Download PDF
Synlett 2024; 35(09): 1047-1051
DOI: 10.1055/a-2219-6907
cluster
Chemical Synthesis and Catalysis in Germany

Visible-Light-Mediated Selective Allylic C–H Oxygenation of Cycloalkenes

Tabea Rohlfs‡
a   University of Münster, Organic Chemistry Institute, Corrensstraße 36/40, 48149 Münster, Germany
,
Leon Gerken‡
a   University of Münster, Organic Chemistry Institute, Corrensstraße 36/40, 48149 Münster, Germany
,
Jose L. Nova-Fernández
b   Organic Chemistry Department (Módulo 1), Facultad de Ciencias, Universidad Autónoma de Madrid, 28049 Madrid, Spain
,
Sara Malagón
a   University of Münster, Organic Chemistry Institute, Corrensstraße 36/40, 48149 Münster, Germany
,
Mustafa Uygur
a   University of Münster, Organic Chemistry Institute, Corrensstraße 36/40, 48149 Münster, Germany
,
Silvia Cabrera
c   Inorganic Chemistry Department (Módulo 7), Facultad de Ciencias, Universidad Autónoma de Madrid, 28049 Madrid, Spain
d   Institute for Advanced Research in Chemical Sciences (IAdChem), Universidad Autónoma de Madrid, 28049 Madrid, Spain
,
José Alemán∗
b   Organic Chemistry Department (Módulo 1), Facultad de Ciencias, Universidad Autónoma de Madrid, 28049 Madrid, Spain
d   Institute for Advanced Research in Chemical Sciences (IAdChem), Universidad Autónoma de Madrid, 28049 Madrid, Spain
,
Olga García Mancheño∗
a   University of Münster, Organic Chemistry Institute, Corrensstraße 36/40, 48149 Münster, Germany
› Author Affiliations

This work was funded by the Deutsche Forschungsgemeinschaft (DFG) within the projects GA 1594/6-2, GA 1594/7-1, and GRK 2678–437785492, and the DFG is gratefully acknowledged for this generous support. This work was also supported by the Spanish Ministry of Science and Innovation (project PID2021-122299NB-I00, TED2021-130470B-I00, TED2021-129999B-C32), the Comunidad de Madrid for European Structural Funds (S2018/NMT-4367) and proyectos sinérgicos I+D (Y2020/NMT-6469). We want to thank the ERASMUS+ and the ERASMUS program for enabling T.R. to do a research stay at U.A.M. and S.M. to conduct Master’s thesis work at the University of Münster, respectively. T.R. also thanks the IRTG 2678 for a PhD contract.
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

A visible-light-mediated selective allylic C–H bond oxygenation of cyclic olefins is presented. Hence, the selective, mild monooxygenation of simple cycloalkenes has been achieved using an acridinium photoredox catalyst in combination with a phosphate base and a disulfide HAT reagent under air atmosphere at room temperature. The combination of both photocatalyst and HAT reagent, which can operate through a single or two different concurrent mechanistic pathways for the formation of the allyl radical, proved highly efficient, while the reaction with exclusively one or the other mediator performs in significantly lower yields. The formed allyl radical further reacts with a molecule of oxygen to build the corresponding peroxyradical that can abstract a hydrogen atom of another cycloalkene substrate, generating the known hydroperoxide intermediate in the formation of the ketone moiety. The advantages of this method rely on the easy use of air as oxygen source, as well as the selective monooxygenation of cycloalkenes without substitution in one of the allylic positions. Besides simple cyclic olefins, the method was also successfully applied in the oxidation of natural products such as the terpene valencene or cholesterol derivatives.


#

Key words

photocatalysis - allylic oxidation - C–H oxygenation - acridinium salts - visible light

The selective allylic C–H bond oxidation of olefins represents a fundamental transformation in organic chemistry,[1] leading to allylic alcohols or α,β-unsaturated carbonyls that are key synthetic building blocks and motifs in bioactive and natural products.[2] Most of the available methodologies rely on metal catalysis, use of additives and, generally, harsh conditions such as high temperatures (Scheme [1a]).[3] [4] Moreover, these methods often encounter poor chemo- and regioselectivity. Therefore, efficient, selective strategies for the direct allylic C–H bond oxygenation are still highly demanded. Recently, the development of metal-free approaches has attracted great attention, such as the use of in situ generated nitroxides (Scheme [1b], top).[5] In this regard, visible-light photoredox catalysis has emerged as a powerful tool towards mild, selective oxidation reactions.[6] [7] In 2022, Cai and coworkers reported an efficient rose bengal catalyzed allylic oxidation to enones using n-BuNBr as hydrogen atom transfer (HAT) cocatalyst under oxygen atmosphere (Scheme [1b], bottom).[8] An energy-transfer mechanism within this catalytic system was proposed. Nonetheless, the regioselectivity issue with simple alkenes has not been yet solved, since the second allylic position needed to be blocked as a quaternary center to achieve a good performance.

Continuing our program on photocatalytic C–H oxygenations,[9] we aimed at solving some of the current limitations by developing an allylic oxygenation reaction (Scheme [1c]). To this purpose, we envisioned the use of acridinium dyes,[10] which exhibit highly oxidative excited states upon visible-light irradiation, in combination with a HAT reagent.

For the optimization of the reaction conditions, simple cyclohexene, not suitable for the previous photocatalytic system,[8] was chosen as model substrate and DCE as appropriate solvent based on our experience in the related benzylic oxygenation[9] (Table [1], see the Supporting Information for further screening). The reaction with commercially available Fukuzumi catalyst PC-I (E red* = 2.06 V vs. SCE),[10a] 1 equiv of tributylmethyl ammonium dibutyl phosphate [(n-Bu3MeN)O2P(On-Bu)2] as base, and diphenyldisulfide (I) as HAT reagent under air atmosphere and blue LED irradiation at room temperature (20 °C) led to the formation of the desired cyclohexenone (2) in 6% yield (entry 1). Upon exploring other HAT reagents (entries 2–4), bis(4-nitrophenyl)disulfide (II) was found as the most suitable species, providing 2 in 49% yield (entry 2). Conversely, n-Bu4NBr showed a significantly lower performance (29%, entry 4). Next, the base was screened (entries 5–9; see Table S5 for the complete base screening). While the inorganic base K3PO4 (entry 9) and i-Pr2NEt (entry 8) were not efficient, other ammonium phosphate and benzoate bases provided the product in similar yields (entries 5 and 6; 49% and 47%, respectively).

Zoom Image
Scheme 1 Allylic C–H bond oxygenation approaches

Interestingly, when 2,4,6-collidine was used as base, a good 42% yield on 2 was achieved, but the reaction was less chemoselective (entry 7), detecting the corresponding epoxide as side product. Therefore, we continued our investigation with (n-Bu3MeN)O2P(On-Bu)2 as base, for which the use of 1 equiv proved more efficient (entries 2, 10, 11). Moreover, the use of a base turned out to be important to reach high conversions (entry 12), while the amount of the HAT reagent had less influence on the yield (entries 13–15).

Control experiments (entries 15–22) showed the need of light (entry 18) and air oxygen (entry 2 vs. entry 22) to promote the reaction, while the outcome was highly dependent of the concentration of oxygen in the reaction media. Thus, an excess of oxygen (i.e., 1 atm O2; entry 21) proved detrimental, most probably due to undesired overoxidation side reactions. We could observe that the reaction was sensitive to moisture and the stirring speed employed, the latter influencing the amount of oxygen dissolved in solution during the reaction (see the Supporting Information for more details).[11] [12] Accordingly, the use of pre-dried DCE over 4 Å MS (≤ 6 ppm H2O content) as solvent and 700 rpm stirring speed were found optimal, leading to the product 2 in 51% GC yield (entry 23). No improvement could be observed when decreasing or increasing the catalyst loading (entries 24 and 25) or by increasing the amount of both base and HAT reagent to 2 equiv (entry 26), leading to significant lower yields. We also realized that the determination of the yield of 2 by NMR using CH2Br2 was more suitable, as GC-FID analysis seemed to underestimate the yield of the product (70% vs. 51%, entry 23; see also Scheme [2]). Additionally, small amounts of 2-chlorocyclohexanol (ca. 10%) were detected as a side product by GC–MS and GC-FID,[13] while certain evaporation of the substrate 1 is assumed during the reaction as it was kept open to air.

Table 1 Optimization of the Reaction Conditionsa

Entry

Base (equiv)

HAT reagent (equiv)

Yield 2 (%)b

 1

(n-Bu3MeN)O2P(On-Bu)2 (1)

I (1)

 6

 2

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

49

 3

(n-Bu3MeN)O2P(On-Bu)2 (1)

III (1)

17

 4

(n-Bu3MeN)O2P(On-Bu)2 (1)

n-Bu4NBr (1)

29

 5

(n-Bu4N)O2P(On-Bu)2 (1)

II (1)

49c

 6

n-Bu4NOBz (1)

II (1)

47c

 7

2,4,6-collidine (1)

II (1)

42c,d

 8

i-Pr2NEt (1)

II (1)

26c

 9

K3PO4 (1)

II (1)

 29

10

(n-Bu3MeN)O2P(On-Bu)2 (0.5)

II (1)

 44c

11

(n-Bu3MeN)O2P(On-Bu)2 (2)

II (1)

 39c

12

–

II (1)

 26

13

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (0.5)

 38

14

(n-Bu3MeN)O2P(O n-Bu)2 (1)

II (2)

 41

15

(n-Bu3MeN)O2P(On-Bu)2 (1)

–

 42

16

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

 33e

17

–

–

 25

18

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

  0f

19

(n-Bu3MeN)O2P(On-Bu)2 (1)

–

  0e

20

–

II (1)

 13e

21

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

<10c,g

22

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

<10c,h

23

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

 51i (70)i,j

24

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

 32i,k

25

(n-Bu3MeN)O2P(On-Bu)2 (1)

II (1)

 31i,l

26

(n-Bu3MeN)O2P(On-Bu)2 (2)

II (2)

 12i

a Reaction were performed on a 0.1 mmol scale in dry DCE and open to air.

b GC yields using hexadecane as internal standard.

c GC yields using 2-methylnaphthalene as internal standard.

d Epoxide formation as side product was observed with this base.

e No photocatalyst.

f No light.

g Reaction under O2 atmosphere.

h Reaction under N2 atmosphere.

i Stirring speed of 700 rpm vs. 300–400 rpm.

j NMR yield using CH2Br2 as internal standard in brackets.

k Use of 5 mol% of PC-I.

l Use of 15 mol% of PC-I.

The effect of the stirring speed (700 rpm) in the reaction promoted by only the PC-I (30% NMR yield) or only the HAT-II reagent (31% NMR yield) was then analyzed (Scheme [3]A), leading to significant lower yields and showing the beneficial combination of photoredox and HAT processes for this transformation (up to 70% NMR yield). However, these results also indicate two possible competitive mechanisms for the formation of the key allyl radical II, one involving a photoredox SET oxidation by the acridinium catalyst (PC-I cycle) and the other only a HAT process with the corresponding thiyl radical from HAT-II (HAT cycle, Scheme [3]B).

Zoom Image
Scheme 2 Scope of the reaction. a NMR yield using CH2Br2 as internal standard. b Isolated yield upon column chromatography on silica gel. c Complete decomposition of 5 took place.
Zoom Image
Scheme 3 (A) Reactivity and stability of Int-OH and product 2 vs. 1 as substrate under standard reaction conditions (Table [1], entry 23). NMR yields using CH2Br2 as internal standard. (B) Plausible concurrent mechanisms. (C) Reaction monitoring. NMR yields using CH2Br2 as internal standard.

The formed allyl radical further reacts with a molecule of oxygen to build the corresponding peroxyradical III that can abstract a hydrogen atom via HAT to another cycloalkene substrate to initiate a radical chain, leading to the hydroperoxide IV as known precursor in the formation of the ketone moiety via either direct loss of H2O or hydroxyl radical towards the allylic alcohol Int-OH.

The reactivity and stability of the allylic alcohol intermediate (Int-OH) and final oxygenated product 2 under the optimized reaction conditions (Table [1], entry 23) was next studied (Scheme [3]A).

To our delight, the oxidation of the allylic alcohol proceeded in a good 98% yield, while the amount of cyclohexanone slightly reduced, being recovered in 82% yield.

Next, the reaction course of the oxygenation of 1 was monitored by 1H NMR (Scheme [3]C). While cyclohexene was almost fully converted after 16 h, we could observe a decrease of the yield of the product 2 after 18 h from 70% to 60% after 24 h, which is in concordance with the observed partial decomposition of 2 under the reaction conditions (Scheme [3]A).

With the optimized conditions in hand (use of PC-I, (n-Bu3MeN)O2P(On-Bu)2, and HAT-II for 18 h),[14] the scope of our methodology was tested with various cycloalkenes (Scheme [2]). We observed a strong effect of the ring size on the efficiency of the oxidation reaction. To our delight, cyclopentene also underwent the desired oxidation to cyclopentenone 3 (47% NMR yield). However, the monooxygenation process with cycloheptene was less efficient, resulting in the formation of the desired enone 4 in a lower 37% NMR yield. Furthermore, the formation of the desired product was not observed with larger rings such as cyclooctene 5. For this case, we assume a polymerization of intermediates to occur, as besides the base and HAT reagent only traces of the corresponding epoxide could be detected in the crude mixture.[15] Finally, other substitution at the cyclohexene core was compatible. Thus, a tert-butyl and a methyl ester group on the alkene provided the cyclic enones 6 and 7 with a good 65% and 59% isolated yield, respectively. The substitution at one of the allylic positions with a methoxy and a n-butyl rest exhibited a significant less reactivity, leading to the desired product 8 in 23% yield, while observing unreacted starting material after 18 h reaction. Moreover, this method could also be applied to more complex molecules such as terpenoids like (+)-valencene bearing more than one unsaturation.

In this case, the compound 7 was built as only isomer in a 46% yield. Furthermore, the reaction with cholesterol derivatives also proceeded smoothly, providing the enone products 10 and 11 in 35% and 39% yield, respectively.

In conclusion, we have developed a visible-light-mediated catalytic allylic oxygenation of cyclic alkenes presenting two allylic positions. Using commercially available Fukuzumi acridinium photocatalyst under air in the presence of (n-Bu3MeN)O2P(On-Bu)2 as base and bis(4-nitrophenyl)disulfide, chemoselective, monooxygenation to the corresponding enones was achieved under mild conditions. Cyclohexenes led to the desired products in yields up to 70%, while 5- and 7-membered cyclic substrates were also compatible. The method could also be applied to more complex substrates such as cholesterol derivatives, as well as a terpene presenting two double bonds, observing complete selectivity towards the allylic oxidation of the endocyclic olefin vs. the exo moiety.


#

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

Dr. J.H. Kuhlmann is acknowledged for providing some of the screened photocatalysts.

Supporting Information

    Supporting information for this article is available online at https://doi-org.accesdistant.sorbonne-universite.fr/10.1055/a-2219-6907.
  • Supporting Information

Corresponding Author

José Alemán
Inorganic Chemistry Department (Módulo 7), Facultad de Ciencias, Universidad Autónoma de Madrid
28049 Madrid
Spain   

Olga García Mancheño
University of Münster, Organic Chemistry Institute
Corrensstraße 36/40, 48149 Münster
Germany   

Publication History

Received: 03 October 2023

Accepted after revision: 29 November 2023

Accepted Manuscript online:
29 November 2023

Article published online:
05 January 2024

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Scheme 1 Allylic C–H bond oxygenation approaches
Zoom Image
Scheme 2 Scope of the reaction. a NMR yield using CH2Br2 as internal standard. b Isolated yield upon column chromatography on silica gel. c Complete decomposition of 5 took place.
Zoom Image
Scheme 3 (A) Reactivity and stability of Int-OH and product 2 vs. 1 as substrate under standard reaction conditions (Table [1], entry 23). NMR yields using CH2Br2 as internal standard. (B) Plausible concurrent mechanisms. (C) Reaction monitoring. NMR yields using CH2Br2 as internal standard.