|Home | About | Journals | Submit | Contact Us | Français|
Iridium(I) catalyzes the intramolecular benzylic C–H bond amination of ortho-homobenzyl-substituted aryl azides to produce indolines at 25 °C.
Iridium-catalyzed functionalization of aryl C–H bonds provides regioselective access to valuable aromatic- and heteroaromatic boronic esters.1 Analogous Ir-catalyzed C–H bond functionalizations of aliphatic substrates, however, are rare.2 Since iridium complexes are known to decompose azides,3 we were interested in exploring the N-atom-transfer ability of these compounds as a means to achieve aliphatic C–H bond amination.4,5 Herein, we report an iridium(I)-catalyzed benzylic C–H bond amination reaction that transforms ortho-homobenzyl-substituted aryl azides into indolines at 25 °C (eq 1).
Aryl azide 4 (R = H) was initially selected to investigate the potential for intramolecular Ir(I)-catalyzed C–H bond amination.6 2-Phenylindole and aniline 5 were produced when 4 was exposed to [(cod)Ir(Cl)]2 (5 mol %) at 100 °C (Table 1, entry 1).7 Further screening showed that conversion of 4 to aniline, indole, or indoline was dependent on the identity of the iridium catalyst. No reaction was observed with either [(coe)2Ir(Cl)]2 or [(cod)2Ir]BF4 (entries 2 and 3), but 2 mol % of [(cod)Ir(OMe)]2 catalyzed the transformation at 25 °C to form indoline 7 as the major product (entry 4).8 Further optimization revealed that the addition of an electron-withdrawing substituent to 4 (R = CF3) provided only the indoline product in high yield (entry 5). Control experiments show that iridium is essential for this C–H bond amination process (entries 6 – 8). Metal-free thermolysis exhibited the opposite electronic trend as the iridium-catalyzed process: better indoline conversions were observed with the electron-rich aryl azide substrate (4, R = H, 70%) than the electron-deficient 4 (R = CF3, 15%; entries 6 and 7).9 Low yields (25%) of 7 were also reported by Murata and co-workers when 4 (R = H) was irradiated with a high pressure mercury lamp (entry 8).9c In contrast to our earlier studies,10 exposure of 4 to rhodium(II) complexes resulted in no reaction (entry 9). Rhodium(I) complexes, e.g. [(cod)Rh(OMe)]2, were also incompetent catalysts (entry 10).6 These results emphasize the importance of the metal ligand combination present in [(cod)Ir(OMe)]2 to trigger the decomposition of ortho-homobenzyl-substituted aryl azides at room temperature.
Investigation of the conversion of aryl azide 4 into indoline 7 provided the optimal conditions to examine the scope of the intramolecular Ir-catalyzed C–H bond amination. As suggested by the effect of the fluoride substituent in entry 5 of Table 1, the transformation was sensitive to the electronic nature of the aryl azide substituents: the yield of indoline increased as stronger electron-withdrawing R1- and R2-groups were added to aryl azide 8 (Table 2, entries 1 – 7). In contrast, the electronic identity of the homobenzylic aryl group did not influence the yield of reaction (entries 8 – 11). Presently, our Ir(I)-catalyzed reaction is limited to the amination of secondary benzylic C–H bonds as substrates with R3-alkyl groups or tertiary C–H bonds did not react (entries 12 and 13).11,12
A variety of vinyl- and aryl azides were examined to determine if [(cod)Ir(OMe)]2 could catalyze the amination of aryl- or vinyl C–H bonds (Scheme 1). While all azidoacrylates tested (cf. 10a) were found to be unreactive towards [(cod)Ir(OMe)]2, aryl azides were cleanly converted to indoles or carbazoles (11b – 11i) by [(cod)Ir(OMe)]2 in comparable yields to Rh2(O2CC3F7)4. Enhanced regioselectivity was observed in the reaction of 10i with [(cod)Ir(OMe)]2 as compared to Rh2(O2CC3F7)4. These substrates, however, required higher reaction temperatures (40 °C) and increased catalyst loading (5 mol %) of iridium than aryl azides 8. The reactivity of aryl azides 10 was also not dependent on the electronic nature of their substituents. These differences suggest that a different mechanism (or rate-determining step) might be operating for aromatic N-heterocycle formation than for indoline formation.
Several intermolecular competition experiments were performed to determine if C–H bond activation accounted for indoline formation (Scheme 2).13 In this mechanism, activation of the benzylic C–H bond2 by the iridium catalyst produces η3-benzyl 12.14 Subsequent nucleophilic attack by the pendant azide then forms the C–N bond in the indoline. If this mechanism was occurring, the rate of benzylic C–H bond activation should be retarded when the electron-withdrawing R2-substituents are present.15 Analogously, the rate of nucleophilic addition of azide should be attenuated with electron-deficient R1-substituents. In contrast to these expectations, aryl azides 8b and 8i exhibited nearly equal reactivity towards [(cod)Ir(OMe)]2 and the more electron-deficient 8g reacted faster than 8b. These results suggest that a benzylic C–H bond activation/nucleophilic addition mechanism does not account for N-heterocycle formation.
The faster rate of indoline formation by the electron-deficient aryl azide 8g as well as the production of aniline, a common nitrene decomposition product, suggests that an electrophilic iridium nitrenoid (14) is generated in the mechanism (Scheme 3).3,16 This species could be produced by coordination of the aryl azide with the iridium catalyst (to form α-13 or γ-13)17 followed by extrusion of N2. Carbon–nitrogen bond formation could then occur by two different pathways: a concerted insertion of the nitrenoid via 15 or hydrogen-atom abstraction (to form 16) followed by radical recombination.18,19
To examine the mechanism of C–H bond cleavage, two competition experiments were performed (eq 2). When 8g-d2 and 10h-d5 were exposed to reaction conditions, intramolecular kinetic isotope effects (KIE) of 5.06 and 1.04 were observed to suggest that different mechanisms (or change in the rate-determining step) are occurring for aryl azides 8 and 10. Our measured value for 8g-d2 (5.06) is smaller than the intramolecular KIE of the photochemical reaction of 8g-d1 (14.7)9e and is comparable to the KIE measured for the reaction of a rhodium nitrenoid with cyclohexane (5).18d,20 While the photochemical KIE was interpreted as evidence for a triplet nitrene intermediate, the combination of radical clock studies, Hammett correlation studies with isotope experiments indicate that the electronic state of the rhodium nitrenoid is a singlet.18,20 The magnitude of our KIE (5.06) is similar to those observed for the Rh2(II)-nitrenoid insertion mechanism,18 but further experiments are necessary to rule out alternative mechanisms,21 including one involving triplet 14.22,23
In conclusion, we have demonstrated that iridium(I)-complexes can catalyze the functionalization of benzylic C–H bonds to produce indolines at 25 °C. Future mechanistic experiments are aimed at determining the electronic nature of the reactive intermediates. The resulting mechanistic insight will be exploited in the development of new asymmetric methods to form N-heterocycles from azides by metal-mediated nitrogen atom transfer.
We are grateful to the National Institutes of Health NIGMS (R01GM084945), Petroleum Research Fund administered by the American Chemical Society (46850-G1), and the University of Illinois at Chicago for their generous support. We also thank Profs. L. Anderson (UIC) and M. Newcomb (UIC) for insightful discussions, Mr. B. J. Stokes (UIC) for repeating entry 7, Table 2 and Mr. Furong Sun (UIUC) for mass spectrometry data.