Abstract
Alkyl and alkenyl arenes are precursors to polymers, elastomers, fragrances and pharmaceuticals. For example, styrene is used for the production of polymers and is produced on a scale of ~40 million tons per year through a multiple-step and energy-consuming process. The predominant route to produce styrene involves ethylbenzene dehydrogenation, which requires temperatures > 550 °C. Production of ethylbenzene is itself inefficient, operating through acid-catalyzed benzene ethylation. Acid-catalyzed benzene ethylation produces substantial quantities of polyethylbenzenes as a result of overalkylation. Polyethylbenzene side products are then converted to ethylbenzene by heating with benzene at temperatures > 350 °C in a trans-alkylation reaction. Many of the inefficiencies in styrene production are the direct consequence of mechanistic limitations of acid-catalyzed arene alkylation.
An alternative process to produce styrene and related alkenyl arenes is through transition metal-catalyzed oxidative arene alkenylation. When combined with an oxidant such as Cu(II) carboxylates or dioxygen, previous studies by the Gunnoe group and others have found that catalyst precursors Pd(OAc)2 (OAc = acetate) and [(η2-C2H4)2Rh(µ-OAc)]2 are active for alkenyl arene production at temperatures 120-200 °C. The proposed mechanism of alkenyl arene formation involves metal-mediated arene C–H activation by a metal carboxylate, olefin insertion into the formed metal-aryl bond, β–hydride elimination, and reaction of the formed metal hydride intermediate with two oxidizing equivalents to regenerate the starting catalyst. In addition to providing a single-step route to alkenyl arenes at lower temperatures than currently employed processes, transition metal-catalyzed arene alkenylation can access products unattainable to acid-catalyzed reactions when employing substituted arenes and olefins.
Pd(OAc)2 is a catalyst precursor for styrene production from benzene and ethylene using Cu(OPiv)2 as the oxidant. When performing catalytic benzene ethenylation reactions, an induction period of approximately one hour was observed. In situ 1H NMR spectroscopy studies found that Pd(OAc)2 is consumed during the induction period and converted to PdCu2(μ-OPiv)6, which was identified by single crystal x-ray diffraction. Use of PdCu2(μ-OPiv)6 as the catalyst precursor resulted in no induction period, consistent with its identification as an active catalyst. DFT calculations by the Goddard group find that the barrier for benzene C–H activation through a concerted metalation deprotonation mechanism was substantially lower for PdCu2(μ-OAc)6 relative to Pd3(μ-OAc)6, which is consistent with experimental observations.
In an extension of the studies of Rh and Pd catalyst speciation, in previous work in the Gunnoe group the selectivity and activity of Rh and Pd catalysis was compared. Rh-catalyzed benzene ethenylation reactions proceed ~20-fold more rapidly than Pd-catalyzed benzene ethenylation at otherwise identical conditions. Also, Rh catalysis was found to be ~98% selective for styrene over undesired vinyl ester side product while Pd catalysis was ~82% selective. To probe differences in arene C–H activation mechanism, the regioselectivity of mono-substituted arene ethenylation was studied. Rh catalysis resulted in ~2:1 meta:para selectivity with minimal ortho functionalization with anisole, toluene and chlorobenzene, while Pd catalysis gave ~0.5:1:1 ortho:meta:para selectivity with each of these substrates. With trifluoromethylbenzene, both Pd and Rh catalysts gave ~3:1 meta:para selectivity with minimal ortho functionalization. The propensity of Pd catalysis to form ortho and para products when electron-rich arenes are used suggested that the Pd arene C–H activation step has significant electrophilic character. This hypothesis was supported by kinetic intermolecular competition experiments using equimolar quantities of benzene and mono-substituted arenes. While Rh catalysis gave similar relative rates between electron-rich and electron-deficient substrates, the rate of Pd catalysis decreased as substituent donor ability is decreased. These findings indicated that Pd activates C–H bonds through a mechanism with electrophilic aromatic substitution character while the C–H bond step for Rh catalysis has minimal electrophilic character.
The effect of olefin substituent on the reaction rate and selectivity of Rh-catalyzed arene alkenylation was studied. This included comparative kinetics and regioselectivity using 1-butene, cis- and trans-2-butene, isobutene, 2-methyl-2-butene and tetramethylethylene as the olefin, in addition to the analogous phenyl-substituted isomers and isomers of propenylbenzene. The studies found that mono-substituted olefins react more rapidly than di-substituted olefins, with the relative rates among di-substituted olefins dependent upon substituent identity, and tri-substituted olefins are minimally reactive. Also, we probed Markovnikov to anti-Markovnikov selectivity using methyl, ethyl, isopropyl, tert-butyl and cyclohexyl-substituted mono-substituted olefins and found that larger substituents generally result in increased selectivity for anti-Markovnikov products.
A disadvantage of using Cu(II) carboxylates as the oxidant for Rh-catalyzed arene alkenylation is that they mediate a stoichiometric side-reaction with benzene to form phenyl ester (e.g., phenyl acetate) byproducts. While the catalysis can be carried out in the absence of Cu(II) carboxylates using dioxygen as the oxidant, the turnover frequency is approximately two orders of magnitude slower relative to catalysis with Cu(II). Our group previously speculated this to be the result of kinetically challenging Rh–H oxidation by dioxygen. With the goal of identifying an alternative oxidant to Cu(II) carboxylates, we speculated that Fe carboxylates might be capable oxidants. Aerobic benzene ethenylation reactions performed in the presence of Fe(II) carboxylates results in a turnover frequency ~20-fold more rapid than catalysis performed only in the presence of dioxygen. Mechanistic studies determined that Fe(II) carboxylates react with dioxygen and carboxylic acid to form FeIII6(µ-OH)2(µ3-O)2(µ-OPiv)12(HOPiv)2, which likely serves as the direct oxidant.
In order to quantify differences in Rh-catalyzed arene alkenylation as a function of in situ oxidant identity, catalysis using Cu(II) carboxylates under anaerobic conditions, Cu(II) carboxylates under aerobic conditions, Fe(III) carboxylates under anaerobic conditions, Fe(II) carboxylates under aerobic conditions, and dioxygen as the oxidant were studied. This study included comparisons of the rate of styrene formation and the selectivity for styrene versus undesired side products such as vinyl esters, benzaldehyde, phenyl esters and trans-stilbene. The dependence of reaction rate on reagent concentrations was probed to understand differences in reaction pathway as a function of oxidant identity. Additionally, ortho:meta:para regioselectivity was probed using mono-substituted arenes, and Markovnikov:anti-Markovnikov selectivity was compared using mono-substituted olefins.
