Report

A rhodium catalyst for single-step styrene production from benzene and ethylene

See allHide authors and affiliations

Science  24 Apr 2015:
Vol. 348, Issue 6233, pp. 421-424
DOI: 10.1126/science.aaa2260

A more direct way to synthesize styrene

Foam cups, foam pellets, plastic cutlery: All are made of polystyrene, which in turn is made of styrene. The massive manufacturing scale of this commodity chemical places a premium on the efficiency of its synthesis. The current industrial route requires three steps to make styrene from benzene and ethylene. Vaughan et al. present a rhodium catalyst that achieves the coupling in a single step by using a recyclable copper salt as an oxidant. Although the catalyst is slow for industrial application, it demonstrates the viability of a more direct process.

Science, this issue p. 421

Abstract

Rising global demand for fossil resources has prompted a renewed interest in catalyst technologies that increase the efficiency of conversion of hydrocarbons from petroleum and natural gas to higher-value materials. Styrene is currently produced from benzene and ethylene through the intermediacy of ethylbenzene, which must be dehydrogenated in a separate step. The direct oxidative conversion of benzene and ethylene to styrene could provide a more efficient route, but achieving high selectivity and yield for this reaction has been challenging. Here, we report that the Rh catalyst (FlDAB)Rh(TFA)(η2–C2H4) [FlDAB is N,N′-bis(pentafluorophenyl)-2,3-dimethyl-1,4-diaza-1,3-butadiene; TFA is trifluoroacetate] converts benzene, ethylene, and Cu(II) acetate to styrene, Cu(I) acetate, and acetic acid with 100% selectivity and yields ≥95%. Turnover numbers >800 have been demonstrated, with catalyst stability up to 96 hours.

Vinyl arenes are important precursors for fine chemical synthesis, as well as for the preparation of plastics and elastomers (15). For example, styrene is produced globally on a scale of ~18.5 million tons (2). Current methods for the large-scale production of vinyl arenes involve multiple steps, typically beginning with arene alkylation using a Friedel-Crafts (e.g., AlCl3 with HF) or zeolite catalyst followed by energy-intensive dehydrogenation of the alkyl group (Fig. 1) (16). Friedel-Crafts catalysis suffers from the use of harsh acids, including HF, low selectivity for the monoalkylated product (polyalkylation is inherent to the mechanism), and the generation of stoichiometric waste (2). Zeolite catalysts have improved the process for benzene alkylation, yet these catalysts still require high temperatures (generally 350° to 450°C) and give polyalkylated products (2, 710).

Fig. 1 Comparison of the current route to styrene production and the single-step route described herein.

An alternative method for the production of vinyl arenes is a direct and single-step oxidative arene vinylation (Fig. 1). If the terminal oxidant is oxygen from air (either introduced in situ or used to recycle a different in situ oxidant), the net reaction is the conversion of benzene, ethylene, and oxidant to styrene and water (11). Acid-based (i.e., Friedel-Crafts or zeolite catalysts) catalysis occurs by electrophilic aromatic substitution and does not offer a viable pathway to directly generate vinyl arenes. Transition metal complexes that catalyze ethylene hydrophenylation by benzene C–H activation followed by ethylene insertion into a metal-phenyl bond have been reported as alternatives to acid-based catalysts (Fig. 2) (1225). For these catalysts, β-hydride elimination from a M–CH2CH2Ph intermediate and dissociation of styrene provides a route for the direct oxidative vinylation of benzene (Fig. 2).

Fig. 2 Proposed cycle for transition metal–catalyzed styrene production from benzene and ethylene using CuX2 as an oxidant.

The cuprous (CuX) product could be recycled back to the cupric state using air, as shown at the upper left. Potential side reactions that a selective catalyst must avoid are shown in red.

Previously, our groups have studied the use of platinum(II) catalysts for the hydrophenylation of ethylene to produce ethylbenzene (1619, 2628). Through a combination of experimental and computational mechanistic studies, we discerned a competing β-hydride elimination pathway from Pt–CH2CH2Ph intermediates to form a Pt–styrene hydride complex, which can lead to the formation of free styrene (28). Unfortunately, the formation of styrene leads to catalyst decomposition (27). We proposed that this catalyst decomposition is the result of unstable Pt(II)–hydride complexes, which are formed from β-hydride elimination, that react to release H2 and produce metallic Pt. The thermodynamic driving force for the formation of Pt(s) presents a substantial challenge to achieving long-lived vinyl arene production with these catalysts (Fig. 2, inset) (11). Given that the formation and decomposition of Pt(II)–H species is problematic, we sought to design catalysts using isoelectronic Rh(I) in anticipation that Rh(I)–H would exhibit greater stability compared with related Pt species (Fig. 2, inset) (11).

Figure 2 shows a targeted catalytic cycle for the direct oxidative vinylation of benzene to produce styrene. Despite precedent for the key steps in this catalytic cycle, designing a selective catalyst represents a substantial challenge because many competing side reactions (shown in red) are likely to have activation barriers that are similar to or lower than those of the reactions along the desired catalytic cycle. In addition to these possible side reactions, designing a molecular catalyst that achieves high turnover numbers (TONs) is difficult because the oxidative conditions and the presence of potentially reactive metal-hydride intermediates could be anticipated to result in catalyst decomposition.

Table 1 compares previously reported homogeneous catalysts for direct oxidative styrene synthesis from ethylene and benzene (2934). Generally, all suffer from one or more of the following drawbacks: low selectivity, low yield, low TON, and/or use of oxidants that cannot be regenerated using oxygen. Notable results include the work of Hong and co-workers, who reported a Rh4(CO)12 catalyst that gave, to our knowledge, the highest TON of styrene (472). In tandem with this process, liberated dihydrogen is consumed by two equivalents of ethylene and one equivalent of CO to produce 3-pentanone with 809 turnovers (TOs) (29). Sanford and co-workers reported that (3,5-dichloropyridyl)Pd(OAc)2 catalyzes styrene production with 100% selectivity and 6.6 TOs for styrene (33% overall yield) using PhCO3tBu, an oxidant that cannot be recycled with oxygen (33). Here, we report a rhodium catalyst for the selective one-step production of styrene from benzene, ethylene, and Cu(II) salts. We chose a Cu(II) salt as the in situ oxidant because of industrial precedent for recycling reduced Cu(I) using oxygen. In the commercial Wacker-Hoechst process for ethylene oxidation (35, 36), use of oxygen to reoxidize Cu(I) to Cu(II) has proven viable both in situ and in a second step (37).

Table 1 Comparison of previously reported catalysts for styrene production.

Selectivity is defined as the ratio of turnovers of styrene to total turnovers (all products) and is given as a percentage. Yield of styrene is reported relative to the limiting reagent. acac, acetylacetonate; DBM, dibenzoylmethane; DCP, 3,5-dichloropyridine; HPA, H3PMo12O40·30H2O; TFA, trifluoroacetate.

View this table:

We recently reported the synthesis of an electron-deficient Rh(I) complex (FlDAB)Rh(TFA)(η2-COE) [FlDAB is N,N′-bis(pentafluorophenyl)-2,3-dimethyl-1,4-diaza-1,3-butadiene; TFA is trifluoroacetate; COE is cyclooctene] and demonstrated that this complex is an active catalyst for arene H/D exchange in trifluoroacetic acid (38). Given that arene C–H activation is a key step in transition metal–catalyzed oxidative arene vinylation, we hypothesized that this Rh(I) complex might be an effective catalyst precursor for styrene production. Because the COE ligand would likely exchange for ethylene, the ethylene analog (FlDAB)Rh(TFA)(η2-C2H4) (1) was independently synthesized as our catalyst precursor (Fig. 3).

Fig. 3 Synthesis of (FlDAB)Rh(TFA)(η2-C2H4) (1) (RT, room temperature).

Heating a 20-mL benzene solution of 1 [0.001 mole percent (mol %) relative to benzene] with ethylene and Cu(OAc)2 (120 equivalents relative to 1) to 150°C affords 58 to 62 TOs of styrene after 24 hours (for all TOs reported, two runs were performed, and both results are given). Samples of the reaction mixture were analyzed by gas chromatography–flame ionization detector (GC/FID) using relative peak areas versus an internal standard (decane). This corresponds to quantitative yield based on the Cu(II) limiting reagent. The calculated yield here assumes that two equivalents of Cu(II) are consumed to produce two equivalents of Cu(I) per equivalent of styrene. No other products were observed upon analysis of the reaction mixture by GC–mass spectrometry or GC/FID, indicating high selectivity for styrene production. Detection limits for the instruments were equivalent to ~1 TO of product. Specifically, we looked for evidence of stilbene, biphenyl, and vinyl acetate production, because these are the most commonly observed by-products in previously reported catalysis (Table 1). Control reactions with [Rh(μ-TFA)(η2-C2H4)2]2, a precursor to complex 1, afforded <5 TOs of styrene after 24 hours, with or without Cu(OAc)2, highlighting the importance of the FlDAB ligand. Control reactions with Cu(OAc)2 alone also afforded no styrene formation.

With a competent catalyst in hand, we next sought to optimize reaction conditions. The effect of oxidant identity on catalysis with 1 was the first parameter investigated. Both soluble {copper 2-ethylhexanoate [Cu(OHex)2] and copper pivalate [Cu(OPiv)2]} and insoluble {copper acetate [Cu(OAc)2] and copper trifluoroacetate hydrate [Cu(TFA)2]} Cu(II) salts were screened. Figure S2 shows a plot of turnovers versus time for the various Cu(II) oxidants. Using a turnover frequency (TOF) calculated after 4 hours of reaction, soluble Cu(OHex)2 gives the fastest initial rate with a TOF of 2.8 × 10−3 s–1, but the reaction does not reach 100% yield relative to oxidant until 28 hours, which may indicate that catalyst deactivation occurs. Cu(OAc)2 affords a slower initial rate than Cu(OHex)2, with a TOF of 2.8 × 10−4 s–1 after 4 hours, but this oxidant provides a more stable catalytic process. Both Cu(TFA)2 and Cu(OPiv)2 afford slower initial rates; reactions with Cu(OPiv)2 reach 92% yield after 28 hours, whereas reactions with Cu(TFA)2 produce only 19 TOs of styrene (32% yield) after 20 hours.

To study catalyst longevity, we varied the amount of Cu(OAc)2. Between 60 and 240 equivalents (relative to 1), the yield of styrene relative to oxidant was always >95% (fig. S3). These near-quantitative yields demonstrate that the catalytic process using 1 as a precursor is stable and long-lived. For a reaction using 0.0001 mol % 1 and 2400 equivalents of Cu(OAc)2, the catalyst remained active over a period of 96 hours and afforded a TON of 817 to 852. A plot of TO versus time shows that the Rh catalyst is stable through at least 96 hours (fig. S1). The tolerance of 1 to a large excess of oxidant without any decrease in activity is promising. The effect of temperature on catalysis was also examined (fig. S4). Generally, the rate of reaction increased with temperature; however, at 180°C, rapid catalyst deactivation led to a low TON (<10 TOs) after 12 hours. Minimal activity (<1 TO) was also observed at temperatures <100°C. The optimal temperature proved to be 150°C.

We also observed that the reaction rate increased with increasing ethylene pressure. To determine the TOF, we measured TO after 4 hours of reaction. Figure 4 shows a plot of TOF versus ethylene pressure. A linear correlation is observed. Thus, the reaction rate appears to have a first-order dependence on ethylene concentration. This is in contrast to previously reported Pt(II) and Ru(II) catalysts for the hydrophenylation of ethylene, which show an inverse dependence on ethylene pressure (14, 17). For the Pt and Ru catalysts, M(CH2CH2Ph)(η2-C2H4) complexes were identified as the likely catalyst resting states. The opposite dependence on ethylene pressure using 1 as catalyst precursor signals a likely change in the catalyst resting state.

Fig. 4 Effect of ethylene pressure on catalysis with (FlDAB)Rh(TFA)(η2-C2H4) (1).

Reaction conditions: 0.001 mol % 1, 120 equivalents Cu(OAc)2, 150°C, 4 hours. Data for two independent reactions are shown.

To gain further insight into the reaction mechanism, we ran the reaction in a 1:1 molar mixture of C6H6 and C6D6. After 1 hour, a kH/kD (ratio of rate of reaction of protio-benzene and perdeutero-benzene) of 3.1(2) was determined by examining the ratio of undeuterated styrene [mass/charge ratio (m/z) = 104] to styrene-d5 (m/z = 109) in the mass spectra from three independent experiments (fig. S5). After 2 hours, the observed isotope effect was 3.0(2), statistically equivalent to the data at 1 hour (fig. S5). Thus, the observed kH/kD of ~3.1 likely reflects a kinetic isotope effect (KIE) for the catalytic cycle. The KIE is consistent with other transition metal–mediated C–H activation reactions. (39, 40) The primary KIE supports the hypothesis that a Rh catalyst is facilitating a metal-mediated C–H activation process, which occurs before or during the turnover-limiting step. No change in the isotopic distribution for benzene was observed over the course of the reaction, and no styrene-d6-8 products were observed except those predicted by the natural abundance of deuterium in ethylene.

Although more detailed studies are required to understand the reactivity profile of 1, we believe that the highly electron-withdrawing perfluorophenyl groups on the FlDAB ligand help suppress irreversible oxidation to inactive Rh(III) in the presence of Cu(II), possibly facilitate associative ligand exchange between free ethylene and coordinated styrene, and facilitate rapid ethylene insertion into Rh–Ph bonds. Challenges that remain for the continued development of this class of catalyst include increasing activity with the aim of achieving higher conversions of benzene.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/348/6233/421/suppl/DC1

Materials and Methods

Figs. S1 to S4

Table S1

References (4143)

References and Notes

  1. Acknowledgments: The authors acknowledge support from the U.S. Department of Energy, Office of Basic Energy Sciences [DE-SC0000776 (T.B.G.) and DE-FG02-03ER15387 (T.R.C.)] for studies of styrene catalysis; the Center for Catalytic Hydrocarbon Functionalization, an Energy Frontier Research Center (award DE-SC0001298), which funded the initial catalyst discovery; and an AES Corporation Graduate Fellowship in Energy Research (M.S.W.-G). The authors also thank B. McKeown, G. Fortman, S. Kalman (University of Virginia), and R. Nielsen (California Institute of Technology) for helpful discussions.
View Abstract

Navigate This Article