Direct activation of relatively unstrained carbon–carbon bonds in homogeneous systems

Abstract

New modes of chemical reactivity are of high value to synthetic organic chemistry. In this vein, carbon–carbon (C–C) activation is an emerging field that offers new possibilities for synthesizing valuable complex molecules. This review discusses the pioneering stoichiometric discoveries in this field up to the most recent synthetic applications that apply catalytic transformations. Specifically, the review focuses on C–C activation in relatively unstrained systems, including stoichiometric reactions, chelation-directed and chelation-free catalytic reactions. While the field of C–C activation of relatively unstrained systems is underdeveloped, we expect that this review will provide insight into new developments and pave the path for robust, practical applications.

New modes of catalytic chemical reactivity can provide powerful transformations for synthetic chemists to attain higher efficiency and more direct routes to desired molecules. Two areas of research that fit the criteria for novel mechanisms and transformations include carbon–hydrogen (C–H) and carbon–carbon (C–C) activation. While catalytic C–H activation has garnered significant recent attention by the synthetic community, reports of C–C activation methods are less prominent in the literature. C–C Activation is often viewed as a destructive mode of reactivity, counter to the primary focus on developing methods to form C–C bonds rather than to break them. This review will show how selective functionalization of the metal center in conjunction with C–C activation can lead to unique methods for constructive bond formation.

When compared to C–H activation, C–C activation is kinetically challenging. C–H bonds are typically more abundant and C–C bonds are hindered. Therefore, C–H bonds more readily display proper orbital overlap trajectories with the metal center than C–C bonds. Thermodynamics play less of a role as C–C oxidative addition can in fact be exothermic, although cleavage of C–C bonds can also be thermodynamically uphill. Methods for transition-metal-mediated C–C activation employ strategies mainly involving strain-release, aromatization, and chelation-assistance (proximity effect) to overcome these barriers. The latter two have proven useful for the activation of unstrained C–C bonds, which remain particularly challenging in the context of C–C activation of unstrained C–C bonds.

This review primarily focuses on the direct activation of C–C bonds in relatively unstrained systems, i.e. non-three or four-membered ring systems. While not intended to comprehensively cover all literature references, it rather provides a perspective on the scope of reactivity through selected examples to highlight representative reactions types. Chronological order is roughly followed to loosely track the progress of research for the activation of unstrained C–C bonds. Three main sections will be discussed: (1) stoichiometric C–C activation reactions, (2) chelation-directed catalytic C–C activation reactions and (3) chelation-free catalytic C–C activation reactions. Olefin metathesis, retro-aldol reactions, β-carbon elimination reactions, C–CN activation,¹ decarboxylation,² deallylation,³ solid-state photochemical⁴ reactions, diazoalkane-carbonyl homologation,¹ Baeyer–Villiger oxidation and other oxidative C–C cleavage reactions, while all involving C–C bond breaking or rearrangement, will not be discussed here as these topics are either beyond the scope of the review or have been thoroughly reviewed elsewhere.⁵

Stoichiometric reactions

One of the first examples of C–C activation in an unstrained system was reported by Rusina and co-workers in 1965.⁶ They observed that when RhCl₃ and PPh₃ were heated in a number of different alcoholic solvents, they obtained golden-yellow crystals of Rh^ICl(CO)(PPh₃)₂. The same complex was isolated when conducted in cyclohexanone and acetophenone at reflux. Although no mechanism was proposed, it is likely that Rh^ICl-(CO)(PPh₃)₂ formation is the result of a C–C activation event. Furthermore, they suggest that elevated temperatures were required as no complex was formed when the reaction was conducted in acetone (b.p. 57 °C).

(1)

A few years after the work of Rusina, Müller and coworkers^7,8 showed that Wilkinson's complex RhCl(PPh₃)₃ can cleave the sp–sp² bonds of diynones (1), resulting in decarbonylation and reductive elimination to produce diynes (2) and RhCl(CO)(PPh₃)₂ (eqn (2)). A stoichiometric amount of RhCl-(PPh₃)₃ was required for full conversion as the RhCl(CO)-(PPh₃)₂ carbonyl byproduct does not catalyze the reaction. This method was demonstrated to work on a number of different symmetrical and unsymmetrical diynones with varying electronic and sterics properties to produce the corresponding diynes (20–93%). In one example, a monoynone produced a disubstituted alkyne in 8% yield. A mechanistic discussion in the catalytic version of this reaction will be presented in a later section of this review.

Wilkinson's complex effectively decarbonylates unstrained 1,2- and 1,3-diketones (3), as shown by Teranishi and coworkers⁹ who observed monoketone products (4) under refluxing toluene (eqn (3)). They reported the isolation of an acetylacetonato complex, which proved inactive as a catalyst. However, the Rh-carbonyl complex isolated from the reaction was found to be active in the decarbonylation of acetylacetone to provide methyl ethyl ketone. Their studies highlighted specificity of Rh in the decarbonylation of 1,2- and 1,3 diketones as other catalysts, such as RhCl₃–3H₂O, IrCl₃–3H₂O, PdCl₂, and CoCl₂–6H₂O, were screened but failed to promote decarbonylation chemistry.

In 1971, King and co-workers showed that 5-acetyl-1,2,3,4,5-pentamethylcyclopentadiene (5) undergoes C–C bond cleavage when treated with Co₂(CO)₈ in cyclohexane at 110 °C for 22 h to form stable complex 6 (eqn (4)). Eilbracht and Dahler¹⁰ later demonstrated that alkyl substituents can also participate in C–C activation via aromatization. Diene 7 reacts with Fe₂(CO)₉ to form a stable complex (8), which when heated with additional Fe₂(CO)₉ in benzene at 80 °C provides a cyclic Fe species 9. Furthermore, Crabtree and co-workers¹¹ also showed that iridium complexes can form Cp-complex via C–C bond cleavage in various gem-dialkyldienes (eqn (6)). Initially, [IrH₂(Me₂CO)₂((p-FC₆H₄)₃P)₂] underwent dehydrogenation with compound 10 to form diene complex 11, and demethylation via C–C activation provided complex 12. While not all examples of aromatization strategies for C–C activation are covered here,¹² these examples demonstrate that aromatization can drive C–C activation with a variety of different metals and substrates.

(2)

(3)

(4)

(5)

(6)

Chelation-assisted activation of α C–C bonds to a carbonyl group was first demonstrated by Suggs and co-workers¹³ through insertion of Rh into a series of quinoline-derived substrates under mild conditions (rt to 40 °C) (13, eqn (7)). This strategy was found applicable for the activation of alkynyl and alkyl ketones; however, styryl ketones did not react even at high temperature. Deuterium labeling studies showed that in this case C–H activation of the alkyl substituents did not occur prior to C–C activation. Furthermore, this approach was used to synthesize chiral rhodium alkyl species through the incorporation of an α chiral substituent (15, Scheme 1). When the chiral metal-complex 16 was treated with P(OMe)₃, reductive elimination returned the starting material with no loss of optical purity. Given that reductive elimination typically occurs with retention of configuration¹⁴ C–C activation must also occur through retention of configuration. The authors proposed that the reaction followed a mechanism similar to the Baeyer-Villiger reaction through a tetrahedral intermediate.

Scheme 1.

Rh-promoted C–C activation of a chiral quinoline derived substrate.

(7)

In 1989 Bergman and co-workers¹⁵ reported the synthesis of a highly reactive Ru benzyne complex 17 that was able to active C–H, C–N, N–H, O–H and C–C bonds under relatively mild conditions.¹⁶ When the catalyst is heated with acetone at 45 °C for 1.5 days, Ru complex 18 was obtained in 28% isolated yield and ¹H NMR analysis identified methane as a byproduct (eqn (8)). Bergman later identified a hydroxyruthe-nium complex 19, also capable of promoting C–C activation under mild conditions. When hexafluoroacetone was treated with 19 in benzene at room temperature for 1 h, complex 20 was formed along with an equivalent of fluoroform. By the mechanism shown in Scheme 2, dissociation of hydroxide permits reversible formation of cationic complex 21, allowing the substrate to bind giving intermediate 22. Hydroxide ion attack on the carbonyl forms species 23, which undergoes C–C cleavage to produce fluoroform and complex 20.

Scheme 2.

Ru-promoted C–C activation of hexafluoroketone.

(8)

In 1992¹⁷ and subsequent papers¹⁸ in following years, Rosenthal and co-workers reported that titanium and zirconium could be used to activate the internal C–C bond of diynes (24, Scheme 3). A general scheme is provided to illustrate the basic mechanism of this reaction. Most characteristic to the mechanism is that the reaction goes through a metallocyclocumulene to give the observed dimeric products (25). Crystallographic data was obtained to support the proposal of this mechanism. This work demonstrates that early transition metals are also capable of C–C bond activation.

Scheme 3.

Zr- and Ti-mediated C–C activation of diynes.

The activation of strong unstrained C–C bonds was demonstrated by Milstein and co-workers¹⁹ with pincer complexes. When bisphosphine 26 was reacted with HRh(PPh₃)₃ at room temperature in THF, a thermally stable C–H activation product was obtained in quantitative yield (27, Scheme 4). When this complex was heated at 90 °C in the presence of H₂, the C–C activation product (28) was formed quantitatively. The mechanism of C–C activation likely involves reversible C–H activation followed by subsequent C–C activation. These results suggest that the C–C activation and formation of a strong Rh–Ar bond is thermodynamically more favorable than C–H activation. Milstein and co-workers^20,21 indicated that the C–C activation event proceeds through a three-centered mechanism rather than an η²-arene complex suggesting that the metal center is perfectly positioned for direct C–C activation. Mil-stein and co-workers²² also studied an unsymmetrical pincer ligand bearing a phosphine and amine chelate (29). C–C Activation occurs exclusively (eqn (9)) under mild conditions (rt to 45 °C) to provide complex 30 with no evidence for C–H activation or initial complexation of Rh to the ligand. This observation may be explained by the less sterically demanding amine ligand with its hard electronic influence versus a soft phosphine promoting rapid reversible C–H activation. These experiments demonstrate the feasibility of activating unstrained C–C bonds, and support the notion that the cleavage can be favored from both kinetic and thermodynamic respects under the correct circumstances.

Scheme 4.

Activation of strong C–C bonds in pincer ligands reported by Milstein and co-workers.

(9)

In 1994, Ito and co-workers showed that Rh-catalyzed C–C activation can take place with strain-free cycloalkanones to provide strained ring-contracted products.²³ Cyclopentanone derivative 31 was decarbonylated using an equimolar amount of Wilkinson's catalyst to provide a strained cyclobutane (32) in 57% yield after refluxing for 8 days (eqn (10)). Larger rings, such as cyclododecanone, were decarbonylated after heating at higher temperature (150 °C) for 3 d to provide an eleven-membered carbocycle, albeit in much lower yield (20%, eqn (11)).

(10)

(11)

Jones and co-workers²⁴ presented a unique case where Pt can cleave the sp²–sp C–C bonds of diphenyl acetylene under photochemical conditions.²⁵ Shedding UV light onto a series of air- and moisture-sensitive Pt-η²-alkyne-complexes (33a–e) afforded unsymmetrical Pt^II complexes 34a–e with a phenyl group σ-bonded to the metal center (eqn (12)). The structures were confirmed by single crystal X-ray diffraction, and ¹H NMR analysis shows platinum satellites for a single set of ortho-phenyl protons indicating that the phenyl group is indeed σ-bonded to the Pt center. Interestingly, when specific complexes 34b and 34e were thermally activated, a reductive elimination took place to reform the Pt-η²-alkyne complexes (33b and 33e, eqn (13)). The reversible nature of the reaction under thermal conditions suggests that oxidative cleavage of sp²–sp C–C bonds in these particular complexes is thermodynamically uphill.²⁶

(12)

(13)

In 2004, Daugulis and Brookhart²⁷ reported the use of a Rh complex capable of decarbonylation. For example, 3,3′-bis(tri-fluoromethyl) benzophenone (35) was decarbonylated with Cp*Rh(C₂H₄)₂ (36) in refluxing toluene to give biaryl product 37 in 82% yield (based on catalyst, eqn (14)). Catalyst 36 was also found to successfully decarbonylate chalcone. Heating chalcone with Cp*Rh(C₂H₄)₂ at 120 °C forms complex 38, however upon further heating at elevated temperatures (150 °C) decarbonylation occurs to form stilbene in 36% yield (Scheme 5). A number of other benzophenones and aceto-phenones were decarbonylated and while electronic factors did not influence the rate of decarbonylation with benzophenones, more electron-withdrawing substituents (e.g. CF₃) on aceto-phenones gave substantially faster reaction rates.

Scheme 5.

Decarbonylation of chalcone by unique Rh complex 36.

(14)

Following the work of Daugulis and Brookhart, Ozerov and co-workers²⁸ presented an electron-rich pincer-ligated Ru complex (PNP)(Ru)H₃, 39) that can successfully decarbonylate acetone. Heating of acetone (4 equiv.) with 39 in fluorobenzene at 75 °C for 18 h resulted in the decarbonylation of acetone to form 40 in 95% conversion (eqn (15)). The proposed mechanism is shown in Scheme 6, where 39 cleaves a single α C–C bond of acetone to provide ruthenium-acyl complex 41. Reductive elimination produces an equivalent of methane (observed by ¹H NMR) and complex 42, which can react with hydrogen to extrude another equivalent of methane and result in the formation of complex 40. Ozerov and co-workers conducted DFT studies, in addition to ¹H NMR experiments, to support the mechanistic rationale and the exothermic nature of the reaction (−51.5 kcal mol⁻¹). They proposed two main thermodynamic driving forces for the reaction. The first involves the favorable binding of CO over H₂ by 38.9 kcal mol⁻¹ to the metal center. The second is preferable formation of C–H bonds over C–C bonds by 19.9 kcal mol⁻¹, supporting the production of two equivalents of CH₄ rather than one equivalent of C₂H₆.

Scheme 6.

Demethylation of acetone via C-C activation.

(15)

Another form of chelate-assisted C–C activation was reported by Ruhland and co-workers showing that sp²–sp² C–C bonds in biaryls and benzophenones can be successfully cleaved with phosphonite directing groups.²⁹ For instance, when biaryl compound 43 was treated with Ni(PPh₃)₂(CO)₂ at room temperature, the two phosphinite groups coordinated to the metal center and extruded two equivalents of PPh₃ to provide complex 44 (Scheme 7). Upon heating of 44 under 5 bar CO, benzophenone complex 45 was afforded via C–C activation and CO insertion The reverse reaction was also shown to be possible in the absence of CO, where ligand 46 was treated with Ni(PPh₃)₂(CO)₂ to afford complex 45 and upon heating (95 °C) decarbonylation occurred to give complex 44 (20%, 4 days). Decarbonylation was also observed when 46 was treated with Ni(COD)₂, to give a 4 : 1 mixture of complexes 47a and 47b, respectively (Scheme 8). The formation of 47a and 47b arises from two degenerate intermediates 48a and 48b, which are in equilibrium as observed by NMR analysis. Extensive experimentation and further NMR analysis provided a mechanistic understanding of the carbonylation and decarbonylation reactions. In the case of carbonylation reaction, they proposed that CO dissociation occurred prior to C–C activation, whereas in the decarbonylation pathway CO deinsertion occurred after electron-rich Ni(0) oxidatively inserted into the C–C bond. Similar C–C activation of benzophenone derivatives was observed by Obenhuber and Ruhland with other transition metals such as Ir (e.g. Vaska complex)³⁰ and Rh (e.g. [Rh-(COE)₂Cl]₂).³¹

Scheme 7.

Phosphonite-directed C–C activation and carbonylation with Ni.

Scheme 8.

C–C activation and decarbonylation of benzophenone with Ni as the metal center.

Catalytic chelation-directed C–C activation of unstrained substrates

In 1998, Milstein and co-workers³² described a catalytic version of the C–C activation with their bidentate phosphine substrates in converting 49 to 50 under H₂ pressure or with excess HSi(OEt)₃ (eqn (16)). Although not optimized at the time, [RhCl(COE)₂]₂ showed 100 turnovers with H₂. The proposed catalytic cycle is shown in Scheme 9. The first step involves hydrogenation of the COE ligand and complexation to the bidentate substrate to form 51. At this stage C–C activation can ensue to achieve Rh^III intermediate 52, at which point hydrogenolysis of the Rh–methyl bond produces methane and complex 53. Reductive elimination results in Rh^I 54, which can undergo exchange with the substrate to produce 50 and allow Rh^I to reenter the catalytic cycle.

Scheme 9.

Proposed catalytic cycle for C–C activation of PCP ligand.

(16)

A year after Milstein's report on catalytic C–C activation, Murai and co-workers reported a chelation-assisted decarbonylation reaction via activation of unstrained C–C bonds.³³ They reported the use of Ru₃(CO)₁₂ (5 mol%) under 5 atm CO in toluene at 160 °C to undergo oxazoline-directed decarbonylative C–C cleavage of alkyl phenyl ketones (55) to the corresponding products in 34–96% yield (56, eqn (17)). They also demonstrated that the directing group is necessary, as in substrate 57 decarbonylation only occurred ortho to the directing group to furnish 58 in 85% yield (eqn (18)).

(17)

Murai and co-workers propose two possible mechanistic pathways that lead to the observed product (Scheme 10). Coordination of Ru to the starting material (55) followed by nucleophilic attack at the carbonyl provides complex 59. At this point there are two possible C–C bonds that can be cleaved, the Ar–CO C–C bond or the CO–alkyl C–C bond. Cleavage through pathway A would lead to six-membered rhodacycle 60, which after decarbonylation (to give 61) and β-hydride elimination would provide the desired product. Alternatively, pathway B would provide the five-membered rhodacycle 62, which can undergo decarbonylation to provide the same intermediate (61) or it can first undergo β-hydride elimination to give complex 63, which would release a molecule of ketene and reductively eliminate to give the desired product. Murai and co-workers suggest that pathway B is more plausible. Benzyl ketone derived substrate (R = Ph) was reacted in methanol to give 73% yield of the desired product (56) isolated along with 34% of methyl phenylacetate through methanolysis of intermediate 62 or trapping of phenyl ketene.

Scheme 10.

Proposed mechanism for oxazoline-directed C–C activation.

(18)

In the same year, Jun and co-workers³⁴ reported a catalytic C–C activation of unstrained ketones using 2-amino-3-picoline as a cofactor.³⁵ As shown in eqn (19), Wilkinson's catalyst and 2-amino-3-picoline promote C–C activation and alkyl group transfer in the presence of ketones (64) and terminal olefins with good yields overall (42–98%). Internal olefins typically did not work as well giving lower yields. The proposed mechanism of the reaction explains the preferred C–C cleavage on the homobenzyl side of the ketone (Scheme 11). The first step of the reaction involves condensation of 2-amino-3-picoline to provide imine 66 and H₂O. Coordination of Rh(i) to the substrate followed by C–C activation gives rhodacycle 67, which can undergo β-hydride elimination, possible only on the homobenzylic side, to form an equivalent of styrene and complex 68. Migratory insertion into 1-hexene affords intermediate 69, which after reductive elimination provides ketimine 70 and regenerates the Rh(i) catalyst. Hydrolysis of ketimine 70 provided product 65 and 2-amino-3-picoline which can reenter the catalytic cycle. This process was also demonstrated with the C–C activation of secondary alcohols, where alkyl-group-exchange takes place after initial oxidation of the secondary alcohol to the ketone via transfer hydrogenation.³⁶

Scheme 11.

Mechanism of alkyl-exchange via C–C bond activation.

Jun and co-workers further developed a picoline-directed C–C activation strategy for ring-opening and skeletal rearrangement of several different cycloalkanone imines.³⁷ For example, when cycloalkanoketimine 71 is subjected to [(COE)₂RhCl]₂ (3 mol%) and PCy₃ (6 mol%) in toluene at 150 °C for 6 h a mixture of symmetrical and unsymmetrical ring-opened products (72 and 73, respectively) were obtained in good yields (76–89%, eqn (20)). Smaller ring systems such as cyclopentanoketimine and cyclohexanoketimine reacted poorly and gave yields of less than 10%. The mechanism of this reaction (Scheme 12) is similar to that of the previously discussed reaction, where 2-amino-3-picoline directs Rh(i) to cleave the α C–C bond of the ketimine providing intermediate 74. β-Hydride elimination gives Rh–H 75, which can then add across an equivalent of 1-hexene to provide complex 78. Hydrolysis of 78 gives product 73; alternatively, 78 can undergo another cycle of alkyl-group transfer to give the symmetrical ketimine 79, ultimately leading to ketone 72.

Scheme 12.

Proposed mechanism for Rh-catalyzed ring-opening of cycloalkanoketimines.

(19)

(20)

When the reaction with seven-membered ketimine 80 was run without 1-hexene, a rearrangement occurred followed by hydrolysis to produce cyclohexanone 81 and cyclopentanone 82 in 82% yield with 76 : 24 ratio, respectively. Other rings sizes gave lower yields (0–21%). The mechanism of this rearrangement follows a similar sequence as that in Scheme 12. For example, rather than addition to 1-hexene, the Rh^III intermediate 75 reinserts into the appended olefin in a 6-exo fashion to give an intermediate that can reductively eliminate to provide cyclohexanone product 81. On the other hand, for product 82, after reinsertion of the appending olefin, a second β-hydride elimination (alkene chain-walk) followed by another addition across the olefin and reductive elimination would result in cyclopentanone 82. Jun and co-workers also showed that when bicyclic systems such as bicycle[3.2.1]octan-2-one 83 were subjected to the reaction conditions, a 5,5-fused bicyclic ketone 84 is obtained in 25% yield following the mechanism proposed below (Scheme 13).

Scheme 13.

Rh-catalyzed skeletal rearrangement of bicycle[3.2.1]octan-2-one.

(21)

In 2009, Douglas and Dreis³⁸ reported an intramolecular carboacylation reaction. Quinolines were utilized as directing groups for catalytic C–C σ bond activation with Rh. Refluxing substrate 85 with 5 mol% [RhCl(C₂H₄)₂]₂ in toluene effectively results in all-carbon quaternary stereocenter-containing compound 86 in moderate to excellent yields (63–94%, eqn (22)). A number of different substrates with various linkers participate. An intermolecular version of this transformation was reported later that year by the same group on insertion of norbornene-type alkenes.^39,40

(22)

Johnson and co-workers published mechanistic studies regarding the catalytic cycle for these transformations (Scheme 14).⁴¹ They found slightly different mechanistic pathways by investigating the use of RhCl(PPh₃)₃ and [RhCl-(C₂H₄)₂]₂ as catalysts, respectively. With RhCl(PPh₃)₃ the proposed catalytic cycle is shown in Scheme 13, wherein the reaction was found to be zero-order in substrate and first-order in catalyst, and the C–C activation step was determined to be rate-limiting. The resting state of the reaction is when catalyst is bound to the substrate in complex 87. Next, C–C activation and acylation of the metal center provides intermediate 88, which rearranges to the quaternary center in complex 89. Reductive elimination would then produce 90, where the product can be exchanged with another equivalent of substrate allowing the catalyst to reenter the catalytic cycle. PPh₃ may also intercept 90 to free the catalyst. On the other hand, when the reaction was studied for [RhCl(C₂H₄)₂]₂ a slightly different catalytic cycle was proposed. The reaction is second-order overall, first-order in both substrate and catalyst, suggesting that the resting state of the catalyst is when it is not bound to either substrate or product. The rate-limiting step was determined to be C–C activation with 1,1′-disubstituted alkenes or substrates with longer linkers influencing the reaction rate. The main intermediates of catalytic cycle for [RhCl(C₂H₄)₂]₂ are similar to RhCl(PPh₃)₃, with the exception of the resting state of the catalyst.

Scheme 14.

Mechanism for the synthesis of all carbon quaternary centers.

In 2012, Shi and co-workers^42,43 published a pyridine-directed C–C activation strategy for the decarbonylation of biaryl ketones and alkyl/alkenyl aryl ketones (91). The optimized conditions employ 5 mol% [(CO)₂Rh(acac)] in refluxing chlorobenzene to provide the decarbonylated products (92) in 80–97% yield (eqn (23)). A number of different aryl, alkyl, and alkenyl groups participate under the reaction conditions and both electron-donating and electron-withdrawing groups afforded good yields. Nitrogen-based directing groups other than pyridine (e.g. pyrazolyl and oxazolyl) provided decarbonylation products in lower yields (52 and 44%, respectively). The proposed mechanism for the decarbonylation is shown in Scheme 15. The reaction is initiated by coordination and subsequent oxidative addition to form either intermediate 93 or 94, and decarbonylation forms 95 followed by reductive elimination giving the observed product. Shi and co-workers propose the five-membered rhodacycle (93) as the favored pathway, although the other pathway cannot be ruled out. The reaction was not promoted by photoirradiation (as the reaction takes place in the dark).

Scheme 15.

Proposed catalytic cycle for the C–C activation and decar-bonylation of biaryl ketones.

(23)

Catalytic C–C activation without chelation

Fillion and co-workers⁴⁴ showed that unstrained C–C σ bonds in benzylic Meldrum's acid derivatives (96) could be hydrogenolyzed to produce various benzylic products 97 (eqn (24)). When subjected to 15 mol% of 10% Pd/C under 1 atm of H₂ for 24 h in MeOH at room temperature, benzylic products were obtained in good to excellent yields (65–96%). Electronic effects were observed on the phenyl moiety. meta-Substituted substrates (OC₈H₁₇) gave no product whereas the ortho- and para-substituted analogs gave facile hydrogenolysis and high yields. At the benzylic position, sterics played an important factor such that incorporation of an i-Pr group hampered the reactivity giving only modest conversion; whereas methyl substitution gave near full conversion and high yields. When enantioenriched substrates were hydrogenolyzed, an inversion of stereochemistry is observed with only slight erosion of the enantiomeric ratio. Furthermore, deuterium labeling studies showed that full incorporation of deuterium was observed only when both the solvent and gas were labeled and only partial incorporation was observed when run separately. Therefore, it is likely that the reaction follows an S_N2 type mechanism that involves nucleophilic attack by palladium to obtain a benzylic organopalladium intermediate, which undergoes protonation by methanol.

In 2012, Arisawa, Yamaguchi and co-workers⁴⁵ reported a method for preparing unsymmetrical ketones through a Rh-catalyzed acyl-transfer reaction that requires no chelation assistance. Under optimized conditions RhH(CO)(PPh₃)₃ with 1,2-bis(diphenylphosphino)benzene (dppBz) in N,N′-dimethyl-imidazolidinone (DMI) at 150 °C for 12 h results in the transfer of acyl groups between a variety of benzyl ketones (98) and thioesters/aryl esters (99) to provide the respective products 100 and 101. In general, a combination of different substrates gave good yields with the exception of 1,2-diphenylethanone, which gave 21% yield of unsymmetrical benzyl ketone and 18% yield of the thioester. The reaction was also applied to the acyl transfer of benzyl ketones (102) with aryl esters (103) and gave modest to good yields (39–71%) of the benzyl ketone and thioester. Multiple equivalents of the starting thioester favor the forward direction in this equilibrium driven reaction. While the exact mechanism is still unclear, the authors propose that the reaction first undergoes CO–benzyl bond cleavage by a low-valent Rh, followed by exchange with the thioester and reductive elimination to give the product.

In 2013, Dong and co-workers⁴⁶ reported a catalytic version of the Rh-mediated decarbonylation of diynones initially reported by Muller. The key factor for promoting this catalytic process was a bidentate phosphine ligand, which is believed to assist in CO elimination from the metal center allowing regeneration of the active catalyst. The reaction works well for a number of symmetrical and unsymmetrical diynones 106 using 2.5 mol% [Rh(COD)Cl]₂ and 6 mol% 1,1′-bis(diphenyl-phosphino)ferrocene (dppf) in refluxing chlorobenzene to obtain the diyne products 107 (21–91%). The reaction is amenable to substrates that present both electronic and steric variations. This method has been used for natural product modification (eqn (28)). It was also applied to the synthesis of a highly conjugated ynediyne 108 through a decarbonylation followed by Sonagashira coupling, in which the orthogonality of this process to other Pd and Cu-catalyzed methods (i.e. tolerance of aryl halides) was demonstrated, making it a complimentary strategy (Scheme 16).

Scheme 16.

Utilization of the Rh-catalyzed decarbonylation to synthesize highly conjugated rod-like structures.

The proposed mechanism for the decarbonylation is illustrated in Scheme 17. The first step involves coordination of the substrate to the metal center (109), bringing the metal into close proximity for C–C bond cleavage. Oxidative addition of Rh^I gives Rh^III-complex 110, which undergoes decarbonylation and CO migration to form complex 111. Rapid reductive elimination provides the product and Rh carbonyl complex, which is regenerated into its active form through ligand-assisted release of CO upon substrate binding. The first three steps are in principle all reversible. Also, in the case of unsymmetrical diynones, only unsymmetrical diynes are obtained, suggesting that the reaction is strictly intramolecular without observed intermolecular exchange of acetylene units.

Scheme 17.

Proposed mechanism for the Rh-catalyzed decarbonylation of diynones.

(24)

(25)

(26)

(27)

(28)

Chan and co-workers have developed a rhodium phorphyrin for the catalytic C–C activation of aliphatic [2.2]paracyclophane (112, PCP).⁴⁷ Similar chemistry was developed by the group previously in the context of C–C activation using stoichiometric rhodium phorphyrin complexes.^48,49 In this transformation, PCP is catalytically converted to 4,4′-dimethyl-bibenzyl 113 under two possible conditions (eqn (29)). Rh^III(ttp)I (ttp = tetratolylporphyrinato dianion) in the presence of KOH (1 equiv.) provided the cleaved product 113 in 83% yield in 25 h; with Rh^III(ttp)Me, the product was obtained in 78% after 54 h. Independent deuterium labeling experiments showed that water is the source of hydrogen and that the C–C activation does not involve C–H activation intermediates. The proposed catalytic cycle, which is supported by experimental data, is shown in Scheme 18. Hydrolysis of either precatalyst 114 or 115 provides Rh^III(ttp)OH (116), which can decompose to Rh^II(ttp) radical and exist in an equilibrium with Rh^II(ttp)₂. Two equivalents of Rh^II(ttp) react with PCP to form complex 117; kinetic data supports an overall third-order reaction, first order in PCP and second order in Rh^II(ttp). Finally, reaction with H₂O provides 113 and regenerates the initial catalyst 116.

Scheme 18.

The proposed mechanism for Rh-catalyzed cleavage of PCP.

(29)

Conclusion

In summary, a variety of stoichiometric and catalytic C–C activation reactions have been presented with a discussion of the differing mechanistic pathways that are involved. Although many examples may appear to be special cases, they demonstrate the potential to develop generalized methodologies and strategies. With improvements in this field, it should become possible to construct compounds non-traditionally using novel bond disconnections. In fact, as the field has moved forward from stoichiometric reactions to the development of new catalytic C–C activation transformations, the synthetic utility of this approach has become more apparent. Future improvements in this field would involve moving away from specific examples and developing more general methods for catalytic C–C activation of more common substrates. This will likely require the evolution of new types of catalysts or strategies. We expect, in the coming years, C–C activation will emerge to have a profound impact, in both academia and industry, on accessing synthetically useful molecules.

Abbreviations

Acac

Acetylacetone

C–C

Carbon–carbon

C–H

Carbon–hydrogen

C₂H₄

Ethylene

COD

Cyclooctadiene

COE

Cyclooctene

DMI

N,N′-Dimethylimidazolidinone

dppBz

1,2-Bis(diphenylphosphino)benzene

dppf

1,1′-Bis(diphenylphosphino)ferrocene

NMR

Nuclear magnetic resonance

PCy₃

Tricyclohexylphosphine

PCP

[2.2]paracyclophane

P(OMe)₃

Trimethylphosphite

PPh₃

Triphenylphosphine

THF

Tetrahydrofuran

ttp

Tetratolylporphyrinato dianion

Biographies

Alpay Dermenci was born in Huntington Beach, CA (USA) in 1984. He received his B.S. from University of California, Irvine and carried out his doctoral studies under the guidance of Prof. Scott J. Miller at Yale University. His research focused on the application and development of amino acids and peptides as catalysts for the synthesis of natural products. Following his Ph.D., he conducted post-doctoral studies with Prof. Guangbin Dong at UT Austin focusing on the development of C–C activation methodology. Alpay is currently a research scientist at Pfizer, Inc. in Groton, CT.

Jotham W. Coe, a Research Fellow at Pfizer, is a medicinal chemist in BioTherapeutics working towards treatments for inflammatory conditions. Coe spent 15 years in the Neuro-science Department (CNS) pursuing medicines for depression, schizophrenia, ADHD and addiction, where he was directly involved in the discovery of Chantix (Champix), an aid to smoking cessation. Coe earned his A.B. degree in chemistry from Harvard College in 1981, worked 2 years at SISA Pharmaceutical Laboratories, then entered MIT to study with Professor William R. Roush. He received his Ph.D. degree in 1988 and joined Pfizer Central Research. Coe has published extensively in the area of organic and medicinal chemistry with a focus on reaction mechanism and synthesis and has received numerous patents in the areas of oncology, neuroscience and inflammation.

Guangbin Dong received his B.S. degree from Peking University and completed his Ph.D. degree in chemistry from Stanford University with Professor Barry M. Trost, where he was a Larry Yung Stanford Graduate fellow. In 2009, He began to research with Prof. Robert H. Grubbs at California Institute of Technology, as a Camille and Henry Dreyfus Environmental Chemistry Fellow. In 2011, he joined the department of chemistry and biochemistry at the University of Texas at Austin as an assistant professor and a CPRIT Scholar. His research interests lie at the development of powerful chemical tools for addressing questions of biological importance.