Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jun 5.
Published in final edited form as: J Am Chem Soc. 2013 May 24;135(22):8205–8212. doi: 10.1021/ja4031648

Aerobic Dehydrogenation of Cyclohexanone to Cyclohexenone Catalyzed by Pd(DMSO)2(TFA)2: Evidence for Ligand-Controlled Chemoselectivity

Tianning Diao 1, Doris Pun 1, Shannon S Stahl 1,*
PMCID: PMC3795849  NIHMSID: NIHMS481927  PMID: 23662700

Abstract

The dehydrogenation of cyclohexanones affords cyclohexenones or phenols via removal of one or two equivalents of H2, respectively. We recently reported several PdII catalyst systems that effect aerobic dehydrogenation of cyclohexanones with different product selectivities. Pd(DMSO)2(TFA)2 is unique in its high chemoselectivity for the conversion of cyclohexanones to cyclohexenones, without promoting subsequent dehydrogenation of cyclohexenones to phenols. Kinetic and mechanistic studies of these reactions reveal the key role of the dimethylsulfoxide (DMSO) ligand in controlling this chemoselectivity. DMSO has minimal kinetic influence on the rate of Pd(TFA)2-catalyzed dehydrogenation of cyclohexanone to cyclohexenone, while it strongly inhibits the second dehydrogenation step, conversion of cyclohexenone to phenol. These contrasting kinetic effects of DMSO provide the basis for chemoselective formation cyclohexenones.

Introduction

Dehydrogenation of saturated C–C bonds represents an important class of C–H functionalization reactions. Homogeneous catalysts, such as iridium/PCP-pincer complexes, have been investigated extensively for such transformations, and they have proven to be quite effective in the conversion of alkanes to alkenes or arenes.1,2 These reactions are performed with a sacrificial H2 acceptor, such as tert-butylethylene, or under “acceptorless” conditions, in which H2 is physically removed from the reaction mixture. Oxidative dehydrogenation, using O2 as the hydrogen acceptor, represents an appealing alternative method to introduce sites of unsaturation into organic molecules. The majority of precedents in this area, however, feature high-temperature, gas-phase conditions for commodity chemical applications. Prominent targets include the conversion of ethane to ethylene, propane to propylene or ethylbenzene to styrene.3 The reaction methods and conditions for these transformations are unsuitable for use with fine-chemicals, pharmaceuticals, or related molecules bearing diverse functional groups. Homogeneous PdII catalysts could find utililty in such applications; however, precedents are quite limited.4–6 Until recently, little effort has been made to expand the scope and synthetic utility of such reactions.7

We recently reported three different catalyst systems that promote aerobic dehydrogenation of cyclohexanones and other carbonyl compounds (Chart 1). A Pd(TFA)2/2-Me2Npy (TFA = trifluoroacetate, 2-Me2Npy = 2-dimethylaminopyridine) catalyst system exhibits good activity for the dehydrogenation of cyclohexanones and cyclohexenones to phenols.8 A Pd(TFA)2/4,5-diazafluorenone catalyst exhibits similar reactivity, but is especially useful for α,β-dehydrogenation of heterocyclic carbonyl compounds and exhibits some success in the dehydrogenation of acyclic carbonyl compounds.9, 10 Finally, a Pd(DMSO)2(TFA)2 catalyst enables selective dehydrogenation of cyclohexanones to cyclohexenones. 11 The reactions mediated by these catalysts provide compelling routes to substituted phenols and/or enones, and they serve as an important foundation for the development of other aerobic dehydrogenation methods. A number of related transformations, including reactions for the synthesis of aryl ethers and anilines from cyclohexanones and cyclohexenones, have been reported by other groups over the past year.12

Chart 1.

Chart 1

Open in a new tab

Pd Catalyst Systems for Aerobic Dehydrogenation of Carbonyl Compounds.

Further development of oxidative dehydrogenation reactions of this type would benefit from mechanistic insights. The catalytic dehydrogenation of a cyclohexanone could lead to cyclohexenone and/or phenol products, and, in the reaction of unsubstituted cyclohexanone, the three catalyst systems in Chart 1 exhibit different selectivity patterns (Scheme 1). The first two catalysts systems, Pd(TFA)2/2-Me2Npy and Pd(TFA)2/diazafluorenone, favor formation of phenol, whereas Pd(DMSO)2(TFA)2 promotes highly selective formation of cyclohexenone. Kinetic modeling of the reaction time courses has been used to obtain relative rate constants for the first and second dehydrogenation steps with each of these catalyst systems (Scheme 1B).13 The results show that the first two catalyst systems promote dehydrogenation of cyclohexenone to phenol more rapidly than cyclohexanone to cyclohexenone (i.e., k1/k2 < 1; Scheme 1B). The opposite trend is observed with Pd(DMSO)2(TFA)2 as the catalyst (k1/k2 = 10–33, depending on the solvent14). These kinetic differences have important synthetic implications, as the k1/k2 ratio with Pd(DMSO)2(TFA)2 is sufficiently high that excellent yields of cyclohexenone products can be obtained for a wide range of substrates, with minimal phenol byproducts.11

Scheme 1.

Scheme 1

Open in a new tab

Relative Rate Constants for Different Catalyst Systems in the Dehydrogenation of Cyclohexanone.

A plausible catalytic cycle for dehydrogenation of cyclohexanone derivatives involves formation of a PdII-enolate followed by β-hydride elimination (steps 1 and 2, Scheme 2). This sequence is analogous to that proposed for Saegusa-type dehydrosilylation of silyl enol ethers.15 Oxidation of the PdII–hydride intermediate is expected to proceed via Pd0, as has been demonstrated for well-defined model systems (steps 3–5).16 A similar catalytic cycle can be proposed for the dehydrogenation of cyclohexenone to phenol.

Scheme 2.

Scheme 2

Open in a new tab

Proposed Mechanism for Pd-Catalyzed Dehydrogenation of Cyclohexanones

The simplified analysis above does not account for the different selectivity patterns observed with the different catalyst systems in Scheme 1. Here, we present a kinetic and mechanistic investigation of Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone to cyclohexenone, and this work sheds light on three key mechanistic issues: (1) the identity of the turnover-limiting step; (2) the role of ligand in the catalyst system; (3) the origin of the chemoselectivity for enone relative to phenol. The results show that DMSO serves as a ligand that stabilizes a homogeneous PdII catalyst and that inhibits conversion of cyclohexenone to phenol. These results complement a companion study of Pd(TFA)2/2-Me2Npy-catalyzed dehydrogenation of cyclohexanone, in which the PdII catalyst is found to undergo in situ conversion into Pd nanoparticles that facilitate full dehydrogenation of cyclohexanone to phenol.17

Results and Discussion

Kinetic Studies of Pd(DMSO)2(TFA)2-Catalyzed Oxidation of Cyclohexanone

The catalytic dehydrogenation of cyclohexanone with Pd(DMSO)2(TFA)2 in EtOAc (eq 1) proceeds smoothly and reaches complete conversion after approx 24 h at 60 °C under 1 atm O2 (Figure 1). The time course data fit well to an A → B → C kinetic model (Figure 1). Good mass balance reveals that negligible side reactions occur. The rate constants for the first (k1) and second (k2) oxidation steps are estimated to be 0.13 h−1 and 0.013 h−1, respectively, resulting in a ratio of k1/k2 = 10. Higher selectivity can be achieved in other solvents (e.g., k1/k2 = 33 in AcOH), but use of the lower-selectivity solvent EtOAc was chosen for the present studies to facilitate study of both dehydrogenation steps.

Figure 1.

Figure 1

Open in a new tab

Reaction time course of Pd(DMSO)2(TFA)2-catalyzed aerobic dehydrogenation of cyclohexanone. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: [cyclohexanone] = 0.8 M (0.8 mmol), [Pd(TFA)2] = 0.04 M (0.04 mmol), [DMSO] = 0.08 M (0.08 mmol), O2 (1 atm), EtOAc (1 mL), 60 °C.

graphic file with name nihms481927e1.jpg (1)

The lack of an induction period in the reaction time course for Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone allowed us to employ initial-rates methods to determine the kinetic orders of each reaction component. The conversion of cyclohexenone to phenol is negligible within the first two turnovers, and the concentration of cyclohexenone was monitored by gas chromatography to determine the initial rates. The reaction exhibits a first order dependence on [cyclohexanone] and [Pd(DMSO)2(TFA)2] (Figure 2A and 2B, respectively). Systematic analysis of the O2-pressure dependence was complicated by Pd black formation at lower pO2; however, increasing the O2 pressure from 1.7 to 3.2 atm had minimal impact on the rate, consistent with a zero-order dependence on O2 (Figure S1).

Figure 2.

Figure 2

Open in a new tab

Kinetic orders for Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone to cyclohexenone: dependence of the initial rate on (A) substrate concentration (B) Pd(DMSO)2(TFA)2 concentration. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: EtOAc (1 mL), O2 (1 atm), 60 °C; (A) [Pd(TFA)2] = 0.01 M (0.01 mmol), [DMSO] = 0.02 M (0.02 mmol); (B) [cyclohexanone] = 0.2 M (0.2 mmol).

Variation of [DMSO] at fixed [cyclohexanone] and [Pd(TFA)2] reveals that increasing the quantity of DMSO has only a minor inhibitory effect on the reaction rate (Figure 3A). Comparison of the time courses in the absence and presence of 2 equiv DMSO, however, reveals that the catalyst undergoes rapid deactivation in the absence of DMSO (Figure 3B). Concomitant formation of Pd black is observed under these ligandless conditions. Minimal Pd black formation is observed with ≥2 equiv DMSO.

Figure 3.

Figure 3

Open in a new tab

(A) Effect of the DMSO ligand on Pd-catalyzed dehydrogenation of cyclohexanone to cyclohexenone evident from the dependence of the initial rate on the DMSO/Pd ratio. (B) Comparison of time courses for the dehydrogenation of cyclohexanone in the absence and presence of DMSO. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: EtOAc (1 mL), O2 (1 atm), 60 °C; (A) [cyclohexanone] = 0.2 M (0.2 mmol), [Pd(TFA)2] = 0.01 M (0.01 mmol); (B) [cyclohexanone] = 0.8 M (0.8 mmol), [Pd(TFA)2] = 0.04 M (0.04 mmol), (DMSO = 0.08 M (0.08 mmol) (blue)).

Deuterium kinetic isotope effects (KIE) were determined by independent measurements of the initial rates with the protio, α-deuterated (cyclohexanone-d4) and fully deuterated (cyclohexanone-d10) substrates (Figure 4). Cyclohexanone reacts 2.9 (± 0.27) times faster than cyclohexanone-d4, reflecting a primary KIE for cleavage of the α-C–H. The reaction time courses for cyclohexanone-d4 and cyclohexanone-d10 are nearly identical, and the ratio of the rates, 1.1 (± 0.20), reflects a negligible KIE for cleavage of the β-hydrogen atom (Figure 4). GC-MS and 1H NMR spectroscopic analyses reveal that no proton incorporation takes place into the α position under the reaction conditions. The intrinsic KIE for α-C–H cleavage was obtained from an intramolecular competition experiment with spiro[4,5]decan-6-one-7-d1 (eq 2). The corresponding enone is formed slowly as the sole product. After 24 h, a 13 % yield of enone is obtained, with the H/D products obtained in a ratio of 1:2.7, corresponding to a nearly identical KIE relative to that obtained from the independent rate measurements.

Figure 4.

Figure 4

Open in a new tab

Deuterium kinetic isotope effects (KIE) derived from independent initial rates measurements for dehydrogenation of cyclohexanone (performed in triplicate). The plots are labeled with the corresponding cyclohexanone-dn substrate. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: [Substrate] = 0.2 M (0.2 mmol), [Pd(TFA)2] = 0.01 M (0.01 mmol), [DMSO] = 0.02 M (0.02 mmol), EtOAc (1 mL), O2 (1 atm), 60 °C.

graphic file with name nihms481927e2.jpg (2)

Replacement of Pd(TFA)2 with Pd(OAc)2 leads to reduced initial rate (Figure 5). When a mixture of Pd(TFA)2 and Pd(OAc)2 is used as the Pd source, maintaining the total [Pd] at 5 mol %, increased initial rates are observed. The maximum rate was observed at an acetate:trifluoroacetate ratio of 1:1.

Figure 5.

Figure 5

Open in a new tab

Dependence of the initial rate of Pd-catalyzed oxidation of cyclohexanone on different anionic ligands of Pd. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: [cyclohexanone] = 0.2 M (0.2 mmol), [Pd]total = 0.01 M (0.01 mmol), [DMSO] = 0.02 M (0.02 mmol), EtOAc (1 mL), O2 (1 atm), 60 °C.

Pd(DMSO)2(TFA)2-Catalyzed Dehydrogenation of Cyclohexenone to Phenol

Cyclohexenone-to-phenol dehydrogenation proceeds very slowly under the conditions of Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanones (cf. Scheme 1), but independent insights into cyclohexenone reactivity could be obtained by increasing the catalyst loading from 5 to 10 mol % (eq 3), and increasing the substrate concentration from 0.2 to 0.8 M. In contrast to the well-behaved kinetics observed for cyclohexanone-to-cyclohexenone dehydrogenation, the cyclohexenone-to-phenol reaction exhibits an induction period and a sigmoidal time course for formation of phenol (Figures 6 and S6). Control reactions with 10 mol % phenol or H2O added to the initial reaction mixture exhibit identical rates (Figure S8), indicating that the sigmoidal time course does not arise from product-induced autocatalysis.

Figure 6.

Figure 6

Open in a new tab

Time courses of Pd(TFA)2-catalyzed dehydrogenation of cyclohexenone to phenol in the presence (blue squares) and absence (black circles) of DMSO. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: [cyclohexenone] = 0.8 M (0.8 mmol), [Pd(TFA)2] = 0.08 M (0.08 mmol), ([DMSO] = 0.16 M (0.16 mmol)), EtOAc (1 mL), O2 (1 atm), 60 °C.

graphic file with name nihms481927e3.jpg (3)

Pd black is observed after the reaction, but the reaction mixture remains deep yellow, suggesting that the PdII catalyst is partially retained in solution. If Pd(TFA)2 is used as the catalyst (i.e., in the absence of DMSO), cyclohexenone dehydrogenation proceeds without an induction period to high conversion. More Pd black formation is observed relative to the reaction in the presence DMSO, and the solution is colorless after 24 h, suggesting near-complete conversion of PdII into Pd black during the reaction. The Pd black obtained from the reaction mixture is not an effective catalyst for the reaction.18 More thorough analysis of the effect of DMSO on the reaction reveals that increasing the [DMSO] has a significant inhibitory effect on the rate of phenol formation (Figure 7). Similarly, higher [DMSO] increases the length of the induction period (Figure 8 and S6 and Table S1). Collectively, these results are consistent with formation of Pd nanoparticles or intermediate-sized aggregates under the reaction conditions that exhibit higher activity than PdII for the dehydrogenation of cyclohexenone (see further discussion below). The inhibitory effect of DMSO can be attributed to its ability to stabilize the homogeneous PdII catalyst and slow formation of these particles.

Figure 7.

Figure 7

Open in a new tab

Dependence of the initial rate (i.e. during the induction period) of Pd-catalyzed dehydrogenation of cyclohexenone to phenol on the DMSO/Pd ratio. Estimated std. dev. for individual points: ≤ 5%. Reaction conditions: [cyclohexenone] = 0.8 M (0.8 mmol), [Pd(TFA)2] = 0.08 M (0.08 mmol), EtOAc (1 mL), O2 (1 atm), 60 °C.

Figure 8.

Figure 8

Open in a new tab

Dependence of the length of induction period on the Pd-catalyzed dehydrogenation of cyclohexenone to phenol on [DMSO]. Estimated std. dev. for individual points: ≤ 10 %. Reaction conditions: [cyclohexenone] = 0.8 M (0.8 mmol), [Pd(TFA)2] = 0.08 M (0.08 mmol), EtOAc (1 mL), O2 (1 atm), 60 °C.

Cyclohexenone dehydrogenation by PdII could be initiated by activation of the C–H bond at the 4- or 6-position of the ring, as relevant precedents exist for both. For example, deprotonation of cyclohexenone by lithium diisopropylamide (LDA) affords 2-oxylcyclohexa-1,3-dienyl lithium (eq 4),19,20 and deuterated Brønsted acids lead to isotopic scrambing of the α-C–H position, leaving the allylic C–H position unaffected.21 In contrast, cyclohexenone reacts with sodium tetrachloropalladate to afford a dimeric π-allyl-PdII species, corresponding to cleavage of the allylic C–H bond (eq 5).22,23 And, Imahori et al. recently reported a Pd-catalyzed C–H arylation-aromatization of cyclohexenone that furnishes 4-arylphenol derivatives, again consistent with Pd-mediated cleavage of the allylic C–H bond.12b

graphic file with name nihms481927e4.jpg (4)
graphic file with name nihms481927e5.jpg (5)

Isotopically labeled cyclohexenones were prepared to probe the site of C–H cleavage in Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexenone.24 Initial rates were monitored for reactions of cyclohexenone-d0, 2,6,6-cyclohexenone-d3, and 2,4,4,6,6-cyclohexenone-d5 under the standard reaction conditions (Figure 9). The unlabeled cyclohexanone exhibits a rate 2.9-fold faster than the two deuterium-labeled substrates, both of which react with identical rates. These data indicate the presence of a primary kinetic isotope effect (2.9 ± 0.2) associated with cleavage of the α-C–H bond in the conversion of cyclohexenone to phenol, which not only establishes the site of C–H cleavage, but also reveals that this step is rate-determining in Pd-mediated dehydrogenation of cyclohexenone.

Figure 9.

Figure 9

Open in a new tab

Deuterium kinetic isotope effects for dehydrogenation of cyclohexenone to phenol derived from independent measurement of initial rates (performed in triplicate). The plots are labeled with the corresponding cyclohexenone-dn substrate. Reaction conditions: [substrate] = 0.8 M (0.4 mmol), [Pd(TFA)2] = 0.04 M (0.02 mmol), [DMSO] = 0.08 M (0.04 mmol), EtOAc (1 mL), O2 (1 atm), 60 °C.

Mechanistic Analysis

(A) DMSO ligand-controlled chemoselectivity in the dehydrogenation of cyclohexanone

The mechanistic data presented above provide a number of insights into the Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone. Perhaps most striking is the dramatically different effects of DMSO on the cyclohexanone-to-cyclohexenone and cyclohexenone-to-phenol dehydrogenation steps. DMSO has negligible impact on the rate of the first reaction (Figure 3A), but it strongly inhibits the second (Figure 7). These contrasting effects provide the basis for selective formation of cyclohexenone, rather than phenol, in the catalytic reaction.

The inhibitory effect of DMSO on cyclohexenone-to-phenol dehydrogenation is attributed to two factors: (1) a dramatic decrease in the initial rate (Figure 7) and (2) an increase in the length of the induction period (Figure 8) with increasing [DMSO]. The effect of [DMSO] on the initial rate can be explained by DMSO dissociation from PdII prior to activation of cyclohexenone, while the effect of [DMSO] on the induction period appears to be associated with Pd aggregation into Pd nanoparticles and Pd black. DMSO serves as a stabilizing ligand for the homogeneous catalyst, inhibiting Pd aggregation. This effect is needed to achieve high conversion in the dehydrogenation of cyclohexanone to cyclohexenone (cf. Figures 3B). In contrast, dehydrogenation of cyclohexenone to phenol appears to benefit from catalyst aggregation, as reflected by the enhanced rate following the induction period. The latter reaction proceeds most rapidly under ligand-free conditions, in the absence of DMSO.

The sigmoidal time course observed for cyclohexenone-to-phenol dehydrogenation in the presence of DMSO (Figure 6) resembles reactions that undergo in situ transformation of a molecular catalyst-precursor into catalytically active nanoparticles.25 Pd nanoparticle formation has been characterized recently in the Pd(TFA)2/2-Me2Npy-catalyzed dehydrogenation of cyclohexanone to phenol,17 and the Pd nanoparticles were found to be more active than the PdII catalyst precursor for dehydrogenation of cyclohexenone. Extrapolation of these results to the present catalyst system supports the proposal that the inhibitory effect of DMSO on the cyclohexenone-to-phenol step (cf. Figure 7) arises from DMSO stabilization of homogeneous Pd and inhibition of Pd-nanoparticle formation. Thus, the high selectivity for enone formation with Pd(DMSO)2(TFA)2 appears to correlate with the presence of a (relatively) stable homogeneous PdII catalyst.

(B) Mechanism of Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone

We recently characterized the solution-phase structure of Pd(DMSO)2(TFA)2 in a number of different solvents.26 In EtOAc, this complex exists as an equilibrium mixture of S,S- and S,O-ligated bis-DMSO complexes (eq 6). Pd(DMSO)2(TFA)2-catalyzed dehydrogenation of cyclohexanone is envisioned to proceed by a series of three steps (Scheme 3): (i) formation of a PdII-cyclohexanone adduct, (ii) cleavage of the α-C–H bond, with concomitant loss of TFAH, to afford a PdII-enolate species, and (iii) β-hydride elimination to afford the cyclohexenone product and a PdII–hydride.

Scheme 3.

Scheme 3

Open in a new tab

Proposed Mechanism for Pd(DMSO)2(TFA)2-Catalyzed Dehydrogenation of Cyclohexanone.

graphic file with name nihms481927e6.jpg (6)

An alternative pathway for the formation of the PdII-enolate species could involve reversible ketone-enol tautomerization, followed by activation of the enol by PdII (eq 7). This mechanism would resemble that of Saegusa-type oxidations of silyl enol ethers;15 however, it is not consistent with the significant primary kinetic isotope effect observed here (cf. Figure 4).

graphic file with name nihms481927e7.jpg (7)

The experimental data reveal a first-order kinetic dependence on [cyclohexanone] and [Pd(DMSO)2(TFA)2]. Whereas a primary KIE is observed for cleavage of the α-C–H bond, no KIE is observed for cleavage of the β-hydrogen atom. These data are consistent with pre-equilibrium or steady-state formation of the PdII-cyclohexanone adduct followed by turnover-limiting cleavage of the α-C–H bond, as proposed in Scheme 3, steps i and ii.

The minimal effect of DMSO on the reaction rate suggests that DMSO does not dissociate prior to the turnover-limiting step. Application of the steady-state approximation to the PdII-cyclohexanone adduct results in the rate law shown in eq 8, which is consistent with the first-order dependence on [cyclohexanone] and [Pd(DMSO)2(TFA)2]. Subsequent β-hydride elimination27 and aerobic oxidation of the PdII–hydride16 are rapid and kinetically invisible under the catalytic conditions.

graphic file with name nihms481927e8.jpg (8)

The comparison of acetate and trifluoroacetate as anionic ligands shows that the Pd(TFA)2-derived catalyst is approximately three-fold more active. This effect is attributed to the enhanced electrophilicity of the TFA-ligated catalyst, which should favor formation of the cyclohexanone adduct and possibly enhance the rate of C–H cleavage. The maximum rate of cyclohexanone is observed, however, with a 1:1 TFA:OAc ratio. Studies of PdII-mediated C–H activation have highlighted “concerted metalation-deprotonation” mechanisms in which Pd–C bond formation takes place in concert with C–H deprotonation by a coordinated carboxylate,28 and we have recently observed a similar beneficial effect of mixed TFA/OAc anionic ligands in Pd-catalyzed C–H activation of arenes.29 These observations potentially reflect a need to balance the electrophilicity of the PdII center and the basicity of the carboxylate ligand. On the basis of these considerations and the lack of an inhibitory effect by DMSO, we propose the five-coordinate transition structure 1 for α-C–H cleavage. This structure closely resembles the five-coordinate transition state 2 proposed for Pd(py)2(OAc)2-catalyzed aerobic oxidation/dehydrogenation of alcohols, which is supported by experimental and DFT computational studies.30 The initially formed O-bound enolate should be able to isomerize readily under the reaction conditions into a C-bound enolate that can undergo β-hydride elimination.

graphic file with name nihms481927u1.jpg

A mechanism analogous to Scheme 3, involving a soluble PdII species, appears to be operative in the dehydrogenation of cyclohexenone during the induction period (cf. Figure 6). The alkene in cyclohexenone is probably a better ligand for Pd than the carbonyl oxygen atom. The resulting alkene adduct may not be on the pathway for α-C–H cleavage, however, and could slow the net dehydrogenation process for this substrate. Furthermore, the loss of DMSO in this step (evident from the inhibitory effect of DMSO) could make the catalyst more susceptible to aggregation into Pd nanoparticles upon reduction to Pd0. For reasons that have not yet been elucidated, the resulting Pd nanoparticles exhibit higher activity than the molecular PdII catalyst precursor.

Conclusion

Pd(DMSO)2(TFA)2 catalyzes the chemoselective formation of cyclohexenone, rather than phenol, in the aerobic dehydrogenation of cyclohexanone. The selectivity of this reaction originates from the different kinetic effect of DMSO on the two sequential dehydrogenation steps: DMSO has little impact on the rate of the first dehydrogenation step, whereas it strongly inhibits the second step. DMSO is found to stabilize the homogeneous PdII catalyst, which mediates efficient ketone-to-enone dehydrogenation. The same catalyst is not effective for cylohexenone-to-phenol dehydrogenation. Efficient catalysis of the latter reaction is only observed under conditions that enable conversion of the homogeneous Pd into Pd nanoparticles. These findings may be compared to our companion study of the Pd(TFA)2/2-Me2Npy catalyst system, which promotes full dehydrogenation of cyclohexanone to phenol.17 In this reaction, the PdII catalyst precursor transforms rapidly into Pd nanoparticles that mediate both dehydrogenation steps. Together, these studies represent a rare demonstration of different selectivity patterns that can arise from homogeneous versus nanoparticle catalysis.31

Supplementary Material

1_si_001

Acknowledgments

Financial support of this work was provided by the NIH (R01-GM100143) and the NSF (CHE-1041934 - fellowship to DP). The O2 dependence experiment was performed on a HEL high pressure (HP) ChemSCAN unit, which was funded by the NSF (CHE-0946901).

Footnotes

Supporting Information. Experimental details for data acquisition and additional kinetic data are available free of charge via the Internet at http://pubs.acs.org.

References

  • 1.For recent reviews: Dobereiner GE, Crabtree RH. Chem Rev. 2010;110:681. doi: 10.1021/cr900202j.Choi J, MacArthur AHR, Brookhart M, Goldman AS. Chem Rev. 2011;111:1761. doi: 10.1021/cr1003503.
  • 2.Recent efforts to use transfer dehydrogenation catalysts with cyclohexanone substrates: Yi CS, Lee DW. Organometallics. 2009;28:947. doi: 10.1021/om8010883.Zhang X, Wang DY, Emge TJ, Goldman AS. Inorg Chim Acta. 2011;369:253.
  • 3.For a leading review article, see: Cavani F, Ballarini N, Cericola A. Catal Today. 2007;127:113.Gärtner CA, van Veen AC, Lercher JA. ChemCatChem. 2013 doi: 10.1002/cctc.201200966.
  • 4.Muzart J. Eur J Org Chem. 2010:3779. [Google Scholar]
  • 5.For general reviews of Pd-catalyzed aerobic oxidation reactions, see: Stahl SS. Angew Chem Int Ed. 2004;43:3400. doi: 10.1002/anie.200300630.Gligorich KM, Sigman MS. Chem Commun. 2009:3854. doi: 10.1039/b902868d.Shi Z, Zhang C, Tang C, Jiao N. Chem Soc Rev. 2012;41:3381. doi: 10.1039/c2cs15224j.
  • 6.For leading primary references, see: Theissen RJ. J Org Chem. 1971;36:752.Trost BM, Metzner PJ. J Am Chem Soc. 1980;102:3572.Muzart J, Pete JP. J Mol Catal. 1982;15:373.Wenzel TT. J Chem Soc Chem Comm. 1989:932.Sheldon RA, Sobczak JM. J Mol Catal. 1991;68:1.Shvo Y, Arisha AHI. J Org Chem. 1998;63:5640.Tokunaga M, Harada S, Iwasawa T, Obora Y, Tsuji Y. Tetrahedron Lett. 2007;48:6860.Bercaw JE, Hazari N, Labinger JA. J Org Chem. 2008;73:8654. doi: 10.1021/jo8016296.Williams TJ, Caffyn AJM, Hazari N, Oblad PF, Labinger JA, Bercaw JE. J Am Chem Soc. 2008;130:2418. doi: 10.1021/ja076740q.Giri R, Maugel N, Foxman BM, Yu JQ. Organometallics. 2008;27:1667.Stang EM, White MC. J Am Chem Soc. 2011;133:14892. doi: 10.1021/ja2059704.
  • 7.Complementary studies have been reported describing hydrogen-atom-abstraction reactions that install sites of unsaturation into organic molecules: Bigi MA, Reed SA, White MC. Nat Chem. 2011;3:216. doi: 10.1038/nchem.967.Voica AF, Mendoza A, Gutekunst WR, Fraga JO, Baran PS. Nat Chem. 2012;4:629. doi: 10.1038/nchem.1385.
  • 8.Izawa Y, Pun D, Stahl SS. Science. 2011;333:209. doi: 10.1126/science.1204183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Diao T, Wadzinski TJ, Stahl SS. Chem Sci. 2012;3:887. doi: 10.1039/C1SC00724F. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.For a concurrent report on Pd/diazafluorenone-catalyzed dehydrogenation, see: Gao W, He Z, Qian Y, Zhao J, Huang Y. Chem Sci. 2012;3:883.
  • 11.Diao T, Stahl SS. J Am Chem Soc. 2011;133:14566. doi: 10.1021/ja206575j. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.(a) Simon MO, Girard SA, Li CJ. Angew Chem Int Ed. 2012;51:7537. doi: 10.1002/anie.201200698. [DOI] [PubMed] [Google Scholar]; (b) Imahori T, Tokuda T, Taguchi T, Takahata H. Org Lett. 2012;14:1172. doi: 10.1021/ol300145g. [DOI] [PubMed] [Google Scholar]; (c) Xie Y, Liu S, Liu Y, Wen Y, Deng GJ. Org Lett. 2012;14:1692. doi: 10.1021/ol3002442. [DOI] [PubMed] [Google Scholar]; (d) Hajra A, Wei Y, Yoshikai N. Org Lett. 2012;14:5488. doi: 10.1021/ol302568b. [DOI] [PubMed] [Google Scholar]; (e) Girard SA, Hu X, Knauber T, Zhou F, Simon MO, Deng GJ, Li CJ. Org Lett. 2012;14:5606. doi: 10.1021/ol3027279. [DOI] [PubMed] [Google Scholar]; (f) Izawa Y, Zheng C, Stahl SS. Angew Chem Int Ed. 2013;52:3672. doi: 10.1002/anie.201209457. [DOI] [PMC free article] [PubMed] [Google Scholar]; (g) Sutter M, Sotto N, Raoul Y, Métay E, Lemaire M. Green Chemistry. 2013;15:347. [Google Scholar]
  • 13.Kinetic fitting was carried with COPASI software: Hoops S, Sahle S, Gauges R, Lee C, Pahle J, Simus N, Singhal M, Xu L, Mendes P, Kummer U. Bioinformatics. 2006;22:3067. doi: 10.1093/bioinformatics/btl485.
  • 14.The mechanistic insights obtained in the present study suggest that solvents that stabilize homogeneous Pd catalysts will provide better selectivity for enone versus phenol formation. The mechanistic basis for such stabilization effects is not fully known, but multiple factors could contribute. Possible effects include the relative O2 solubility, the coordinating ability of the solvent, and/or other effects that contribute to the partitioning between catalyst reoxidation by O2 versus aggregation of Pd0 into nanoparticles.
  • 15.(a) Ito Y, Hirao T, Saegusa T. J Org Chem. 1978;43:1011. [Google Scholar]; (b) Larock RC, Hightower TR, Kraus GA, Hahn P, Zheng D. Tetrahedron Lett. 1995;36:2423. [Google Scholar]; (c) Porth S, Bats JW, Trauner D, Giester G, Mulzer J. Angew Chem, Int Ed. 1999;38:2015. doi: 10.1002/(SICI)1521-3773(19990712)38:13/14<2015::AID-ANIE2015>3.0.CO;2-#. [DOI] [PubMed] [Google Scholar]
  • 16.(a) Popp BV, Stahl SS. J Am Chem Soc. 2007;129:4410. doi: 10.1021/ja069037v. [DOI] [PubMed] [Google Scholar]; (b) Konnick MM, Stahl SS. J Am Chem Soc. 2008;130:5753. doi: 10.1021/ja7112504. [DOI] [PubMed] [Google Scholar]; (c) Popp BV, Stahl SS. Chem Eur J. 2009;15:2915. doi: 10.1002/chem.200802311. [DOI] [PubMed] [Google Scholar]; (d) Decharin N, Popp BV, Stahl SS. J Am Chem Soc. 2011;133:13268. doi: 10.1021/ja204989p. [DOI] [PMC free article] [PubMed] [Google Scholar]; (e) Konnick MM, Decharin N, Popp BV, Stahl SS. Chem Sci. 2011;2:326. [Google Scholar]
  • 17.Pun D, Diao T, Stahl SS. submitted for publication. [Google Scholar]
  • 18.The heterogeneous Pd was isolated by filtration through Celite. Efforts to use this material in the dehydrogenation of cyclohexenone under the standard reactions conditions led to disproportionation of cyclohexenone into cyclohexanone and phenol. For a precedent for this disproportionation reactivity with Pd/C, see: Horning EC, Horning MG, Walker GN. J Am Chem Soc. 1949;71:169. For similar observations with cyclohexene, see refs 6b and 6i.
  • 19.Rubottom GM, Gruber JM. J Org Chem. 1977;42:1051. [Google Scholar]
  • 20.Formation of the linear conjugated dienolate can be achieved by using KN(TMS)2 as the base: Kawanisi M, Itoh Y, Hieda T, Kozima S, Hitomi T, Kobayashi K. Chem Lett. 1985:647.
  • 21.Forsyth DA, Botkin JH, Osterman VM. J Am Chem Soc. 1984;106:7663. [Google Scholar]
  • 22.Kasahara A, Tanaka K, Asamiya K. Bull Chem Soc Jpn. 1967;40:351. [Google Scholar]
  • 23.For related Pd-mediated cleavage of allylic C–H bonds to form π-allyl-Pd complexes, see ref. 6b and Parshall GW, Wilkinson G. Inorg Chem. 1962;1:896.
  • 24.The syntheses of 2,6,6-cyclohexenone-d3 and 2,4,4,6,6-cyclohexenone-d5 are based on procedures described in following paper: Lambert JB, Clikeman RR. J Am Chem Soc. 1976;98:4203.In this paper, 2,6,6-cyclohexenone-d3 is inaccurately assigned to 2,4,4-cyclohexenone-d3. Our 1H NMR and 1D NOESY data confirm that the resulting deuterated substrate is 2,6,6-cyclohexenone-d3.
  • 25.See, for example: Widegren JA, Finke RG. J Mol Catal A: Chem. 2003;198:317.Crabtree RH. Chem Rev. 2011;112:1536. doi: 10.1021/cr2002905.
  • 26.Diao T, White P, Guzei I, Stahl SS. Inorg Chem. 2012;51:11898. doi: 10.1021/ic301799p. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.For a fundamental study of β-hydride elimination from a PtII-enolate, see: Alexanian EJ, Hartwig JF. J Am Chem Soc. 2008;130:15627. doi: 10.1021/ja8056908.
  • 28.(a) Davies DL, Donald SMA, Macgregor SA. J Am Chem Soc. 2005;127:13754. doi: 10.1021/ja052047w. [DOI] [PubMed] [Google Scholar]; (b) García-Cuadrado D, Braga AAC, Maseras F, Echavarren AM. J Am Chem Soc. 2006;128:1066. doi: 10.1021/ja056165v. [DOI] [PubMed] [Google Scholar]; (c) Gorelsky SI, Lapointe D, Fagnou K. J Am Chem Soc. 2008;130:10848. doi: 10.1021/ja802533u. [DOI] [PubMed] [Google Scholar]; (d) Balcells D, Clot E, Eisenstein O. Chem Rev. 2010;110:749. doi: 10.1021/cr900315k. [DOI] [PubMed] [Google Scholar]; (e) Ackermann L. Chem Rev. 2011;111:1315. doi: 10.1021/cr100412j. [DOI] [PubMed] [Google Scholar]
  • 29.Izawa Y, Stahl SS. Adv Synth Catal. 2010;352:3223. doi: 10.1002/adsc.201000771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Steinhoff BA, Guzei IA, Stahl SS. J Am Chem Soc. 2004;126:11268. doi: 10.1021/ja049962m. [DOI] [PubMed] [Google Scholar]
  • 31.Homogeneous vs. nanoparticle catalysis was recently invoked to explain a switch in selectivity between alcohol oxidation and Wacker-type oxidation of alkenes: Mifsud M, Parkhomenko KV, Arends IWCE, Sheldon RA. Tetrahedron. 2010;66:1040.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1_si_001
NIHMS481927-supplement-1_si_001.pdf (860.4KB, pdf)

RESOURCES