phase geometries are used, the energy order of the two proton-transfer TSs is reversed to G‡smd/gas = 16.2 vs 19.5 kcal/mol, but each of these TSs remains substantially higher than both 6-TS-OH (9.6 kcal/mol) and 7-TS-Hyd (11.7 kcal/mol). Based on these results, we can pursue a search for other routes for the same H/OH metathesis that can avoid the proton-transfer TSs and the uphill gem-diol intermediate altogether. H/OH Metathesis through Ion-Pair Slippage. The outersphere PES in the reaction between the aldehyde and 2-Ru−OH starts with a localized coupling step leading to 7-ip-OH. The MLC mechanism in Figure 5 invokes formation and reaction of a free gem-diol to rearrange 7-ip-OH into 7-ip-CH to enable transfer of a hydride to the metal to produce the carboxylic acid and 2-Ru−H. In this section, we investigate a more direct mechanism for the same rearrangement through slippage TSs taking place within the intact ion pair.

We first considered the α-OH/H slippage mode that would achieve the rearrangement in a least-action mode. The search for a TS for this mode encountered convergence problems because of the flat nature of the PES and was complicated because the anionic oxygen tends to tip toward the metal during geometry optimization. On the other hand, we identified 7-TS-Slip-1 for an α-OH/O− slippage mode leading to the J


Article

Inorganic Chemistry

Figure 13. Ion-pair-slippage PES for an OH/H metathesis of 2-Ru−OH and 6. Energies are given in kcal/mol relative to the separated reactants.

Figure 14. Molecular structures of the carbonyl insertion and deinsertion TSs and products.

Figure 15. Relaxed 3D PES for the three possible ion-pair-slippage modes of the gem-diolate (in kcal/mol and Å). R1 = 7-ip-OH, P1 = 2-Ru− OCH(OH)R, and P2 = 7-ip-CH.

The results highlight the flat nature of the slippage PESs, and they satisfactorily support the conclusion that the lowestenergy pathway in each case passes through 7-TS-Slip-1 and 7-TS-Slip-2 (denoted as TS1 and TS2, respectively, in the plots). To obtain some information on the OH/H slippage mode that can connect -ip-OH directly to 7-ip-CH, we also calculated the 3D PES defined by the Ru···OH and αCH···Ru bond distances. The plot suggests that there is a TS at approximately Ru···OH = 3.0 A and αCH···Ru = 2.7 Å. The point optimized for a minimum at the given parameters has only one imaginary frequency with coordinates fully consistent with α-OH/H slippage (7-TS-Slip-3fix; eq 16). 7-TS-Slip-3fix has E‡smd = −7.1 kcal/mol, 1.9 kcal/mol higher than the value for 7-TS-Slip-1. The slippage Gibbs free energy profiles are shown in Figure 16. For the geometries optimized in the SMD water continuum, the G‡smd values of 7-TS-Slip-1 and 7-TS-Slip-2 are 8.6 and

The coordinates of the imaginary frequency in the slippage TSs are the same in the gas and SMD geometries, and their animations reveal motions that are fully characteristic of the assignments made in Figure 13. Attempts to characterize the PES around these TSs in more detail using IRC calculations were not successful as the calculations stopped after a few steps in either one or two directions. This seems to be an artifact that follows from the flat nature of the slippage PESs. In Figure 3, the IRC computed in the gas phase for the α-O−/H slippage of the unsubstituted alkoxide is associated with a gradual increase in the Ru···OCH2R bond (r1) taking place against a gradual decrease in the αCH···Ru bond (r2). Thus, as a possible substitute for the IRC analyses, we calculated two relaxed threedimensional (3D) PESs (in the SMD water continuum), one as a function of the Ru···OH and Ru···O− distances and the other as a function of the Ru···O− and αCH···Ru distances (Figure 15).68 K


Article

Inorganic Chemistry

coupling reactions,69 Yang has proposed mechanisms that give varied roles for free ions, as described in Scheme 2, for example. In this section, we evaluate the possibility of the involvement of free ions by dissociation of 7-ip-OH. As outlined in Scheme 5, Scheme 5. Ion-Pair Dissociation and Reassociation

9.4 kcal/mol, respectively. These values are slightly lower than the already low G‡smd value of the TS of the first coupling step (9.9 kcal/mol for 7-TS-OH). The highest-energy point on the slippage Gsmd profile is for the hydride-transfer TS (13.5 kcal/mol). In Figure 12, the energies of 7-TS-Prt-1 and 7-TS-Prt-2 are 20.7 and 17.6 kcal/mol, respectively. The Gsmd results indicate strongly, therefore, that ion-pair slippage should provide a much more facile path for aldehyde oxidation by 2-Ru−OH than MLC. The thermodynamics of the carbonyl insertion step on the slippage path is slightly endoergic at Gsmd = 3.3 kcal/mol, which changes to Gsmd = −4.4 kcal/mol if the entropies are scaled by 0.5. The latter value is similar to Gsmd value of the separated 2-Ru−H and carboxylic acid products (−4.0 kcal/mol, using scaled entropies). Thus, the insertion intermediate 2-Ru−OCH(OH)R is not a thermodynamic well on the slippage PES. When the gas-phase geometries are used in single-point SMD calculations, 7-TS-Slip-1 (17.5 kcal/mol) and 7-TS-Slip-2 (14.8 kcal/mol) become higher in energy than 7-TS-Hyd (11.7 kcal/mol; G‡smd/gas values in Figure 16). Nevertheless, the slippage TSs on the G‡smd/gas scale remain lower than that of 7-TS-Prt-2 (19.5 kcal/mol), which is the highest-energy point on the MLC PES (Figure 12). Thus, although the highestenergy point on the slippage PES changes depending on how the geometries are optimized, the two sets of data still agree that slippage should be more favorable than MLC in the given H/OH metathesis. Ion-Pair Dissociation. In a number of computational investigations on metal-catalyzed hydrogenation and dehydrogenative

if this occurs, the ions can combine again in different modes. Recombination through the αCH bond would yield the C−Hbound ion pair on the path to hydride transfer. The slightly more favorable recombination through the anionic oxygen produces the insertion product 2-Ru−OCH(OH)R. Dissociation and recombination thus provide a variation on the ion-pairmediated mechanism to H/OH metathesis that skips the slippage TSs. Using geometries optimized in the SMD water continuum, the dissociation of 7-ip-CH is computed to be endothermic by 11.8 kcal/mol (eq 17).

The computed unscaled entropy of dissociation is 43 eu, which leads to a slightly negative free energy of dissociation: ΔGsmd = −1.1 kcal/mol. As mentioned in the Computational and Solvation Methods section, the computed entropies of associative/ dissociative reactions can be larger than the experimental

Figure 16. ΩB97X-D Gibbs free energy profiles for the ion-pair-mediated path to H/OH metathesis from 2-Ru−OH and 6 (at 298 K and 1 atm, in kcal/mol). L


Article

Inorganic Chemistry

of 2-Ru catalyzes the dehydrogenative coupling of alcohols into esters.1a However, no esters were observed in the experiments with 2-Ru in water.21 This might be due to kinetic factors that selectively favor carboxylic acid formation or to initial formation and subsequent hydrolysis of the ester by the base during catalysis under reflux conditions. It becomes of interest to examine whether calculations can discriminate between the two possibilities. Application of either an MLC or a slippage mechanism to ester formation requires an aldehyde and the octahedral ruthenium alkoxide (2-Ru−OR′, R′ = CH2R = nBu). Li and Hall reported the MLC mechanism to follow formation and dehydrogenation of a hemiacetal through two concerted TSs matching TS-MLC1 and TS-MLC-2 described in Scheme 1.25 We calculate the aldehyde to react with 2-Ru−OR′ by a localized electronically barrierless coupling TS that gives a hemiacetaloxide ion pair as a minimum on the PES. The localized coupling step is in line with our previous calculations on ester hydrogenation/ formation by the amino PNN catalyst. In Figure 17, we follow two outer-sphere pathways: an MLC path leading to a hemiacetal and a slippage path leading directly to the ester. To put the energy profiles for ester and carboxylic acid formation on the same scale, the separated aldehyde and 2-Ru−OR′ in Figure 17 are given a value of 0.1 kcal/mol based on eq 20 (see also eqs 8 and 9 in Scheme 4).

entropies measured in solution. When the computed entropies are scaled by 0.5 as recommended in some studies, the dissociation becomes endoergic by 5.4 kcal/mol. A more serious complication in evaluating the entropy of dissociation of 7-ip-OH arises from the possibility of increased organization of the solvent around the separated ions.70 Experimentally, the entropy of ion-pair dissociation is typically either slightly positive (less than 10 eu at 298 K) or, more often, negative (with magnitudes as large as −50 eu).71 This is true in several aprotic and protic solvents, including water. Obviously, even a small solvation entropy term would strongly disfavor dissociation of 7-ip-OH, yet it is not taken into account in conventional electronic structure calculations. Direct slippage from 7-ip-OH to 2-Ru−OCH(OH)R through 7-TS-Slip-1 has a barrier of only 3.4 kcal/mol (ΔG‡smd) and is thermodynamically slightly more exoergic than dissociation even when the entropies are not scaled. Thus, rearrangement through full dissociation and reassociation cannot be expected to be much more favorable than slippage. Intermolecular slippage and intramolecular dissociation are two intimately related modes that are distinguished by different extents of metal−ligand cleavage. If dissociation and recombination is more facile than slippage, it will simply enforce the most important finding in the present study, namely, that hydride transfer is the highest-energy point on the dehydrogenative coupling PES and not ligand deprotonation. Isotope Labeling. Milstein and co-workers designed experiments to rule out a role of residual O2 in catalysis by using 18 O-enriched water as the solvent and unlabeled alcohols.21 Interestingly, these experiments afforded carboxylate products having two labeled oxygen atoms (eq 18).

R′OH + 2‐Ru‐OH ⇌ 2‐Ru‐OR′ + H 2O ΔH = −2.4,

ΔS = −8.6,

ΔGsmd = 0.1

(20)

The coupling step on the MLC route (path a) affords an OR-bound ion pair in which the terminal anionic oxygen of the hemiacetaloxide is in position to abstract a proton from the axial CH bond of the ligand through 10a-TS-Prt-1. The coupling TS on path b is distinguished from that on path a by exchanging the positions of the CH and carbonyl groups of the aldehyde and yields an OR-bound ion pair with the α-CH bond pointed toward the metal at 3.2 Å (10b-ip-OR). 10b-ip-OR thus has the correct stereoconfiguration to be able to slip into the C−H-bound ion pair (10b-ip-CH) in preparation for hydride transfer. The OR- and CH-bound ion pairs have Gsmd values of 3.7 and 1.4 kcal/mol, respectively. Our attempts to locate the α-OR/H slippage TS that takes ip-OR directly to ip-CH gave a TS for an internal rotation around the RO−C bond with ν‡smd = 25i cm−1. To circumvent this problem, we conducted the TS search with the O−C−M angle fixed at 120°. With this constraint, geometry optimization for a TS converges to 10b-TS-Slip-3fix characterized by one imaginary frequency with ν‡smd = 46i cm−1 for the correct α-OR/H slippage motion (Figure 17). 10a-TS-OR and 10a-TS-Prt-1 on the MLC path in Figure 17 have G‡smd = 9.6 and 19.2 kcal/mol, respectively. The coupling TS on the slippage path has G‡smd = 7.8 kcal/mol. The estimated G‡smd value of 10b-TS-Slip-3fix is 6.1 kcal/mol. As was found before, therefore, the highest-energy point on the slippage path is for the hydride-transfer TS (11.9 kcal/mol). Hemiacetal formation from the aldehyde and 2-Ru−OR′ is endoergic by 6.7 kcal/mol. In contrast, formation of the ester and 2-Ru−H from the same reactants is exoergic by 5.7 kcal/mol. For this reaction too, therefore, the calculations predict H/OR metathesis by ion-pair formation and slippage to be kinetically and thermodynamically much more favored than hemiacetal formation by MLC. A related computational result was recently reported for ester hydrogenation/coupling by Gusev’s catalyst.72

Li and Hall proposed the Ggas/smd energy profile of the MLC (matching the one in Figure 12) to be consistent with this observation.25 Because catalysis is performed in basic media under reflux there can be other routes to the observed isotopeenriched products. Most simplistically, in the framework of any mechanism involving alcohol dehydrogenation into aldehyde, a free labeled hydroxide would react with an unlabeled aldehyde to give an aqueous gem-diolate that undergoes a proton rearrangement and reverse fragmentation into a labeled aldehyde (eq 19).

We calculate gem-diolate formation to encounter an electronic barrier of only 1.4 kcal/mol (ΔG‡smd = +1.2) and to be slightly exoergic or slightly endoergic (ΔGsmd = +3.4 or −2.2 kcal/mol) depending on whether the computed value of ΔS for the reaction (−38 eu) is scaled. Coupling between labeled aldehyde and a labeled hydroxide should give carboxylates with two labeled oxygens. MLC and Slippage PESs for Ester Formation. In tetrahydrofuran (THF) solutions, the diethyl amino analogue M


Article

Inorganic Chemistry

Figure 17. Gsmd profiles for MLC and H/OR ion-pair slippage in the reaction of 2-Ru−OR′ and 6 (at 298 K and 1 atm, in kcal/mol). Energies were adjusted to the previous PESs using eq 20.

Thermodynamics of Catalyst Regeneration. For catalysis to proceed, the octahedral dihydride product 2-Ru− H must be transformed back into the octahedral 2-Ru−OR′ and 2-Ru−OH complexes (eqs 22 and 23 in Scheme 6). Each of these reactions is computed to be uphill by about 8 kcal/mol.

All of the data presented so far were calculated for standardstate conditions in which each species has a “concentration” of 1.0 atm. Under these conditions, the MLC and slippage Gibbs free energy profiles for ester formation are both approximately 1.6 kcal/mol lower than the corresponding PESs for carboxylic acid formation. In the actual experiments, the initial concentrations of the alcohol and water are 5.0 and 55 M, respectively. Obviously, the larger concentration of water will speed the relative kinetics of direct carboxylic acid formation by increasing the concentration of 2-Ru−OH over that of 2-Ru−OR′, and this can reverse the computed selectivity. One way to adjust for the different concentrations of water and alcohol73 is illustrated in eq 21 using the two hydride-transfer TSs on the two PESs in the comparison. H 2O [55 M]

Scheme 6. Thermodynamic Data Relevant to Catalyst Regenerationa

+ 10b‐TS‐Hyd ⇌ 7‐TS‐Hyd + R′OH [x M] ‡

‡

ΔΔH = +5.9, ‡

[x M]

ΔΔS = +14eu,

a

[5 M]

At 1 atm and 298 K. Units of kcal/mol.

‡

ΔΔG = +1.6

Mechanistically, eqs 22 and 23 (Scheme 6) can take place through a heterolytic MLC elimination of H2 (reverse of eq 11, Scheme 4) as considered in the study by Li and Hall,25 for example. Because of the strong trans influence of the hydride, the Ru−H bond in 2-Ru−H is expected to be hydridic,57,74 making it a potential reducing agent. A likely alternative to the MLC dissociative path of eqs 22 and 23 (Scheme 6) might therefore be a simple direct reaction within the carboxylic acid complex 8-Plx and between the proton of the water or the alcohol with a hydride from 2-Ru−H to give H2 and 2-Ru−OH or 2-Ru−OR. Consistent with the latter possibility, some octahedral dihydride complexes with pincer ligands that cannot undergo MLC have recently been demonstrated to catalyze aldehyde hydrogenation with H2.75 Also significant, while this study was under review, Buil and co-workers reported formation of a carboxylate and H2 by reaction of an aldehyde and a hydrido osmium hydroxide complex that cannot undergo MLC.76 Finally, the carboxylate product from carboxylic acid deprotonation can compete for the catalyst. As mentioned before, catalysis in this system is performed with excess hydroxide. Equation 24 in Scheme 6 predicts coordination of the hydroxide (catalyst regeneration) to the metal to be slightly more favored than coordination of a carboxylate.

‡

ΔΔG′ = ΔΔG + RT ln(5/55) = +1.6 − 1.4 = +0.2 kcal/mol

(21) ‡

ΔΔG = 1.6 kcal/mol in eq 21 is the difference between the G‡ values of 10b-TS-Hyd and 7-TS-Hyd computed at 1 atm (Figure 17). To adjust to a concentration of 5.0 M alcohol in water, a term of RT ln(5/55) is added to ΔΔG‡, which affords ΔΔG′‡ = +0.2 kcal/mol. In the study by Li and Hall,25 the adjustment was made at an alcohol concentration of 1.0 M. Our calculated ΔΔG′‡ value for the hydride-transfer TSs in this case is −0.8 kcal/mol. Taken at face value, these results suggest that the rates of carboxylic acid and ester formation should be competitive in water. Note that this is true for both the slippage and MLC mechanisms. However, the computed ΔΔG‡ value in eq 21 is associated with a large entropy term (ΔΔS‡ = +14 eu; computed using statistical mechanics in the gas phase), which seems unrealistic for an isodesmic reaction in solution. If the given entropy value is scaled by 0.5, the ΔΔG′‡ values at R′OH concentrations of 5.0 and 1.0 M will be 2.3 and 1.4 kcal/mol, respectively (in favor of the ester). At the same time, the solvent adjustment method in eq 21 is too simplistic, as it assumes, for example, ideal-gas behavior and equal concentrations for the two TSs, so the adjustments are not expected to be quantitatively accurate. Thus, given the small energy differences being compared, it is not possible to draw a definitive conclusion on whether the absence of esters in the experiments is due to kinetic or thermodynamic reasons.

■

CONCLUSIONS In basic media, the square-pyramidal 1-Ru catalyzes the dehydrogenative coupling of alcohols and water into carboxylates N


Article

Inorganic Chemistry and H2.21 In the present work, we use DFT to address possible mechanisms for two key reactions implicated in catalysis: (i) alcohol dehydrogenation into an aldehyde and (ii) oxidation of the aldehyde into acid with water as the oxidant. Complexes related to 1-Ru undergo characteristic heterolytic MLC addition of alcohols and water, so we take the respective octahedral 2-Ru−OCH2R and 2-Ru−OH complexes as the starting point of study (Scheme 7).

Scheme 8. Carboxylic Acid Formation from 2-Ru−OCH2R and 2-Ru−OH

Scheme 7. Carboxylic Acid Formation from 2-Ru−OCH2R and 2-Ru−OH

full PES is 17.0 kcal/mol for 7-TS-Hyd (at 1 atm and 298 K; based on unscaled entropies). This means that the slippage mechanism could be chemically relevant even if the concentrations of the octahedral reactants were too small under the complex conditions of catalysis, which can include, for example, ruthenium carboxylate species. If it occurred, dissociation of 7-ip-OH into free ions followed by reassociation would provide a variation on the ion-pair mechanism that continues to have 7-TS-Hyd as the highest-energy point. The generally postulated metal−ligand cooperation mechanism has to proceed from 7-ip-OH through 7-TS-Prt-1 to give a gem-diol that will still need to undergo a dehydrogenation. However, the energy of 7-TS-Prt-1 is 24.2 kcal/mol, 7.2 kcal/mol above that of 7-TSHyd. This large preference for slippage over MLC is also computed for ester formation by reaction of the aldehyde and 2-Ru−OCH2R. Experimentally, no esters have been observed under the reaction conditions, which include excess base. Because of the simplistic treatment of the entropy and concentration in the calculations, it is not possible to make a definitive conclusion on whether this follows from kinetic factors or whether the esters are initially formed but subsequently hydrolyzed. An important finding of the present study is that the gas-phase geometries of the ion-pair TSs and minima in the reactions of Milstein-type catalysts can change substantially when they are optimized in the SMD water continuum. Nonetheless, the computed greater preference for ion-pair slippage over MLC is independent of the choice of the gas-phase- or SMD-optimized geometries, but it is greatly amplified in the latter case. The conclusions of the present study are also independent of the choice of the density functional (ωB97X-D, M06, and M06L).

From 2-Ru−OCH2R, we calculate a β-hydride elimination taking place in an outer-sphere mode that directly eliminates an aldehyde from the octahedral alkoxide instead of first giving a metal-coordinated aldehyde. We find the outer-sphere reaction between the aldehyde product and the octahedral 2-Ru−OH to begin with a localized coupling step leading to an ion-pair minimum in which an α-hydroxyl-alkoxide (gem-diolate anion) is coordinated to the metal through the hydroxyl group (7-ip-OH). A trivial slippage within the intact ion pair generates a new octahedral ruthenium alkoxide [2-Ru−OCH(OH)R]. A second outer-sphere β-hydride elimination from 2-Ru− OCH(OH)R gives the carboxylic acid product. The sequence of aldehyde insertion into the Ru−OH bond and carboxylic acid deinsertion from 2-Ru−OCH(OH)R achieves a H/OH metathesis in which a hydride and a hydroxide are exchanged between a metal center and an acyl group. A 3D PES reveals that the metathesis can also proceed from the OH-bound ion pair more directly through an α-OH/H slippage that orients the CH bond of the gem-diolate toward the metal in preparation for hydride transfer without the intermediacy of 2-Ru− OCH(OH)R (Figure 15 and eq 16). The slippage TSs that prepare the alkoxides for hydride transfer to the metal are all computed to be low in energy, a result that we attribute to two factors: (i) the strong trans influence of the “spectator” hydride and (ii) unique geometrical features of the bipyridyl PNN ligand that create an opening over the metal center allowing the alkoxides to rearrange without steric hindrance. In this mechanism, it is hydride transfer from the organic substrate to the metal that marks the high-energy points on the PESs. The abridged Gibbs free energy profiles summarized in Scheme 8 show aldehyde deinsertion from 2-Ru−OCH2R to be uphill by 3.5 kcal/mol and to encounter a barrier of 13.9 kcal/mol. In contrast, carboxylic acid formation from 2-Ru−OH and the aldehyde is thermodynamically favorable by 4.1 kcal/mol, although it too encounters a barrier of 13.5 kcal/mol. Thus, the net reaction originating from the pair of ruthenium hydroxide and alkoxide octahedral complexes is only slightly exoergic (−0.6 kcal/mol), and the highest-energy point on the

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b00766. Tables of the Cartesian coordinates and energies of the geometries optimized in the SMD water continuum (PDF) mol2.txt file that allows (after the txt extension is dropped) convenient direct visualization of all of the molecules in the figures using GaussView or Mercury, free software available from the Cambridge Crystallographic Data Center (TXT) O


Article

Inorganic Chemistry

■

(17) (a) Qu, S.; Dang, Y.; Song, C.; Wen, M.; Huang, K.-W.; Wang, Z.-X. J. Am. Chem. Soc. 2014, 136, 4974−4991. (b) Li, H.; Wang, X.; Huang, F.; Lu, G.; Jiang, J.; Wang, Z.-X. Organometallics 2011, 30, 5233−5247. (18) (a) Li, J.; Shiota, Y.; Yoshizawa, K. J. Am. Chem. Soc. 2009, 131, 13584−13585. (b) Zeng, G.; Li, S. Inorg. Chem. 2011, 50, 10572− 10580. (c) Cho, D.; Ko, K. C.; Lee, J. Y. Organometallics 2013, 32, 4571−4576. (d) Tomasi, J.; Mennucci, B.; Cances, E. J. Mol. Struct.: THEOCHEM 1999, 464, 211−226. (19) (a) Sandhya, K. S.; Suresh, C. H. Organometallics 2013, 32, 2926−2933. (b) Yang, X. ACS Catal. 2012, 2, 964−970. (20) (a) Cantillo, D. Eur. J. Inorg. Chem. 2011, 2011 (19), 3008− 3013. (b) Filonenko, G. A.; Conley, M. P.; Copeŕet, C.; Lutz, M.; Hensen, E. J. M.; Pidko, E. A. ACS Catal. 2013, 3, 2522−2526. (c) Filonenko, G. A.; Hensen, E. J. M.; Pidko, E. A. Catal. Sci. Technol. 2014, 4, 3474−3485. (21) Balaraman, E.; Khaskin, E.; Leitus, G.; Milstein, D. Nat. Chem. 2013, 5, 122−125. (22) Hu, P.; Diskin-Posner, Y.; Ben-David, Y.; Milstein, D. ACS Catal. 2014, 4, 2649−2652. (23) (a) Navarro, R. M.; Peña, M. A.; Fierro, J. L. G. Chem. Rev. 2007, 107, 3952−3991. (b) Palo, D. R.; Dagle, R. A.; Holladay, J. D. Chem. Rev. 2007, 107, 3992−4021. (c) Cortright, R. D.; Davda, R. R.; Dumesic, J. A. Nature 2002, 418, 964−967. (d) Olah, G.; Prakash, G. K. S.; Goeppert, A. J. Am. Chem. Soc. 2011, 133, 12881−12898. (e) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39, 81−88. (f) Grasemann, M.; Laurenczy, G. Energy Environ. Sci. 2012, 5, 8171−8181. (24) Nielsen, M.; Alberico, E.; Baumann, W.; Drexler, H.-J.; Junge, H.; Gladiali, S.; Beller, M. Nature 2013, 495, 85−89. (25) Li, H.; Hall, M. B. J. Am. Chem. Soc. 2014, 136, 383−395. (26) Kohl, S. W.; Weiner, L.; Schwartsburd, L.; Konstantinovski, L.; Shimon, L. J. W.; Ben-David, Y.; Iron, M. A.; Milstein, D. Science 2009, 324, 74−77. (27) Spasyuk, D.; Smith, S.; Gusev, D. Angew. Chem., Int. Ed. 2013, 52, 2538−2542. (28) Spasyuk, D.; Vicent, C.; Gusev, D. J. Am. Chem. Soc. 2015, 137, 3743−3746. (29) Miller, A. J. M.; Heinekey, D. M.; Mayer, J. M.; Goldberg, K. I. Angew. Chem., Int. Ed. 2013, 52, 3981−3984. (30) Brewster, T. P.; Ou, W. C.; Tran, J. C.; Goldberg, K. I.; Hanson, S. K.; Cundari, T. R.; Heinekey, D. M. ACS Catal. 2014, 4, 3034− 3038. (31) Ou, W. C.; Cundari, T. R. ACS Catal. 2015, 5, 225−232. (32) Yang, X. ACS Catal. 2014, 4, 1129−1133. (33) Hasanayn, F.; Baroudi, A. Organometallics 2013, 32, 2493−2496. (34) Hasanayn, F.; Harb, H. Inorg. Chem. 2014, 53, 8334−8349. (35) Hasanayn, F.; Baroudi, A.; Bengali, A.; Goldman, A. S. Organometallics 2013, 32, 6969−6985. (36) 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.; Keith, T.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; 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 09, revision A.02; Gaussian, Inc.: Wallingford, CT, 2009. (37) Marenich, A. V.; Cramer, C. J.; Truhlar, C. J. J. Phys. Chem. B 2009, 113, 6378−6396. (38) Keith, J. A.; Carter, E. A. J. Chem. Theory Comput. 2012, 8, 3187−3206.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the University Research Board of the American University of Beirut (AUB) and by the Lebanese National Council for Scientific Research. The HPC team at AUB is thanked for system administration. Professor Mazen Al-Ghoul is thanked for help with MatLab.

■

REFERENCES

(1) (a) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. J. Am. Chem. Soc. 2005, 127, 10840−10841. (b) Zhang, J.; Balaraman, E.; Leitus, G.; Milstein, D. Organometallics 2011, 30, 5716−5724. (2) (a) Gunanathan, C.; Ben-David, Y.; Milstein, D. Science 2007, 317, 790−792. (b) Gunanathan, C.; Milstein, D. Angew. Chem., Int. Ed. 2008, 47, 8661−8664. (c) Srimani, D.; Balaraman, E.; Hu, P.; BenDavid, Y.; Milstein, D. Adv. Synth. Catal. 2013, 355, 2525−2530. (3) (a) Zeng, H.; Guan, Z. J. Am. Chem. Soc. 2011, 133, 1159−1161. (b) Gnanaprakasam, B.; Balaraman, E.; Gunanathan, C.; Milstein, D. J. Polym. Sci., Part A: Polym. Chem. 2012, 50, 1755−1765. (4) (a) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2006, 45, 1113−1115. (b) Zell, T.; Ben-David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2014, 53, 4685−4689. (5) (a) Balaraman, E.; Gnanaprakasam, B.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc. 2010, 132, 16756−16758. (b) Balaraman, E.; Gunanathan, C.; Zhang, J.; Shimon, L. J. W.; Milstein, D. Nat. Chem. 2011, 3, 609−614. (c) Balaraman, E.; Ben-David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 11702−11705. (6) Gnanaprakasam, B.; Ben-David, Y.; Milstein, D. Adv. Synth. Catal. 2010, 352, 3169−3173. (7) Gnanaprakasam, B.; Milstein, D. J. Am. Chem. Soc. 2011, 133, 1682−1685. (8) (a) Milstein, D. Top. Catal. 2010, 53, 915−923. (b) Gunanathan, C.; Milstein, D. Top. Organomet. Chem. 2011, 37, 55−84. (9) (a) Ito, M.; Ootsuka, T.; Watari, R.; Shiibashi, A.; Himizu, A.; Ikariya, T. J. Am. Chem. Soc. 2011, 133, 4240−4242. (b) Dub, P. A.; Ikariya, T. ACS Catal. 2012, 2, 1718−1741. (c) Clarke, M. L. Catal. Sci. Technol. 2012, 2, 2418−2423. (10) (a) Balaraman, E.; Milstein, D. Top. Organomet. Chem. 2014, 48, 19−44. (b) Alsabeh, P. G.; Mellmann, D.; Junge, H.; Beller, M. Top. Organomet. Chem. 2014, 48, 45−79. (c) Gunanathan, C.; Milstein, D. Chem. Rev. 2014, 114, 12024−12087. (11) (a) Pritchard, J.; Filonenko, G. A.; van Putten, R.; Hensen, E. J. M.; Pidko, E. A. Chem. Soc. Rev. 2015, 44, 3808−3833. (b) Younus, H. A.; Ahmad, N.; Su, W.; Verpoort, F. Coord. Chem. Rev. 2014, 276, 112−152. (c) Younus, H. A.; Su, W.; Ahmad, N.; Chen, S.; Verpoort, F. Adv. Synth. Catal. 2015, 357, 283−330. (d) Alberico, E.; Nielsen, M. Chem. Commun. 2015, 51, 6714−6725. (12) Werkmeister, S.; Junge, K.; Beller, M. Org. Process Res. Dev. 2014, 18, 289−302. (13) Montag, M.; Zhang, J.; Milstein, D. J. Am. Chem. Soc. 2012, 134, 10325−10328. (14) Khaskin, E.; Iron, M. A.; Shimon, L. J. W.; Zhang, J.; Milstein, D. J. Am. Chem. Soc. 2010, 132, 8542−8543. (15) (a) Gunanathan, C.; Milstein, D. Acc. Chem. Res. 2011, 44, 588− 602. (b) Gunanathan, C.; Milstein, D. Science 2013, 341, 1229712. (c) Khusnutdinova, J. R.; Milstein, D. Angew. Chem., Int. Ed. 2015, 54, 12236−12273. (16) (a) Schwartsburd, L.; Iron, M. A.; Konstantinovski, L.; DiskinPosner, Y.; Leitus, G.; Shimon, L. J. W.; Milstein, D. Organometallics 2010, 29, 3817−3827. (b) Iron, M. A.; Ben-Ari, E.; Cohen, R.; Milstein, D. Dalton Trans. 2009, 9433−9439. P


Article

Inorganic Chemistry (39) (a) Zhao, Y.; Truhlar, D. G. Chem. Phys. Lett. 2011, 502, 1−13. (b) Marenich, A. V.; Ho, J.; Coote, M. L.; Cramer, C. J.; Truhlar, D. G. Phys. Chem. Chem. Phys. 2014, 16, 15068−15106. (c) Zanith, C. C.; Pliego, J. R., Jr. J. Comput.-Aided Mol. Des. 2015, 29, 217−224. (d) Ho, J. Phys. Chem. Chem. Phys. 2015, 17, 2859−2868. (40) (a) Sperger, T.; Sanhueza, I. A.; Kalvet, I.; Schoenebeck, F. Chem. Rev. 2015, 115, 9532−9586. (b) Gusev, D. G. Organometallics 2013, 32, 4239−4243. (41) (a) Cances, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032−3041. (b) Mennucci, B.; Cances, E.; Tomasi, J. J. Phys. Chem. B 1997, 101, 10506−10517. (42) Chai, J.-D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615−6620. (43) (a) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639−5648. (b) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650−654. (c) Rassolov, V. A.; Ratner, M. A.; Pople, J. A.; Redfern, P. C.; Curtiss, L. A. J. Comput. Chem. 2001, 22, 976−984. (44) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990, 77, 123−141. (45) Ehlers, A. W.; Böhme, M.; Dapprich, S.; Gobbi, A.; Höllwarth, A.; Jonas, V.; Köhler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys. Lett. 1993, 208, 111−114. (46) EMSL Basis Set Exchange. https:/bse.pnl.gov/bse/portal, accessed May 2014. (47) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101. (48) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241. (49) Hehre, J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular Orbital Theory; Wiley: New York, 1986. (50) Abu-Hasanayn, F.; Goldman, A. S.; Krogh-Jespersen, K. J. Phys. Chem. 1993, 97, 5890−5896. (51) (a) Baratta, W.; Chelucci, G.; Gladiali, S.; Siega, K.; Toniutti, M.; Zanette, M.; Zangrando, E.; Rigo, P. Angew. Chem., Int. Ed. 2005, 44, 6214−6219. (b) Baratta, W.; Bosco, M.; Chelucci, G.; Del Zotto, A.; Siega, K.; Toniutti, M.; Zangrando, E.; Rigo, P. Organometallics 2006, 25, 4611−4620. (c) Baratta, W.; Ballico, M.; Esposito, G.; Rigo, P. Chem. - Eur. J. 2008, 14, 5588−5595. (52) (a) Hamilton, R. J.; Leong, C. G.; Bigam, G.; Miskolzie, M.; Bergens, S. H. J. Am. Chem. Soc. 2005, 127, 4152−4153. (b) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2006, 128, 13700−13701. (c) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2008, 130, 11979−11987. (d) Takebayashi, S.; Dabral, N.; Miskolzie, M.; Bergens, S. H. J. Am. Chem. Soc. 2011, 133, 9666−9669. (53) (a) Bertoli, M.; Choualeb, A.; Lough, A. J.; Moore, B.; Spasyuk, D.; Gusev, D. G. Organometallics 2011, 30, 3479−3482. (b) Spasyuk, D.; Gusev, D. G. Organometallics 2012, 31, 5239−5242. (54) Takebayashi, S.; Bergens, S. H. Organometallics 2009, 28, 2349− 2351. (55) (a) Darensbourg, D. J.; Rokicki, A.; Darensbourg, M. Y. J. Am. Chem. Soc. 1981, 103, 3223−3224. (b) Darensbourg, D. J.; Kudaroski Hanckel, R.; Bauch, C. G.; Pala, M.; Simmons, D.; White, J. N. J. Am. Chem. Soc. 1985, 107, 7463−7473. (c) Darensbourg, D. J.; Kyran, S. J.; Yeung, A. D.; Bengali, A. A. Eur. J. Inorg. Chem. 2013, 2013, 4024− 4031. (d) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. Organometallics 1996, 15, 5166−5169. (56) (a) Salem, H.; Shimon, L. J. W.; Diskin-Posner, Y.; Leitus, G.; Ben-David, Y.; Milstein, D. Organometallics 2009, 28, 4791−4806. (b) Langer, R.; Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; BenDavid, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 9948−9952. (57) Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.; Hazari, N. J. Am. Chem. Soc. 2011, 133, 9274−9277. (58) Ramakrishnan, S.; Waldie, K. M.; Warnke, I.; De Crisci, A. G.; Batista, V. S.; Waymouth, R. M.; Chidsey, C. E. D. Inorg. Chem. 2016, 55, 1623−1632. (59) The other conformer in which the C−O bond of the alkoxide is aligned along the Ru−CO bond or along the central Ru−N bond has energies similar to those of the conformer used in Figure 1. Preliminary calculations suggest that there is no significant barrier for interchange among these conformers.

(60) (a) Hasanayn, F.; Abu-El-Ez, D. Inorg. Chem. 2010, 49, 9162− 9168. (b) Baroudi, A.; El-Hellani, A.; Bengali, A.; Goldman, A. S.; Hasanayn, F. Inorg. Chem. 2014, 53, 12348−12359. (61) This is the same path as considered in the study by Li and Hall.25 The energies of the key points on the diastereomeric path defined by bringing the aldehyde from the other face are slightly higher than the values presented in the text. (62) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122, 1466−1478. (63) Dub, P. A.; Ikariya, T. J. Am. Chem. Soc. 2013, 135, 2604−2619. (64) Hasanayn, F.; Morris, R. H. Inorg. Chem. 2012, 51, 10808− 10818. (65) Dub, P. A.; Henson, N. J.; Martin, R. L.; Gordon, J. C. J. Am. Chem. Soc. 2014, 136, 3505−3521. (66) A reviewer of this article noted that the pKhydration value of butanal is slightly positive (0.3, see ref 67) meaning that the gem-diol intermediate, if formed as proposed in the MLC mechanism, might, under basic conditions, spontaneously fragment into an aldehyde and water before it can react again with 1-Ru. (67) Hilal, S. H.; Bornander, L. L.; Carreira, L. A. QSAR Comb. Sci. 2005, 24, 631−638. (68) Each surface required at least 100 geometry optimizations. For practical consideration, the geometries in these calculations were optimized in a water continuum using the 6-31G(d,p) basis set in place of the 6-31++G(d,p) basis set used in all other optimizations. The actual energies used in the plots were then obtained by singlepoint calculations employing the 6-311++G(2d,2p) basis set. (69) (a) Yang, X. Inorg. Chem. 2011, 50, 12836−12843. (b) Yang, X. ACS Catal. 2012, 2, 964−970. (c) Yang, X. ACS Catal. 2013, 3, 2684− 2688. (70) (a) Marcus, Y.; Hefter, G. Chem. Rev. 2006, 106, 4585−4621. (b) Marcus, Y. J. Phys. Chem. B 2007, 111, 572−580. (71) (a) Boruń, A.; Bald, A. J. Chem. Eng. Data 2012, 57, 475−481. (b) Boruń, A.; Bald, A. J. Chem. Eng. Data 2012, 57, 2037−2043. (72) Chen, X.; Jing, Y.; Yang, X. Chem. - Eur. J. 2016, 22, 1950−1957. (73) Cramer, C. J. Essentials of Computational Chemistry: Theories and Models, 2nd ed.; John Wiley & Sons: New York, 2004; pp 378−380. (74) Basolo, F.; Pearson, R. G. Prog. Inorg. Chem. 1962, 4, 381−453. (75) (a) Mazza, S.; Scopelliti, R.; Hu, X. Organometallics 2015, 34, 1538−1545. (b) Gorgas, N.; Stöger, B.; Veiros, L. F.; Kirchner, K. ACS Catal. 2016, 6, 2664−2672. (76) Buil, M. L.; Cardo, J. J. F.; Esteruelas, M. A.; Oñate, E. Organometallics 2016, 35, 2171−2173.

Q


Mechanism of AlcoholâWater Dehydrogenative Coupling into

Recommend Documents

Mechanism of AlcoholâWater Dehydrogenative Coupling into