2023 Volume 71 Issue 2 Pages 79-82
Metallaphotoredox-catalyzed allylation represents an emerging synthetic methodology that enables allylic substitution using nucleophilic radical species. The C–H allylation of N-aryl tetrahydroisoquinolines is an innovative example in this area and allows access to synthetically useful precursors for the further derivatization of tetrahydroisoquinolines. However, previous methods have required the use of noble metals, which has hampered their application due to concerns over their sustainability. Here we report the C–H allylation of N-aryl tetrahydroisoquinolines using a cobalt/organophotoredox dual catalyst system. Based on precedent, control experiments and controlled irradiation experiments, a mechanism for the cobalt/photoredox-catalyzed allylation that involves a π-allyl cobalt complex is proposed.
Transition metal-catalyzed allylic substitution has garnered widespread attention both in synthetic chemistry and in organometallic chemistry.1–7) Historical advances in this area have been mainly made using closed-shell nucleophiles, such as malonates and amines, which allow access to a variety of products that possess synthetically useful allyl groups. Recently, rapid progress in metallaphotoredox catalysis8) has led to a new breakthroughs in metal-catalyzed allylic substitution and, allylation protocols using open-shell nucleophiles has been an ongoing focus in this area9–13) (Chart 1A). One such pioneering example of this was reported by Lu, Xiao, and colleagues who developed a C–H allylation of tetrahydroisoquinolines using a palladium catalyst and an iridium-based photoredox catalyst14) (Chart 1B). This protocol accommodates a variety of functional groups and allows access to useful precursors for tetrahydroisoquinoline derivatives without the need to prefunctionalize the amine motif prior to the reaction. However, the requirement to use noble metals reduces the sustainability of the transformation, and thus, a catalytic system that utilizes more earth-abundant elements would facilitate the wider implementation of the reaction in chemical synthesis.
Inspired by the recent development of cobalt-catalyzed allylic substitution reactions,15–23) we anticipated that the C–H allylation of N-aryl tetrahydroisoquinolines might be possible without the use of noble-metals via the use of a cobalt/organophotoredox (OPC) dual catalyst system24–30) (Chart 1C). The design of the dual catalyst system is shown in Chart 2. First, the transfer of an electron from N-aryl tetrahydroisoquinoline to an excited organophotocatalyst affords the reduced photocatalyst (OPC·−). The resulting single electron transfer (SET) between the reduced photocatalyst and Co(II) results in the formation of a Co(I) complex (Chart 2A). The low-valent cobalt complex engages with an allyl carbonate to afford a π-allyl Co(III) complex. Single electron transfer from the reduced organophotocatalyst gives a π-allyl Co(II) complex which couples with the photochemically generated tetrahydroisoquinoline radical (·R2) to afford the allylated tetrahydroisoquinoline product and regenerate the Co(I) complex.
Based on the reaction design described in Chart 2, allylation of N-phenyltetrahydroisoquinoline (2a) using a branched allyl carbonate (1a) was attempted in the presence of a cobalt salt and organic photocatalyst. Pleasingly, the desired product 3aa was obtained in an 83% yield when a combination of CoBr2 and 4CzIPN31) was employed in the presence of MS4A (Table 1, entry 1). The allylation was highly regioselective, and the linear product 3aa was exclusively obtained while the branched product 3aa′ was not observed in the crude reaction mixture. This regioselectivity is consistent with the previous cobalt/photoredox-catalyzed allylation using nucleophilic radical species, which is likely to proceed via the coupling of a π-allyl Co complex and a carbon radical.23,32) Control experiments revealed that the cobalt salt (entry 2), the organophotocatalyst (entry 3) and photoirradiation (entry 4) were all required for the reaction to proceed. This is consistent with the proposed mechanism shown in Chart 2. The reaction can be performed in the absence of MS4A (entry 5), but we continued to employ this additive as it improved reproducibility of the reaction. Other desiccant, CaSO4, was also able to be used and showed a comparable result (entry 6). It is tentatively assumed that the desiccant is responsible for removing trace water occasionally contaminated in the reaction mixture and suppressing the deactivation of Co(I) by water via protonation.33) However, despite exhaustive attempts to purify the reaction mixture using silica gel column chromatography, reverse phase column chromatography and gel permeating column chromatography, 3aa could not be isolated in a pure form. 1H-NMR analysis suggested that 4, which is likely to be formed via the dimerization of a stabilized radical derived from 2a, was formed in the reaction mixture and inseparable from 3aa. Fortunately, the issues with purification could be circumvented when 2b was used instead of 2a,34) and 3ab was obtained in a 77% isolated yield (entry 7).
 | ||
|---|---|---|
| Entry | Change from above conditions | 1H-NMR yield of 3aa |
| 1 | None | 83 |
| 2 | Without CoBr2 | < 5 |
| 3 | Without 4CzIPN | < 5 |
| 4 | In the dark | < 5 |
| 5 | Without MS4A | 82 |
| 6 | CaSO4 (20 mg) instead of MS4A | 79 |
| 7 | 2b instead of 2a and CaSO4 instead of MS4A | 77b) |
a) Reaction conditions: 1a (0.1 mmol), 2a (0.2 mmol), CoBr2 (10 µmol), 4CzIPN (1 µmol) and MS4A (20 mg) in acetonitrile (2 mL) were irradiated with blue LEDs for 24 h. b) Isolated yield of 3ab at a 0.2 mmol scale after 18 h of LED irradiation in the presence of CaSO4 (40 mg).
With the optimized conditions in hand, the substrate scope of the C–H allylation of N-aryl tetrahydroisoquinolines was investigated, and the results are summarized in Chart 3. The reaction has a broad scope with respect to the aromatic allylating reagents as both electron-withdrawing (3bb, 3cb) and electron-donating (3db) groups were tolerated on the aromatic ring. It is of particular note that the bromo group of 3cb was tolerated in this reaction, implying that the oxidative addition of the low-valent cobalt complex to an allyl carbonate outcompetes the potentially competitive oxidative addition to a bromoarene. A 4-tolyl group on the nitrogen atom of the tetrahydroisoquinoline reagent was tolerated, and the respective product was obtained in a 70% yield (3ac). The allylation was also possible when aliphatic allylating reagents were used. While the installation of a prenyl group was successful under the standard conditions (3eb), the installation of an unsubstituted allyl group (3fb)35) or an allyl group with aliphatic substituent (3gb)36) resulted in low yields.37)
To obtain insight into the reaction mechanism, the allylation under controlled irradiation conditions were conducted (Chart 4). While the reaction hardly proceeded in the dark (entry 1), 3aa was obtained in a 42% yield when the reaction mixture was irradiated for 3 h and worked up immediately. When the reaction mixture was irradiated for 3 h and then stirred in the dark for 21 h before worked up, the yield of 3aa was 40%. When the same reaction was performed under irradiation for 24 h, the yield of 3aa was 83%. Comparison of entries 2 and 4 suggests that the reaction requires more than 3 h in order to obtain the C–H allylated product in good yield. Comparison of entries 2 and 3 indicates that, in contrast to cobalt/organophotoredox-catalyzed allylation using malonates,19) contribution of long-lived, thermal chain process is negligible for this reaction. The overall observations in Chart 4 are consistent with the proposed mechanism in Chart 2, in which a photon is required not only in initiation step but also in each product forming step.
In conclusion, we have developed a C–H allylation of N-aryl tetrahydroisoquinolines that utilizes a dual cobalt/organophotoredox catalyst system. While the C–H allylation of N-aryl tetrahydroisoquinolines previously required catalytic systems based on noble metals, the method developed here enables the transformation to proceed without the need for noble metals. It is likely that a π-allyl cobalt complex and a carbon radical generated from the N-aryl tetrahydroisoquinoline via photoredox-induced oxidation are involved in the regioselective C–H allylation. Results of control experiments and controlled irradiation experiments are consistent with the proposed reaction mechanism. The development of other types of allylic substitutions that utilize cobalt/photoredox catalytic systems and radical species is currently underway in our group and the results will be reported in due course.
In an argon-filled glove box, a dried screw-capped vial was successively charged with allyl carbonate 1 (0.20 mmol), N-aryl tetrahydroisoquinoline 2 (0.40 mmol), cobalt bromide (10 mol %), 4CzIPN (1 mol %), CaSO4 (40 mg), and degassed MeCN (4.0 mL, 0.05 M). The vial was capped and removed from the glove box. The reaction mixture was irradiated with a blue LED at room temperature. After stirring for 18 h, the reaction mixture was diluted with water (5 mL), and organic layer was extracted with EtOAc (10 mL×2). The combined organic layer was washed with brine and dried over anhydrous Na2SO4. Filtration and concentration in vacuo furnished the crude product, which was purified by silica gel column chromatography (hexane/toluene) to afford 3. Additional purification of 3 by octadecylsilyl silica gel (ODS) column chromatography (MeCN/H2O) was performed when necessary.
This work was supported in part by JSPS KAKENHI Grant Number JP20H02730 (S.M.), JP22K15239 (M.K.) and JP20K15946 (M.K.). This work was partly supported by Hokkaido University, Global Facility Center (GFC), Pharma Science Open Unit (PSOU), funded by MEXT. T.S. gratefully acknowledges JSPS and Nagai Memorial Research Scholarship from the Pharmaceutical Society of Japan for fellowships.
The authors declare no conflict of interest.
This article contains supplementary materials.
