Selective ortho-C-H Activation in Arenes without Functional Groups.

Aromatic C-H activation in alkylarenes is a key step for the synthesis of functionalized organic molecules from simple hydrocarbon precursors. Known examples of such C-H activations often yield mixtures of products resulting from activation of the least hindered C-H bonds. Here we report highly selective ortho-C-H activation in alkylarenes by simple iridium complexes. We demonstrate that the capacity of the alkyl substituent to override the typical preference of metal-mediated C-H activation for the least hindered aromatic C-H bonds results from transient insertion of iridium into the benzylic C-H bond. This enables fast iridium insertion into the ortho-C-H bond, followed by regeneration of the benzylic C-H bond by reductive elimination. Bulkier alkyl substituents increase the ortho selectivity. The described chemistry represents a conceptually new alternative to existing approaches for aromatic C-H bond activation.

Site-selective activation of aromatic C–H bonds is a challenging step that is important for the synthesis of a range of functionalized aromatic molecules, from pharmaceuticals to polymers.[1] An established way to achieve high regioselectivity is to use arenes with heteroatom-containing functionalities that can direct the reagent attack at the ortho, meta, or even para position.[2] Much more appealing is the activation of nonfunctionalized alkylarenes, which are readily available from petrochemical feedstocks; however, such activation remains challenging because alkyl groups have limited directing capacity, which leads to mixtures of products.[3] The exceptions are symmetrical dialkylarenes, in which functionalization occurs at the least sterically hindered position,[4] and a few monoalkylarenes, in which selective activation of meta-C–H[5] and para-C–H bonds has recently been reported.[6]

Arenes without functional groups are generally modified by electrophilic or transition metal species, yet the yields of ortho-substituted products are less than 67% (Figure ). Electrophilic functionalization usually yields mixtures of ortho- and para-substituted products because the regioselectivity is controlled by electronic factors, resulting in nearly isoenergetic ortho- and para-arenium-cation-like transition states and less stable meta transition states (Figure A). As a result, the ortho/para selectivity does not exceed the statistical ratio of 2:1.[7] In contrast, metal-mediated C–H activation in alkylarenes typically yields mixtures of the meta- and para-substituted products, with the typical ratios of 2:1[4] reflecting the steric accessibility of C–H bonds to the metal center. Few such C–H functionalizations yield more than traces of ortho isomers,[4] and in no case does the ortho regioselectivity exceed 58%.[8]

Figure 1

C–H activation of alkylarenes: (A) electrophilic aromatic substitution; (B) metal-mediated C–H activation; (C) suggested approach for selective iridium-mediated ortho-C–H activation. aFor brevity, only para isomers of key intermediates are shown.

Here we report an approach for regioselective activation of ortho-C–H bonds in alkylarenes using simple iridium complexes (Figure C). The high regioselectivity results from an alkyl substituent acting as an efficient directing group that binds the metal via initial benzylic C–H activation, which triggers subsequent oxidative addition of an ortho-C–H bond, reformation of the benzylic C–H bond, and release of the o-alkylaryl metal species. The key to enabling this approach was the use of rare iridium complexes, Cp*Ir(η4-alkylarene), which bear a nonplanar, “spring-loaded” alkylarene ligand with enhanced reactivity.

We recently demonstrated that Cp*Ir(η4-methylarene) complexes promote selective benzylic C–H activation of the methylarene ligand in the presence of a phosphine ligand.[9]Our attempt to extend this reactivity to primary and secondary alkylarenes led to an unexpected switch in selectivity and formation of ortho-C–H activation products (Figure ). The reaction of isopropylbenzene complex 1a with PMe3 in n-hexane at 100 °C yielded the product of oxidative addition of an ortho-C–H bond, Ir aryl hydride complex 2a, in 99% yield (Figure A). The use of a larger ligand, PPh3, decreased the yield of the ortho-C–H activation product 2a-ph (67%), and the use of no ligand led to a complex mixture of products. Crystal structures of the C–H activation products 2a-ph and 2a as a hydride and bromide species are shown in Figures B and S7.

Figure 2

Scope and selectivity of iridium-mediated oxidative addition of ortho-C–H bonds in alkylarenes: (A) selective ortho-C–H activation in isopropylbenzene via initial η4-arene coordination to Cp*Ir and thermolysis of the resulting complex 1a in the presence of PMe3 or PPh3; (B) crystal structure of 2a-ph; (C) ortho-C–H activation of mono- and dialkylarenes in Cp*Ir complexes; (D) relative order of ortho-C–H selectivity; (E) scope of ortho-C–H activation of alkylarene ligands in complexes 1a–n. Numbers under the arene structures are the total isolated yields of all C–H activation products. Numbers above the arene structures are the ortho selectivities determined by integration of the hydride signals in the 1H NMR spectra. aConditions: 8 equiv. of arene, 1 equiv. of [Cp*IrCl2]2, 4 equiv. of AgBF4, acetone, 24 °C, 16 h, then 2 equiv. of Cp2Co, benzene, 24 °C, 1 h. bSee ref (9).

We explored how the selectivity of ortho-C–H oxidative addition depends on the identity of the alkyl substituent on the arene ring by heating alkylarene iridium complexes 1a–n with PMe3 as an added ligand (Figure C). These complexes are accessible from alkylarenes in 75–97% yield via a simple two-step procedure (see the Supporting Information). C–H activation of alkylarene ligands in 1a–n led to high yields of iridium hydrides 2a–n (87–99%) regardless of the identity of the alkyl substituent (Figure C). The ortho selectivity, however, was the highest with larger alkyl groups (Figure D). As shown in Figure E, arenes with secondary alkyls (i-Pr, s-Bu, 3-Pent, c-Pent, c-Hex) underwent ortho-C–H activation with ≥91% selectivity (2a–e), while arenes with primary alkyls (Et, n-Pr, n-Bu, i-Bu) gave lower ortho selectivities of 72–79% (2f–i). An exception was the bulkiest primary alkylarene in the test, neopentylbenzene, which yielded C–H activation product 2j with 91% ortho selectivity. The arene with the smallest alkyl substituent, methylbenzene, gave no ortho-C–H activation product but instead gave benzyl hydride complex 2k.[9] The observed order of ortho regioselectivity, sec-alkyl > n-alkyl ≫ methyl (Figure D), is opposite to that of classical electrophilic substitution[7a,7b] and contrasts with that of known oxidative additions of C–H bonds in alkylarenes, which favor meta and para but not ortho products.[10] The same counterintuitive trend holds for the C–H activation of para-substituted dialkylarene ligands (Figure C): for example, in p-isopropylmethylbenzene, aromatic C–H activation occurs exclusively next to the isopropyl substituent and not next to the methyl substitutent (2m). Bulkier p-diisopropylbenzene gave ortho metalation product 2l exclusively, while smaller p-dimethylbenzene yielded a 32:68 mixture of ortho and benzyl C–H metalation products 2n.[9]

To rationalize the observed counterintuitive regioselectivity, we probed the reaction mechanism by monitoring the model C–H activation in p-diisopropylbenzene complex 1l in the presence of PMe3 in cyclohexane-d12 (Figure A). Complex 1l was chosen because it exists as a single isomer, which improves the accuracy of kinetic measurements by 1H NMR spectroscopy. The reaction is first-order in 1l and zeroth-order in the phosphine (Figure A). The initial reaction rates for separate thermolyses of 1l and its analogue with a fully deuterated arene ring, 1l-d, were within the experimental uncertainty (Figure B), suggesting that ortho-C–H bonds do not participate in the rate-determining step. Contrary to what was expected, the ortho-C–D activation in 1l-d in n-hexane did not lead to deuterium incorporation in the hydride ligand of 2l-d. Instead, deuterium appeared in the methyl groups of the o-isopropyl group (Figure C). This may result from the intramolecular H/D redistribution between the deuteride ligand and the methyl groups. Intermolecular H/D scrambling between the hydride (deuteride) ligand and the solvent was excluded because heating 1l in cyclohexane-d12 did not lead to incorporation of D into 2l (Figure A). The observed H/D redistribution in 2l-d must have resulted from H/D scrambling in reaction intermediates, not in the starting complex 1l-d, in which no H/D redistribution was observed over the course of the reaction.

Figure 3

Mechanistic experiments using model ortho-C–H activation in 1l: (A) model reaction and rate law measurement upon thermolysis of 1l in cyclohexane-d12 in the presence of PMe3; (B) H/D kinetic isotope effect measured for separate thermolyses of 1l and 1l-d in cyclohexane-d12 at <15% conversion; (C) H/D scrambling upon thermolysis of 1l-d in n-hexane. The values in blue show the D contents at the specified positions. aMeasurements were conducted at 75 °C.

We identified a mechanism that agrees with the experimental observations for the ortho-C–H activation in 1l by computing a range of reaction paths using the M06-2X functional (Figures S8–S11). The lowest-energy path (Figure A, path 1) starts with sliding of the arene ligand to give η2-arene intermediate 3, which then oxidatively adds the benzylic C–H bond of the adjacent isopropyl group. The resulting η3-benzyl hydride complex 4 isomerizes into metallacycle 5 by insertion of iridium into the adjacent ortho-C–H bond. Quick elimination of the benzylic C–H bond in 5 forms coordinatively unsaturated aryl hydride 6, which binds PMe3 to afford the observed product 2l. The similar energies of the four least-stable transition states of the main mechanism (21.2–23.3 kcal/mol) preclude unambiguous identification of the rate-determining step.[11] However, the lack of a primary kinetic isotope effect (KIE) in 1l-d vs 1l suggests that oxidative addition of an aromatic C–H bond in 4 is not the rate-determining step.

Figure 4

Mechanistic insight into C–H activation in 1l. (A) Calculated mechanisms for aromatic and benzylic oxidative addition of ortho-C–H and benzylic C–H bonds in 1l. All of the calculations were done with the M06-2X functional using the def2SVP basis set for geometry optimizations and frequency calculations and the def2TZVPP basis set for single-point energy calculations. All free energies are relative to 1 mol of 1l and 1 mol of PMe3 at 75 °C in cyclohexane (represented in computations by the conductor-like polarizable continuum model). (B) Proposed mechanism for the observed intramolecular H/D scrambling in 1l-d.

This mechanism revealed that the observed selective ortho-C–H activation in iridium η4-arene complexes results from the favorable combination of the kinetic and thermodynamic factors that promote site-selective aromatic C–H activation and disfavor competing benzylic C–H activation. The ortho-C–H activation is kinetically favored because of the specific directing effect of an alkyl group (Figure A). The coordinated alkylarene substrate undergoes the initial benzylic C–H activation of the alkyl group that anchors the metal center next to an ortho position and thus promotes the oxidative addition of the ortho-C–H bond followed by the facile formation of the final product via fleeting iridacyclobutane dihydride intermediate 5. This strained and bulky metallacycle has a high free energy (16.8 kcal/mol above the starting complex 1l) and high reactivity (a barrier of 5.9 kcal/mol for the conversion to 6), which precluded the detection of the intermediate. However, more stable analogues of 5 with less bulky ancillary and hydrocarbyl ligands[12] and their proposed intermediacy in a related isomerization of o-methylaryl to benzyl complexes were reported.[13] As can be seen in Figure A, the ortho-C–H activation via the sequential oxidative addition of two C–H bonds indeed requires traversing much lower barriers (≤23.3 kcal/mol, path 1) than the standard direct ortho-C–H oxidative addition in 3 (31.3 kcal/mol, path 2).

The ortho-C–H activation in 1l (Figure , path 1) is thermodynamically preferable to the competing benzylic C–H activation (Figure , path 3) that occurs via the same intermediate 4 and gives exergonic benzyl complex 7. In contrast, C–H activation in the less bulky iridium methylarene complexes selectively yields benzylic products, which are kinetically and thermodynamically more accessible than the corresponding o-methylaryl products as we reported previously.[14] This comparison of the C–H activation in secondary alkylarene and methylarene iridium complexes suggests that the higher degree of substitution at the benzylic carbon destabilizes the benzyl complex versus the aryl complex and therefore promotes aromatic ortho-C–H activation at the expense of benzylic C–H activation (Figure C,D).

This reactivity contrasts with the established reactivity of metal complexes toward alkylarenes, which favors the activation of meta- and para-C–H bonds over the activation of benzylic and ortho-C–H bonds.[10a,10b,15] The reported double C–H activation mechanism overcomes this limitation: the reversible benzylic C–H activation anchors the metal next to the ortho positions and lessens the barrier for the following ortho-C–H oxidative addition (Figure A).

Finally, the mechanism may explain the remarkable H/D redistribution upon the ortho-C–D oxidative addition in 1l-d (Figure C) that yields hydride, not deuteride, product 2l-d.[16] Complex 2l-d may result from equilibration of the initial deuteride intermediate 6-iso-d with hydride 6-d followed by coordination of PMe3. Although the exact mechanism for this equilibration has yet to be identified, our preliminary calculations suggest that it may occur via five-membered metallacycles 8-iso-d and 8-d and that these metallacycles can be accessed from 1l-d only via 6-iso-d (Figures S8–S15). The equilibrium between 6-iso-d and 6-d lies toward hydride 6-d, which is favored entropically because of the 6:1 H:D ratio and also enthalpically because of the zero-point-energy effect, i.e., the preferred location of deuterium in the highest-frequency oscillator,[17] which is the C(sp3)–D bond, not the metal–D bond. A similar explanation was proposed by Jones and Feher for the 2.7-fold higher stability of the related deuteride complex Cp*Rh(PMe3)(C6D4H)D versus its hydride isomer Cp*Rh(PMe3)(C6D5)H in an equilibrium mixture.[18]

In summary, we have presented a conceptually new method for controlling the site selectivity of C–H activation in arenes without directing groups. This method relies on the use of simple iridium(I) complexes that enable highly selective ortho-C–H activation in primary and secondary alkylarenes without any functional groups. Key to this selectivity is the transient reversible benzylic C–H activation that brings the metal center into close proximity to an ortho-C–H bond and enables smooth metal insertion into the most sterically hindered position of the aromatic ring. This C–H activation occurs in a highly reactive Cp*Ir(η2-alkylarene) intermediate generated by sliding of the arene ligand in a Cp*Ir(η4-alkylarene) precursor. Translation of this stoichiometric reactivity into catalytic ortho-C–H functionalizations may open new avenues for the selective synthesis of value-added chemicals from unactivated aromatic hydrocarbons. Enabling such synthetic applications will require further improvement of the scope and selectivity of the process and the design of a catalytic cycle that involves the straightforward formation of the key unsaturated η2-alkylarene iridium intermediate from the free arene and regeneration of this intermediate after C–H functionalization. Work on addressing these challenges is ongoing in our laboratory.