Introduction
Alkenyl and alkyl arenes are fundamental building blocks for various materials such as plastics, fragrances and pharmaceuticals.[1-3]123 In 2024, 42 million tons of styrene were produced, and these numbers are expected to grow over the next ten years.[4] The production of styrene is energy intensive, first requiring the alkylation of benzene and ethylene via Friedel-Crafts catalysis or a zeolite-catalyzed reaction to produce ethylbenzene (Scheme 1).[5] Benzene alkylation and product purification via distillation is followed by an energy intensive dehydrogenation.[6] There are many disadvantages to these commercial processes including over alkylation, the production of halogenated waste, the inability to form styrene from benzene in a single step, and catalyst dictated selectivity for reactions of substituted arenes and olefins (Scheme 1).[7-11]7891011
Transition metal catalyzed oxidative coupling of arenes and olefins to alkenyl arenes offers potential advantages to current commercial routes. Our group and others have reported Rh,[12-33]12131415161718192021222324252627282930313233Pd,[8,14,25,31,34,35]81425313435 Pt,[36-39]36373839 Ru,[31,40,41]314041 and Ir catalysts for this transformation.[31,42]3142 Often, these catalysts have been proposed to operate via C–H activation, olefin insertion, β-hydride elimination, and reaction of M–H intermediates with an in-situ oxidant to regenerate the starting catalyst (Scheme 2).
Our group has shown that oxidant identity can impact reaction kinetics and selectivity for arene alkenylation.[43] For example, in addition to impact on reaction rate for arene alkenylation, when using mono-substituted olefins such as propylene, differences in Markovnikov versus anti-Markovnikov selectivity, also referred to as linear:branched selectivity, based on oxidant identity have been quantified (Scheme 3). Further, when using mono-substituted arenes, ortho:meta:para regioselectivity for arene alkenylation can also vary with oxidant identity (Scheme 3).
This review will focus on the effect of in situ oxidant identity on reaction rate and product selectivity. Primarily, we discuss Rh-based catalysis since the majority of detailed mechanistic studies are for such reactions, but we include results from different catalysts as warranted.
Overview of oxidants
We will discuss Cu(II) carboxylates, dioxygen, Fe(III) carboxylates, and benzoquinones as oxidants for transition metal catalysis. While the use of dioxygen as the ultimate oxidant is likely required for the economic viability of any commercial process, there are several potential disadvantages to using it as the sole oxidant including the formation of off-cycle species with low valent metal centers,[44-46]444546 and the kinetically challenging oxidation of metal hydrides.[21,47]2147 Thus, use of in situ oxidants that can be recycled by dioxygen is a common strategy. A successful example is the commercial Wacker process in which CuCl2 is the operative oxidant, and CuCl and HCl are converted back to CuCl2.[48] In addition to the Wacker process, Cu(II) carboxylates have been used as oxidants in transition metal catalyzed arene alkenylation reactions including the Fujiwara-Moritani reaction.[49] Given the precedence for commercial use of Cu(II) for catalytic ethylene oxidation, we investigated Cu(II) carboxylates under both aerobic and anaerobic conditions for Rh-catalyzed arene alkenylation.Fe(III) has been studied as a catalyst for hydrocarbon oxidation reactions.[50-53]50515253 Additionally, some complexes of Fe(II) are known to undergo oxidation upon reactivation with dioxygen.[50-54]5051525354 The oxidative coupling of arenes and electron-deficient olefins catalyzed by Pd(OAc)2 using benzoquinones as oxidant has been reported by the Bäckvall and Fujiwara groups.[50-56]50515253545556
Rate of styrene production
We have quantified the kinetics of benzene ethenylation for each oxidant using 0.001 mol% of the catalyst precursor [(η2-C2H4)2Rh(μ-OAc)]2 (relative to benzene per single Rh atom), 960 equiv of HOPiv (relative to single Rh atom), and 70 psig of ethylene at 150 °C.[25] When using 480 equiv of Cu(OPiv)2, an initial turnover frequency (TOF) of 0.101(9) s-1 was observed (Table 1, entry 1), which is the highest TOF observed for any oxidant that we studied for Rh-catalyzed styrene formation. When 1 atm of O2 is added with Cu(OPiv)2, thus rendering Cu(II) catalytic rather than stoichiometric, the TOF is ~3-fold slower (entry 2). We speculate that the slower rate of catalysis in the presence of O2 could be due to O2 reacting with Rh to form off-cycle and catalytically inactive Rh species. When 1 atm O2 and 480 equiv of Fe(OAc)2 is used, a TOF of 0.0084(3) s-1 was observed (entry 3), similar to the rate using anaerobic Fe(III) (see below), indicating that dioxygen is not as detrimental as it is with Cu(OPiv)2. Use of 80 equiv Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2 under anaerobic conditions resulted in a TOF of 0.0073(7) s-1 (entry 4). Use of 1 atm of O2 as the sole oxidant gave the lowest TOF of 0.0006(1) s-1 (entry 5). We attribute the slow rate to be due in part to the Rh–H oxidation step being kinetically challenging.[43] Use of Pd catalysis was also compared to Rh using 480 equiv Cu(OPiv)2 and 960 equiv HOPiv at 120 °C. It was found that using Pd(OAc)2 as the catalyst precursor is > 20-fold slower than Rh and has 82% selectivity towards styrene, compared to 98% for Rh (entries 6 and 7).
Kinetics of styrene production using benzoquinones was also studied using 0.001 mol% of the catalyst precursor [(η2-C2H4)2Rh(μ-OPiv)]2, 960 equiv HOPiv, 70 psig of ethylene, and 240 equiv of benzoquinone derivatives at 170 °C.[30] We demonstrated that ortho-benzoquinones exhibit a faster rate of styrene production than para-benzoquinones with the fastest styrene production observed using 1,2-naphthoquinone (entries 8 and 9).
The formation of undesirable side products during styrene production was quantified for each oxidant (Table 2). Use of Cu(OPiv)2 resulted in 93% selectivity towards styrene with vinyl pivalate, biphenyl, phenyl pivalate, and trans-stilbene formed as side products.[43] When using Cu(OPiv)2 under aerobic conditions, selectivity towards styrene decreases and more phenyl pivalate is formed in comparison to anaerobic conditions (Entries 1 and 2). The decreased selectivity for styrene is likely due to the suppressed rate of formation of styrene while having less impact on rate of phenyl pivalate formation when using Cu(OPiv)2/O2 compared to anaerobic Cu(OPiv)2.[57] Use of only O2 gave the highest selectivity towards styrene, 95%, with the most significant side products observed being benzaldehyde and biphenyl. Use of Fe(OAc)2 as the oxidant under aerobic conditions results in the elimination of phenyl acetate or phenyl pivalate as a side product, although benzaldehyde production is observed (Entry 4). Use of Fe₆(μ-OH)₂(μ₃-O)₂(μ-OPiv)₁₂(HOPiv)₂ as the oxidant under anaerobic conditions results in significant biphenyl production (Entry 5), which is much more prominent than with use of Fe under aerobic conditions.
Table 1: Kinetics of benzene ethenylation to styrene as a function of metal catalyst and oxidant identity. Eq. 1.
| Entry | Catalyst precursor | Oxidant | oxidant(equiv) | O2 (atm) | HOPiv (equiv) | Ethylene pressure (psig) | Benzene amount (mL) | Temp (°C) | TOF (s-1) |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Rh1 | Cu(OPiv)2 | 480 | – | 480 | 70 | 7.5 | 150 | 0.101(9) |
| 2 | Rh1 | Cu(OPiv)2/O2 | 480 | 1 | 480 | 70 | 7.5 | 150 | 0.045(4) |
| 3 | Rh1 | Fe(OAc)2/O2 | 480 | 1 | 480 | 70 | 7.5 | 150 | 0.0084(6) |
| 4 | Rh1 | Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2 | 80 | – | 480 | 70 | 7.5 | 150 | 0.0073(7) |
| 5 | Rh1 | O2 | – | 1 | 480 | 70 | 7.5 | 150 | 0.0006(1) |
| 6 | Pd(OAc)2 | Cu(OPiv)2 | 480 | – | 960 | 50 | 10 | 120 | 0.0026(1) |
| 7 | Rh1 | Cu(OPiv)2 | 480 | – | 960 | 50 | 10 | 120 | 0.0561(7) |
| 8 | Rh2 | 1,2-naphthoquinone | 240 | – | 960 | 70 | 7.5 | 170 | 0.04 |
| 9 | Rh2 | para-chloranil | 240 | – | 960 | 70 | 7.5 | 170 | 0.0028 |
Rh1 = [(η²-C₂H₄)₂Rh(μ-OAc)]₂; Rh2 = [(η²-C₂H₄)₂Rh(μ-OPiv)]₂. Detailed reaction conditions can be found in original publications. Each TOF value represents the average of at least three independent experiments, and the values in parentheses represent the standard deviation in the final digit(s).
Table 2: Product selectivity for benzene ethenylation as a function of catalyst precursor and oxidant identity. Eq. 2.
| Entry | Catalyst precursor | Oxidant | A % Selectivity | B % Selectivity | C % Selectivity | D % Selectivity | E % Selectivity | F % Selectivity |
|---|---|---|---|---|---|---|---|---|
| 1 | Rh1 | Cu(OPiv)2 | 93 | 4 | nd | <1 | <1 | 2 |
| 2 | Rh1 | Cu(OPiv)2/O2 | 90 | 1 | 1 | 3 | 1 | 2 |
| 3 | Rh1 | O2 | 95 | nd | 2 | nd | 2 | nd |
| 4 | Rh1 | Fe(OAc)2/O2 | 92 | 1 | 2 | nd | 1 | 3 |
| 5 | Rh1 | Fe₆(μ-OH)₂(μ₃O)₂(μ-OPiv)₁₂(HOPiv)₂ | 92 | <1 | 1 | nd | 5 | <1 |
| 6 | Rh1 | Cu(OPiv)2 | 97 | 3 | nd | nd | nd | nd |
| 7 | Pd(OAc)2 | Cu(OPiv)2 | 96 | 4 | nd | nd | nd | nd |
Rh1 = [(η²-C₂H₄)₂Rh(μ-OAc)]₂; Rh2 = [(η²-C₂H₄)₂Rh(μ-OPiv)]₂; Ir1 = [Ir(μ-Cl)(coe)2]2; Fe6 = Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2. Detailed reaction conditions can be found in original publications. Use of “nd” indicates that the individual product selectivity’s were not reported. L:B ratios represent the average of at least three independent experiments, and the values in parentheses represent the standard deviation in the final digit(s).
Anti-Markovnikov to Markovnikov selectivity for benzene alkenylation using propylene
When using propylene as the olefin for alkenylation of benzene, the regioselectivity of propylene insertion into a metal-phenyl bond can dictate whether the Markovnikov (branched) or anti-Markovnikov (linear) product is formed (Scheme 2). Using a 0.001 mol% loading of [(η2-C2H4)2Rh(μ-OAc)]2 as the catalyst precursor and 960 equiv of HOPiv at 150 °C, we probed linear to branched selectivity with 50 psig of propylene as the olefin.[43]Using 480 equiv of Cu(OPiv)2 as the sole oxidant (i.e., anaerobic), a linear:branched of 11.0(2):1 was observed (Table 3). The selectivity for linear products significantly decreases to 5.7(1):1 at otherwise identical conditions but in the presence of 1 atm of dioxygen. This finding provides evidence that the active Rh catalyst and/or reaction pathway are likely changed upon introduction of dioxygen. We speculated that this might be attributable to formation of dioxygen-coordinated intermediates. Use of dioxygen as the sole oxidant results in a linear:branched selectivity of 4.6(6):1, indicating the possibility of a distinct active species from the one formed in the presence of Cu(OPiv)2. In the presence of 1 atm of dioxygen and 480 equiv of Fe(OAc)2, a linear:branched ratio of 2.21(3):1 was observed, which is substantially lower than observed with the other oxidants. Use of Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2 as the oxidant under anaerobic conditions gives a similar linear:branched selectivity of 2.15(1):1. The observation that no significant variance in selectivity was observed for processes with Fe-based oxidants under aerobic versus anaerobic conditions could indicate that the active species is likely not sensitive to dioxygen, which contrasts to catalytic reactions using Cu(OPiv)2.
Linear:branched selectivity was also studied with tert-butyl ethylene, styrene and methyl acrylate, and a similar trend in linear:branched selectivity to that observed with propylene was found, linear:branched selectivity changes as function of oxidant: Cu(II) > Cu(II)/O2 > O2 > Fe(II)/O2 ≈ Fe(III). While Cu(OPiv)2 and dioxygen produced trace branched product with electronically biased methyl acrylate or styrene as the olefin, use of Fe(III) yielded more significant branched product formation.[43]
Using Ir(I) in place of Rh(I) as catalyst precursor results in a significant increase in linear:branched ratio, albeit with a significantly reduced reaction rate. With 0.005 mol% of [(η2-COE)2Ir(μ-Cl)]2 as the catalyst precursor in benzene and 240 equiv of Cu(OHex)2 (OHex = 2-ethylhexanoate) (relative to single Ir atom), 960 equiv of HOHex and 30 psig of propylene, a linear:branched selectivity of 18(1.3):1 was observed after heating for three hours at 150 °C.[42] Upon heating for 42 hours, the selectivity increased to 42(2):1.[42] After control experiments indicated that the observed increase was not the result of an isomerization or selective decomposition of the branched product, we speculated that an increase in Cu(OPiv) concentration as the reaction progresses might be the origin of this observation. Use of an initial Cu(I):Cu(II) of 4:1 resulted in a linear:branched selectivity of 65(3):1 after 18 h, and no statistically significant variation was observed over time. From these results, we hypothesized that hetero-multimetallic species of Cu(II) and/or Cu(I) can form with Ir and the Cu(I) containing complexes more selectively form linear propenylbenzene products. [42] This mirrors findings with Rh,[23,25,43]232543 and Pd catalysis that hetero multi-metallic species with Cu(II) form under the reaction conditions and serve as the likely active catalysts.
We studied the use of benzoquinone and impact on linear:branched selectivity using [(η2-C2H4)2Rh(μ-OPiv)]2 at 170 °C using 240 equiv of ortho-benzoquinone derivatives as the in-situ oxidant in the presence of 960 equiv of HOPiv and 50 psig of propylene.[30] Use of ortho-chloranil as the oxidant results in a linear:branched of 2.6(1):1, 9,10-phenanthrene dione gives a selectivity of 3.2(1):1, 1,2-naphthoquinone 4.8(1):1 and 3,5-di-tert-butyl-ortho-benzoquinone 4.7(4):1 (Table 3). Since benzoquinone derivatives can react with Rh(I) to form catecholate or semiquinone complexes, we speculated that ligand effects of the quinones might be the origin of the selectivity variation. Quinone donor ability was approximated using DFT-calculated pKa1 values, and it was found that linear selectivity increases as ligand donor ability is increased.[58] Notably, the selectivity did not correlate with quinone oxidizing ability, indicating that ligand effects, rather than Rh–H oxidation kinetics, likely dictate selectivity patterns.[30] With para-benzoquinones, a linear:branched selectivity between ~1:1 and 4:1 were determined with no clear trend as a function of benzoquinone donor or oxidizing ability. We attributed this observation to para-benzoquinones either not coordinating to Rh or not having a significant influence on linear:branched selectivity if coordinated.[30]
Table 3: Linear to branched selectivity for benzene propenylation as a function of catalyst precursor and oxidant identity. Eq. 3.
| Entry | Catalyst precursor | Oxidant | Temp (°C) | L1 % Selectivity | L2 % Selectivity | L3 % Selectivity | B1 % Selectivity | L:B |
|---|---|---|---|---|---|---|---|---|
| 1 | Rh1 | Cu(OPiv)2 | 150 | 44 | 7 | 41 | 8 | 11.0(2):1 |
| 2 | Rh1 | Cu(OPiv)2/O2 | 150 | 44 | 7 | 34 | 15 | 5.7(1):1 |
| 3 | Rh1 | Fe(OAc)2/O2 | 150 | 7 | 14 | 47 | 32 | 2.21(3):1 |
| 4 | Rh1 | Fe₆ | 150 | 8 | 14 | 46 | 32 | 2.15(1):1 |
| 5 | Rh1 | O2 | 150 | 42 | 7 | 35 | 16 | 4.6(6):1 |
| 6 | Ir1 | Cu(OHex)2 | 150 | - | - | - | - | 89(2):1 |
| 7 | Rh2 | 3,5-di-tert-butyl-ortho-benzoquinone | 170 | 19 | 9 | 54 | 18 | 4.7(4):1 |
| 8 | Rh2 | o-naphthoquinone | 170 | 19 | 5 | 59 | 17 | 4.8(1):1 |
| 9 | Rh2 | 9,10-phenanthrenequinone | 170 | 30 | 3 | 44 | 23 | 3.2(1):1 |
| 10 | Rh2 | o-chloranil | 170 | 28 | 11 | 33 | 28 | 2.6(1):1 |
| 11 | Rh2 | anthraquinone | 170 | 23 | 8 | 44 | 25 | 3.0(1):1 |
| 12 | Rh2 | tetramethylquinone | 170 | 34 | 32 | 14 | 20 | 4.1(4):1 |
| 13 | Rh2 | 2,5-di-tert-butyl-p-quinone | 170 | 14 | 7 | 46 | 33 | 2.1(1):1 |
| 14 | Rh2 | p-benzoquinone | 170 | 27 | 9 | 38 | 26 | 2.85(4):1 |
| 15 | Rh2 | 2-chloroquinone | 170 | 26 | 11 | 39 | 24 | 3.1(1):1 |
| 16 | Rh2 | 2,5-dichloro-p-benzoquinone | 170 | 26 | 9 | 40 | 25 | 3.1(1):1 |
| 17 | Rh2 | p-chloranil | 170 | 27 | 12 | 34 | 27 | 2.65(7):1 |
| 18 | Rh2 | p-fluoranil | 170 | 10 | 5 | 35 | 50 | 1.03(9):1 |
Rh1 = [(η²-C₂H₄)₂Rh(μ-OAc)]₂; Rh2 = [(η²-C₂H₄)₂Rh(μ-OPiv)]₂; Ir1 = [Ir(μ-Cl)(coe)2]2; Fe6 = Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2. Detailed reaction conditions can be found in original publications. Use of “–” indicates that the individual product selectivity’s were not reported. L:B ratios represent the average of at least three independent experiments, and the values in parentheses represent the standard deviation in the final digit(s).
Regioselectivity for alkenylation of mono-substituted arenes
Toluene ethenylation reactions were studied at 150 °C using 0.112 mM of [(η2-C2H4)2Rh(μ-OAc)]2 (relative to single Rh atom) as the catalyst precursor with 480 equiv of HOPiv and 45 psig of ethylene.[43] With each oxidant, trace ortho-functionalization was observed. In the presence of 240 equiv of Cu(OPiv)2 under anaerobic conditions, a meta:para ratio of 1.4(1):1 was observed, and addition of 1 atm of air in combination with Cu(OPiv)2 results in a 1.1(1):1 meta:para ratio (Table 4). Consistent with studies of linear:branched selectivity using mono-substituted olefins (see above), the difference between aerobic and anaerobic conditions with Cu(OPiv)2 is significant. Use of 1 atm dioxygen as the sole oxidant yielded a 1.4(1):1 meta:para ratio. In the combination of Fe(OAc)2 and 1 atm dioxygen, a meta:para of 1.6(1):1 was observed, while use of Fe₆(μ-OH)₂(μ₃-O)₂(μ-OPiv)₁₂(HOPiv)₂ under anaerobic conditions gave a 1.9(1):1 ratio.[28] Similar trends were observed with tert-butylbenzene and chlorobenzene: trace or no ortho functionalization with meta:para ratios between ~1:1 and 2:1 were observed. With anisole as the olefin, Cu(II) carboxylates gave more significant ortho functionalization, possibly as a result of a weak directing effect from the oxygen atom.[28]
In a previous report, we found that meta:para selectivity using Cu(OPiv)2 as the oxidant with [(η2-C2H4)2Rh(μ-OAc)]2 as the catalyst precursor for the alkenylation of mono-substituted arenes (e.g., toluene, ethylbenzene, tert-butylbenzene) with mono-substituted olefins (e.g., ethylene, propylene, vinyl cyclohexane) under anaerobic conditions is likely controlled by a combination of thermodynamic and kinetic contributions (Scheme 4).[24] Since arene C–H activation is reversible, under conditions with high [HOPiv] and low [C2H4], an equilibrium between meta and para Rh–aryl intermediates likely contributes to the observed ortho:meta:para regioselectivity with ratios close to ~2:1. In contrast, reactions performed with low [HOPiv] and high [C2H4] give meta:para ratios close to ~1:1. We attributed this to para C–H activation being kinetically favored due to sterics with higher ethylene concentration decreasing the reversibility of toluene C–H activation by increasing the rate of ethylene insertion.[24] This effect was particularly pronounced with anisole as the arene, which we studied in detail in a subsequent study.[26]
Using Pd(OAc)2 as the catalyst precursor under otherwise identical conditions to the Rh catalysis with Cu(OPiv)2 as the oxidant was performed in the presence of 1 atm of air. For the reaction of toluene with ethylene an ortho:meta:para selectivity of ~0.5:1.0:1.0 was observed. The greater predilection towards ortho functionalization observed for Pd catalysis was attributed to the C–H activation step containing significant electrophilic character for Pd, whereas Rh catalysis possibly activates C–H bonds through an oxidative addition pathway, a rationalization that is consistent with Density Functional Theory (DFT) calculations.[25] The Sanford group studied toluene alkenylation with ethyl acrylate using Pd(OAc)2 as the catalyst precursor in the presence of pyridine ligands using PhCO3tBu as the oxidant and observed an ortho:meta:para ratio of ~1:1.2:1.3, a greater predilection towards ortho-functionalization than observed with Cu(OPiv)2 as the oxidant.[11] Fujiwara and coworkers reported Pd(OAc)2 catalyzed arene alkenylation in the presence of tert-butylhydroperoxide and benzoquinone with a reported 1:1:3 ortho:meta:para ratio for alkenylation of toluene with ethyl acrylate.[56] Similarly, with Pd(OAc)2 as the catalyst precursor, Bäckvall and coworkers reported around a 1:1:2.5 selectivity for reaction of butyl acrylate with toluene using benzoquinone as the direct oxidant and Fe(II) phthalocyanine as a catalyst to re-oxidize hydroquinone to benzoquinone with dioxygen.[55] You and coworkers reported Pd(OAc)2-catalyzed conversion of arenes with electron-deficient olefins using (NH4)2S2O8 as the oxidant, and observed around a 1.3:1.0:2.9 ortho:meta:para ratio for reaction of toluene with methyl acrylate.[59]
Meta to para regioselectivity with quinones was studied with tert-butylbenzene as the arene substrate and ethylene as the olefin using 0.112 mM of [(η2-C2H4)2Rh(μ-OPiv)]2 as the catalyst precursor, 240 equiv of quinone (relative to single Rh atom), 70 psig of ethylene and 960 equiv of HOPiv at 170 °C.[30] Variation of para-benzoquinone identity resulted in meta:para selectivity between 1.31(2):1 and 1.78(6):1 with no observation of ortho product. These ratios did not strongly correlate with quinone oxidizing ability or donor ability of putative Rh-coordinated ligands. In contrast, meta selectivity with ortho-benzoquinones increased as quinone donor ability increased. Use of ortho-chloranil as the oxidant resulted in a meta:para of 1.49(4):1, 9,10-phenanthrene dione gave 2.03(7):1 selectivity, 1,2-naphthoquinone 2.0(1):1 and 3,5-di-tert-butyl-ortho-benzoquinone 2.8(1):1. We attributed this to an ortho-chloranil semiquinone or catecholate ligand forming an electron-deficient Rh complex, which results in a more electrophilic C–H activation process, favoring the para position. In contrast to the case of ortho-chloranil, a comparatively electron-rich Rh complex forms upon reaction with 3,5-di-tert-butyl-ortho-benzoquinone, and a classical concerted metalation-deprotonation could occur, favoring the meta position.[30]
To provide more definitive evidence for differences in C–H activation mechanism, our group has performed intermolecular competition arene ethenylation experiments with equimolar quantities of toluene and α,α,α-trifluorotoluene. [25,30,43]253043 For Rh catalysis using Cu(II) carboxylates as the oxidant under both anaerobic and aerobic conditions, toluene and α,α,α-trifluorotoluene undergo ethenylation at statistically identical rates.[25,43]2543In contrast, for catalysis using Pd(OAc)2 as catalyst precursor, toluene reacts ~10-fold faster than electron-deficient α,α,α-trifluorotoluene, strengthening evidence for the C–H activation containing electrophilic character for the Pd-based catalysis.[25] For Rh catalysis with dioxygen as the oxidant, toluene reacts more rapidly than α,α,α-trifluorotoluene, also consistent with its C–H activation step containing at least some electrophilic character. Use of Fe(OAc)2 under aerobic conditions results in α,α,α-trifluorotoluene reacting slightly faster than toluene, which is potentially consistent with a classical concerted metalation-deprotonation, which favors acidic C–H bonds, occurring.[43] Similarly, with 3,5-di-tert-butyl-ortho-benzoquinone as the oxidant, α,α,α-trifluorotoluene reacts faster than toluene, indicating significant deprotonation character in the C–H activation step. Use of 9,10-phenanthrene dione results in approximately identical rates of toluene and α,α,α-trifluorotoluene alkenylation, and ortho-chloranil ethenylated toluene faster than α,α,α-trifluorotoluene. These results indicate that when coordinated to Rh, ortho-quinone ligands can control C–H activation mechanism on a continuum from concerted metalation deprotonations with electrophilic character to those without electrophilic character.[30,60,61]306061
Table 4: meta:para regioselectivity for toluene ethenylation reactions under different conditions with different oxidants. Eq. 4.
| Entry | Catalyst precursor | Oxidant | Ortho functionalization % Selectivity | Meta functionalization % Selectivity | Para functionalization % Selectivity |
|---|---|---|---|---|---|
| 1 | Rh1 | Cu(OPiv)2 | Trace | 58 | 42 |
| 2 | Rh1 | Cu(OPiv)2/O2 | Trace | 52 | 48 |
| 3 | Rh1 | Fe(OAc)2/O2 | Trace | 61.5 | 38.5 |
| 4 | Rh1 | Fe6 | Trace | 65.5 | 34.5 |
| 5 | Rh1 | O2 | Trace | 58 | 42 |
| 6 | Pd(OAc)2 | Cu(OPiv)2 | 20 | 40 | 40 |
Rh1 = [(η²-C₂H₄)₂Rh(μ-OAc)]₂; Fe6 = Fe6(μ-OH)2(μ3-O)2(μ-OPiv)12(HOPiv)2. Detailed reaction conditions can be found in original publications.
Table 5: Comparison of ortho:meta:para regioselectivity in reported Pd(OAc)₂-catalyzed nondirected C–H alkenylation of toluene with acrylate coupling partners. Detailed reaction conditions can be found in original publications.
| Entry | Arene | Olefin | Catalyst precursor | Oxidant | Ligand/Additive | Time (h) | Temp (°C) | o:m:p ratio |
|---|---|---|---|---|---|---|---|---|
| 1 | toluene | ethyl acrylate | Pd(OAc)2 | PhCO₃tBu | Pyridine | 6 | 100 | 1:1.2:1.3 |
| 2 | toluene | ethyl acrylate | Pd(OAc)2 | t-BuOOH & Benzoquinone | – | 12 | 90 | 1:1:3 |
| 3 | toluene | butyl acrylate | Pd(OAc)2 | Benzoquinone/O₂ | Fe(II) phthalocyanine | 24 | 90 | 1:1:2.5 |
| 4 | toluene | methyl acrylate | Pd(OAc)2 | (NH₄)₂S₂O₈ | – | 24 | rt | 1.3:1:2.9 |
Conclusions
Herein, we have overviewed the effect of oxidant identity on catalytic arene alkenylation with a primary focus on our group’s studies of Rh-catalyzed arene alkenylation and added discussion of Pd- and Ir-based catalysis. We discussed Cu(II) carboxylates, Fe(III) carboxylates, dioxygen and benzoquinones and their influence on reaction rate and selectivity. Overall, we conclude:
- The rate of styrene production follows the trend Cu(II) > Cu(II)/O2 > Fe(II)/O2 = Fe(III) > O2 for Rh-catalysis. Selectivity for styrene follows the trend O2 > Cu(II) > Fe(II)/O2 > Fe(III) > Cu(II)/O2. In addition, ortho-benzoquinones generally give a faster rate of catalysis compared to para-benzoquinones.
- For Pd-catalysis with a Cu(II) oxidant, styrene production is increased under aerobic conditions; however selectivity for styrene decreases.
- Linear:branched selectivity follows the trend Cu(II) > Cu(II)/O2 > O2 > Fe(II)/O2 ≈ Fe(III) for Rh-catalysis when using propylene as the olefin. Ortho-benzoquinones gave linear:branched selectivities ranging between 2.6:1-4.8:1, while para-benzoquinones ranged between 1:1-4:1.
- With Ir-catalysis with a Cu(II) oxidant an increase in linear:branched ratio is observed over time, which we speculate is due to the increase in Cu(I) concentration as the reaction progresses. As a result, Cu(I)-containing hetero multi-metallic complexes are thought to form under the reaction conditions and favor the production of linear alkenyl arenes.
- For Rh-catalyzed alkenylation of mono-substituted arenes, distinct ortho/meta/para regioselectivity is observed as a function of oxidant identity. Cu(II) and Cu(II)/O2 have significantly different regioselectivities, indicating there is a difference in mechanistic pathway and/or catalyst identity. With Pd catalysis, the regioselectivity for alkenylation of substituted olefins is more sensitive to arene substituent than Rh catalysis, and greater ortho and para functionalization is observed with electron-rich arenes compared to Rh catalysis.
- For mono-substituted arenes, meta selectivity increases as ortho-quinone donor ability increases.
