Open Access Article
Yi-Dan
Du
ac,
Cong-Ying
Zhou‡
a,
Wai-Pong
To
a,
Hai-Xu
Wang
a and
Chi-Ming
Che
*abc
aState Key Laboratory of Synthetic Chemistry, Department of Chemistry, The University of Hong Kong, China. E-mail: cmche@hku.hk
bHKU Shenzhen Institute of Research & Innovation, Shenzhen, China
cShanghai-Hong Kong Joint Laboratory in Chemical Synthesis, Shanghai Institute of Organic Chemistry, 354 Feng Lin Road, Shanghai, China
First published on 17th April 2020
Visible light driven nitrene transfer and insertion reactions of organic azides are an attractive strategy for the design of C–N bond formation reactions under mild reaction conditions, the challenge being lack of selectivity as a free nitrene reactive intermediate is usually involved. Herein is described an iron(III) porphyrin catalysed sp3 C–H amination and alkene aziridination with selectivity by using organic azides as the nitrogen source under blue LED light (469 nm) irradiation. The photochemical reactions display chemo- and regio-selectivity and are effective for the late-stage functionalization of natural and bioactive compounds with complexity. Mechanistic studies revealed that iron porphyrin plays a dual role as a photosensitizer and as a catalyst giving rise to a reactive iron–nitrene intermediate for subsequent C–N bond formation.
To investigate whether the metal–nitrene intermediate was involved in this light driven C–H amination, several experiments were conducted. [1] The kinetic isotope effect (KIE) was examined by the reaction of ethylbenzene, d10-ethylbenzene and arylazide 2a under the photocatalytic conditions (Scheme 3A). A KIE value of 6.5 was obtained. This KIE value in the range of 5–11 observed in transition metal porphyrin catalysed sp3 C–H amination is supportive of a metal–nitrene intermediate that reacts with the C–H bond via a H-atom abstraction mechanism.10 Bettinger and co-workers found that a free nitrene (R2BN) generated from the photolysis of azidoboranes can undergo intermolecular insertion into a C–H bond of cyclohexane. The KIE value was measured to be 1.35.10d [2] To examine if the reaction involves a radical chain mechanism, the quantum yields (QY) of these photo-chemical reactions have been estimated by using potassium ferrioxalate as a chemical actinometer. It was found that the QY for the reaction of styrene and 2a was 12% while that of indane and 2a was 5.5%. A lower QY of 3.5% was found with Fe(3,5-DitBu-Chenphyrin)Cl11 instead of Fe(TF4DMAP)Cl as the photocatalyst. Since the QY values of these reactions were less than 100%, the photochemical reaction is less likely to involve a radical chain mechanism.12 [3] The cross-over reaction of indane with 2a and 2,3,4,5,6-pentafluoroaniline gave C–H amination products 3a and 3c in 70% and 15% yield, respectively, suggesting a stepwise mechanism (Scheme 3B). [4] When the photochemical C–H amination of indane with arylazide 2a was performed with the chiral Fe(D4-por*)Cl (H2(D4-por*) = meso-tetrakis-{(1R,4S,5S,8R)-1,2,3,4,5,6,7,8-octahydro-1,4:5,8-dimethano-anthracen-9-yl}-porphyrin) catalyst, the C–H aminated product 3a was obtained in 82% yield and with 20% ee (Scheme 4). With Fe(D4-por*)Cl as the catalyst, the reaction of styrene with TsN3 led to aziridine 11d in 49% yield and with 34% ee. This result is comparable to that in our previous work in which the stoichiometric reaction of styrene with Ru(D4-por*)(NTs)2 gave 11d with 27% ee.13 [5] MALDI-MS analysis of the stoichiometric reaction of Fe(TF4DMAP)Cl and azide 2a showed a m/z signal at 1355.1475, corresponding to the iron–nitrene/imido formulation (Fig. S4†). [6] The aziridination of alkenes and 2a under photocatalytic conditions displayed good stereospecificity, giving cis-aziridines from cis-alkenes and trans-aziridines from trans-alkenes in up to 90% yield (Scheme 5), which favours the involvement of a metal–nitrene intermediate rather than a free triplet nitrene. [7] We attempted to detect if any short-lived species would be generated from the iron(III) porphyrin chloride complexes (Fe(TF4DMAP)Cl and Fe(F20TPP)Cl) upon photo-excitation in the presence of azide 2a by nanosecond time-resolved absorption spectroscopy. A transient signal of spectral change at 350–420 nm was observed 1 μs after laser flash and decayed back to the initial base line after 25 μs; such a signal was not observed for photoexcitation of the iron porphyrin complex in the absence of azide 2a under the same conditions. But as the signal was close to the laser excitation wavelength of 355 nm and relatively weak, it would be difficult to make any conclusive statement. These results altogether suggest that in the iron porphyrin catalysed light driven C–H amination, a reactive metal–nitrene/imido species was generated in the course of photolysis.
![]() | ||
| Scheme 4 Asymmetric C–H amination and aziridination catalysed by chiral iron porphyrin. a 5 mol% catalyst was used. | ||
With the optimal conditions, the scope of organic azides for the iron porphyrin-catalysed light driven sp3 C–H amination was examined by using indane as the substrate. As shown in Table 1, electron-withdrawing substituents (CF3, F and Cl) of aryl azides facilitated the intermolecular C–H amination leading to a high yield of products (93–96%, 3b–3d). For aryl azides having only one electron-withdrawing substituent, C–H amination products were obtained in moderate yields (27–60%, 3e–3h). The electron-donating substituent OMe was found to disfavour the reaction resulting in poor product yield (<5%, 3i). It is worth noting that azidoformate Troc-N3 was also reactive in the photocatalytic C–H amination giving 3j in 54% yield. The use of 3,5-bis(trifluoromethyl)benzoyl azide under photocatalytic conditions led to poor yield of the C–H amination product 3k. No reaction was observed when tosyl azide, diphenyl phosphorylazide or benzylazide was used as the nitrene source.
| a Reactions were run under argon with 0.5 mmol of the azide, 5.0 mmol of indane, 5 × 10−3 mmol of the catalyst (1 mol% with respect to the azide) and 120 mg 4 Å MS in 2.0 mL of anhydrous DCE in a 10.0 mL sealed tube. The tube was irradiated with a blue LED at 25–35 °C. |
|---|
|
With azide 2a as the nitrene source, the scope of hydrocarbons was examined. As shown in Table 2, azide 2a underwent nitrene insertion into the benzylic C–H bonds of a range of substrates including ethylbenzene, tetralin, diphenylmethane, fluorene, xanthene, and isochroman to give the amination products in good to high yields (5a–5l). For substituted ethylbenzenes, both electron-donating and -withdrawing groups (4-methoxy, 4-methyl, and 4-chloride) led to high product yields (5g–5j). When isopentylbenzene which bears both benzylic and tertiary C–H bonds was used as the substrate, the reaction took place regio-selectively at the benzylic C–H bond over the tertiary C–H bond (5k).
| a Reactions were run under argon with 0.5 mmol of the azide, 5.0 mmol of the substrates, 5 × 10−3 mmol of the catalyst (1 mol% with respect to the azide) and 120 mg 4 Å MS in 2.0 mL of anhydrous DCE in a 10.0 mL sealed tube. The tube was irradiated with a blue LED at 25–35 °C. b 5 mol% catalyst was used. |
|---|
|
|
In the case of methyl (S)-2-(4-isobutylphenyl)propanoate which has two electronically different benzylic C–H bonds, the C–H amination took place at the more electron-rich C–H bond (5l).
This photochemical C–H amination is also applicable to allylic C–H bonds. The reaction of cyclohexa-1,4-diene with 2a gave the allylic C–H amination product 5m in 70% yield. Allylbenzene gave a mixture of C–H aminated isomeric products (5na and 5nb). The reaction of 1-(allyloxy)-4-(prop-1-en-2-yl)benzene proceeded with high chemo- and regio-selectivity to give 5o in 35% yield. When 2,3-dimethylbuta-1,3-diene was subjected to photocatalysis, a 7-membered heterocyclic product 5p was isolated in 78% yield. The reaction also took place at the allylic C–H bond of α,β-unsaturated aldehydes and α,β-unsaturated ketones in good product yields (5q–5s). When cyclic 2-norbornene was used, bridged secondary C–H amination product 5t was obtained in 52% yield with an endo
:
exo ratio of 3
:
1 along with the aziridination product in 12% yield. No allylic C–H amination product was observed.
We next examined the photocatalytic C–H amination of tetrahydrofuran and unactivated C–H bonds, the functionalization of which is a challenge. Tetrahydrofuran was reactive under photocatalytic conditions giving the amination product 6a in 61% yield (Table 2). The reaction of cyclooctane with 2a proceeded smoothly to give 6b in 40% yield. When the substrate contains tertiary, secondary and primary C–H bonds, the photocatalytic amination preferentially occurred at the tertiary C–H bond over the secondary and primary C–H bonds (6c–6g). We also investigated the regioselectivity of the photochemical C–H amination of substrates bearing multiple tertiary C–H centers. When dihydrocitronellyl acetate or dihydrocitronellyl bromide was subjected to the photochemical C–H amination, the reaction occurred preferentially at the tertiary C–H bond remote from the electron-withdrawing group (acetate or bromo, 6h or 6i). This regioselectivity could be attributed to the deactivation of the C(3)–H bond by the electron-withdrawing group.
To our surprise, the α-C–H bond of ketone, which is generally thought to be deactivated by a carbonyl moiety, could also be aminated. Treatment of 1-phenylbutan-2-one with 2a and a catalytic amount of iron porphyrin at room temperature under blue LED irradiation afforded α-C–H amination product 6l in 40% yield. Likewise, 2,4-dimethylpentan-3-one and 5-methylhexan-2-one gave α-C–H aminated products 6m and 6n in 50% and 36% yields, respectively. No reaction was found when methyl 2-phenylacetate was used. These results suggest that 2a might react with the enol form of ketone to afford an aziridine which underwent ring-opening to give the final α-C–H amination product.14
| Entry | Substrate | Product | Yield (%) |
|---|---|---|---|
| a Reactions were run under argon with 0.3 mmol of the azide, Boc2O (0.36 mmol), 3 × 10−3 mmol of the catalyst (1 mol% with respect to the azide) and 120 mg 4 Å MS in 2.0 mL of anhydrous DCE in a 10.0 mL sealed tube. The tube was irradiated with a blue LED at 25–35 °C. Isolated yields are shown. b [Fe(F20TPP)Cl] was used as the catalyst. Yields were determined by 1H NMR spectroscopy. | |||
| 1 |
|
|
75 |
| 2 |
|
|
70 |
| 3 |
|
|
78 |
| 4 |
|
|
61; 30b |
| 5 |
|
|
45 |
| 6 |
|
|
98 |
| 7 |
|
|
95 |
| 8 |
|
|
55 |
| Entry | Substrate | Product | Conv. (%) | Yield (%) |
|---|---|---|---|---|
| a Reactions were run under argon with 0.3 mmol of the azide, Boc2O (0.36 mmol), 9 × 10−3 mmol of the catalyst (3 mol% with respect to the azide) and 120 mg 4 Å MS in 2.0 mL of anhydrous DCE in a 10.0 mL sealed tube. The tube was irradiated with a blue LED at 25–35 °C. Isolated yields are shown in the brackets. | ||||
| 1 |
|
|
70 | 46 (42) |
| 2 |
|
|
100 | 90 (80) |
| 3 |
|
|
100 | 92 (85) |
| 4 |
|
|
100 | 82 (77) |
| 5 |
|
|
100 | 84 (77) |
| 6 |
|
|
100 | 90 (82) |
:
1. Fe(F20TPP)Cl also catalysed the light-driven amination of Ac-cholesterol with 52% product yield. Similar to cholesterol, the reaction of Bz-diosgenin gave allylic C–H bond aminated product 16 in 60% isolated yield with a dr ratio of 3
:
1 when 2b was used as the nitrene source (Scheme 6B). Amination of the C–H bond adjacent to the oxygen atom was not observed. When 2a was used as the nitrene source, the corresponding amination product was obtained in 78% yield with a dr ratio of 7
:
1. Treatment of an estrone derivative (Scheme 6C) with 2a and Fe(TF4DMAP)Cl under photocatalytic conditions gave an allylic amination product 18 in 79% yield. Likewise, a double allylic amination of a progesterone derivative (Scheme 6D) was observed with 28% yield. When (−)-ambroxide was subjected to photolysis, the C–H amination occurred at the C–H bond adjacent to the O atom giving imine product 20 in 23% yield with a recovery of 45% starting material (Scheme 6E). Treatment of (+)-nootkatone with 2a under photochemical conditions gave an aziridination product in 30% yield.
Footnotes |
| † Electronic supplementary information (ESI) available: Experimental details and characterization of products. See DOI: 10.1039/d0sc00784f |
| ‡ Present address: College of Chemistry and Materials Science, Jinan University, Guangzhou, China |
| This journal is © The Royal Society of Chemistry 2020 |