Download PDF
Synlett 2023; 34(18): 2175-2180
DOI: 10.1055/s-0041-1738448
cluster
Modern Boron Chemistry: 60 Years of the Matteson Reaction

Photoredox-Catalyzed Radical–Radical Coupling of Potassium Trifluoroborates with Acyl Azoliums

Michael J. Rourke‡
,
Matthew J. McGill‡
,
Daniel Yang
,
Emelia J. Farnam
,
Joshua L. Zhu
,
Karl A. Scheidt  ∗

We gratefully acknowledge support from the National Institute of General Medical Sciences (NIH) for support of this work (R35 GM136440). D.Y. and E.J.F. thank Northwestern for Undergraduate Research Grants. D.Y. thanks the Chemistry of Life Processes Institute at Northwestern for support in the form of the Lambert Fellowship.
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

Potassium trifluoroborates have gained significant utility as coupling partners in organic synthesis, particularly in the Suzuki–Miyaura coupling reaction. Recently, they have also been used as radical precursors under oxidative conditions to generate carbon-centered radicals. These versatile reagents have found new applications in photoredox catalysis, including radical substitution, conjugate-addition reactions, and transition-metal dual catalysis. In addition, this photomediated redox-neutral process has enabled radical–radical coupling with persistent radicals in the absence of a metal, and this process remains to be fully explored. In this study, we report the radical–radical coupling of potassium benzylic trifluoroborate salts with isolated acyl azolium triflates, which are persistent-radical precursors. The reaction is catalyzed by an organic photocatalyst and forms isolable tertiary alcohol species. These products can be transformed into a range of substituted ketone products by simple treatment with a mild base.


#

Key words

potassium trifluoroborate - photoredox catalysis - organophotocatalysis - acyl azolium compounds - radical–radical coupling - ketones

Organoboron compounds have long been recognized for providing chemists with unique strategies for carbon–carbon bond formation.[1] Recently, these methodologies have undergone significant growth, resulting in the discovery of new two-electron nucleophilic displacement, boron enolate, and transition-metal coupling reactions.[2] The Suzuki–Miyaura coupling reaction has contributed significantly to the explosion of boronic acid and ester diversity, providing a rapid and efficient method for assembling diverse chemical libraries.[3] Potassium trifluoroborate salts, first reported by Chambers and co-workers in 1960, have become widely adopted as bench-stable reagents.[4] Vedejs accelerated this adoption by demonstrating the facile transformation of boronic acids into trifluoroborate salts upon treatment with inexpensive KHF2.[5] More recently, the renaissance of free-radical methodology development has capitalized on the unique properties of organoboron compounds, offering new directions in the field of boron chemistry.

Boronic acid derivatives are widely used as radical precursors in various chemical reactions, including conjugate-addition reactions, transition-metal-mediated coupling reactions, and radical–radical coupling reactions (Figure [1]). The Baran group was among the first to report the oxidation of arylboronic acids to generate aryl radicals, which then participated in Minisci reactions with electron-deficient heteroarenes (Figure [1]A).[6] Potassium trifluoroborates were reported by Fensterbank and co-workers in 2010 as radical precursors; they were then employed by the groups of Baran and Molander to access aryl, alkyl, and alkoxy radicals.[7] Recent advances in photoredox catalysis have provided new routes to radicals, including boron-containing precursors.[8] Akita and colleagues reported the photocatalytic oxidation of a variety of borates to generate the corresponding carbon-centered radicals (Figure [1]B).[9] These organoboron-derived radicals then reacted with TEMPO and electron-deficient alkenes to form C–O and C–C bonds. Later, this work was extended to the hydroalkoxymethylation of alkenes.[10]

Zoom Image
Figure 1 Reaction plan overview: (A) boronic acids and potassium trifluoroborate as radical precursors in heteroarylation; (B) Application in photoredox catalysis; (C) Transition-metal/photoredox dual catalysis for ketone synthesis; (D) Radical–radical coupling of potassium trifluoroborates and isolated acyl azoliums.

Photoredox catalysis offers kinetic control over radical generation and provides a unique opportunity to create new C–C bond-forming methodologies.[11] This controlled generation of radicals also enables dual catalytic systems, as demonstrated by the work of the Sanford group and others through photoredox–transition-metal dual catalysis.[12] This approach tackles a long-standing problem by eliminating the rate-limiting step of transmetalation, which hinders the ability to achieve C(sp3)–C(sp2)/C(sp) couplings by Suzuki–Miyaura reactions.[13] The reaction pathway proceeds through oxidative addition to a variety of C–X electrophiles, followed by radical capture through single-electron transmetalation of the boron-derived carbon radical and reductive elimination to yield the desired C–C coupled product. An application related to this work is the nickel/photoredox dual catalytic system applied to acyl chlorides and potassium organotrifluoroborate salts to produce α-alkoxy ketones and dialkyl ketones (Figure [1]C).[14]

Carbon-centered radicals are generally highly reactive and short-lived.[15] Nevertheless, conjugate acceptors and metal catalysts offer numerous opportunities for radical capture and subsequent transformations. On the other hand, direct radical–radical coupling of carbon-centered radicals presents a challenge because these organic radicals tend to undergo self-coupling and off-target degradation pathways.[16] However, under appropriate conditions radical-radical coupling is favored when (a) radicals are generated at equivalent rates, such as in photoredox-mediated reactions, and (b) the radicals exhibit different lifetimes, with one radical being persistent and the other transient. This is because the persistent-radical coupling partner increases in concentration over time, and subsequent generation of the transient radical leads to expedient cross-coupling. This phenomenon is known as the persistent-radical effect.[17]

The recent development of persistent ketyl radicals has facilitated the development of a variety of novel radical–radical coupling methodologies for accessing tertiary alcohols and ketones.[18] Among these methods, a persistent azolium-derived ketyl radical intermediate has demonstrated remarkable utility in radical–radical coupling reactions. In 2019, Ohmiya and co-workers accessed this intermediate through the direct reduction of phthalimide-derived redox-active esters by a strongly reducing Breslow intermediate.[19] The resulting radical pair undergoes C–C bond formation through radical–radical coupling. To complement this approach, our group has developed methodologies to capitalize on the direct reduction of acyl azoliums and the corresponding persistent acyl azolium radical intermediates. Our reductive approach employed photoredox and N-heterocyclic carbene (NHC) dual catalysis to couple N-acyl imidazoles with Hantzsch esters to form ketones.[20] Shortly after, we reported the applications of isolated acyl azoliums and expanded our methods to include both stoichiometric ester azoliums and three-component radical-relay reactions, which employed potassium trifluoroborates as radical precursors in the photoredox-mediated alkoxycarbonylation of trifluoroborates and synthesis of γ-aryloxy ketones.[21] The reactivity of the azolium-derived ketyl radical has also been developed by other groups, leading to a variety of novel transformations including radical relay, α-heteroatom functionalization, and triplet ketone methodologies.[22] Here, we report the use of a stoichiometric acyl azolium to facilitate radical–radical coupling with potassium trifluoroborate-derived alkyl radicals in a photoredox-mediated process to form ketone products.

Our initial investigations into this coupling utilized the hydrocinnamyl acyl azolium salt 1a and potassium benzyl(trifluoro)borate (2a) to afford the benzylated ketone product 3a (Table [1]). We initially investigated the reaction with a 1:1.1 stoichiometry of 1a and 2a, respectively, to permit monitoring of the reaction by Liquid Chromotrography-Mass Spectrometry (LC-MS) analysis of the disappearance of 1a. We observed that the borates were undetectable or underwent degradation under this analysis. The reaction was then monitored by LC-MS for the appearance of the tetrahedral intermediate 4a, the structure of which was later confirmed by single-crystal X-ray crystallography.[23] This intermediate underwent rapid transformation to a ketone product upon treatment with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU). The selection of 1,2,3,5-tetrakis(carbazol-9-yl)-4,6-dicyanobenzene (4CzIPN) as the photocatalyst and MeCN as the reaction solvent were informed by our previous studies (Table [1], entry 9). Monitoring the reaction by LC-MS revealed a byproduct 5 whose structure was assigned as the transesterification/secondary alcohol based on MS analysis and NMR spectroscopic characterization of a related compound. Given that acyl azoliums are competent acylating agents in the presence of alcohols, this was not surprising, but clearly shutting this pathway down was deemed critical. We attempted to attenuate the nucleophilicity of the key ketyl radical oxyanion with Brønsted or Lewis acid additives. Increases in yield were observed in the presence of trifluoroacetic acid (TFA) and Zn(OAc)2; however, the addition of Zn(OAc)2 caused the formation of an undesirable gel over the course of the reaction. We proceeded with the optimization using TFA (entry 1) to maintain reaction homogeneity, isolating 47% of ketone 3a by using these conditions. To optimize further, we considered that the process might be impacted by the acidity or concentration of the acid additive. TFA outperformed triflic and acetic acids (entries 5 and 6), whereas four equivalents of TFA (entry 4) provided a comparable yield of 50%. Furthermore, modulation of the equivalency was explored: two equivalents of 1a (entry 2) provided similar yields to our initial stoichiometry, yet two equivalents of 2a (entry 3) proved deleterious to the yield. The reaction was determined to proceed sufficiently well with a slight excess of borate in the presence of a stoichiometric Brønsted acid additive and 4CzIPN photocatalyst at 0.2 M in MeCN. With optimized conditions in hand, we next sought to explore the scope of trifluoroborate and azolium coupling partners.[24]

Table 1 Optimization of the Reaction Conditionsa

Entry

1a (equiv)

2a (equiv)

Brønsted/Lewis Acid

Yield (%)

1

1

1.1

TFA

47

2

2

1

TFA

48

3

1

2

TFA

27

4

1

1.1

TFA (4 equiv)

50

5

1

1.1

TfOH

6b

6

1

1.1

HOAc

26

7

1

1.1

TFA + Zn(OAc)2

39

8

1

1.1

Zn(OAc)2

48b

9

1

1.1

–

20b

a Optimization of reaction conditions. 1 equiv of the Brønsted and/or Lewis Acid additive was employed unless noted otherwise. The reactions were performed at 0.2 M.

b 1H NMR yield of product vs 1,3,5-trimethoxybenzene in CDCl3.

Three different azoliums 1 were coupled with a variety of substituted benzyl potassium trifluoroborates 2 to access hydrocinnamyl (3a–f),[25] cyclohexyl (3g–j),[26] or benzoyl products (3k–p) (Scheme [1]).[27] The structure of product 3c was confirmed by single-crystal X-ray diffraction[23] [see the Supplementary Information (SI)]. These studies indicated that alkyl- and aryl-substituted benzylic trifluoroborates (3b–d, 3h, 3l–n) performed comparably to the standard unsubstituted products (3a, 3g, 3k). However, the electron-rich p-methoxy-substituted salts performed poorly in the reaction (3e, 3i, 3o), with a dimerized p-methoxybenzyl product observed by GC/MS indicating a possible electronic mismatch of the azolium and p-methoxybenzyl radicals.[28] The π-withdrawing p-CO2Me-substituted trifluoroborate provided the ketone products in reduced yields (3f, 3j, 3p), and significant azolium starting material remained after 24 hours, suggesting the absence of redox reactivity; this was attributed to a reduced rate of oxidation, as the inductive effect of the electron-deficient arene increases the oxidation potential. Fortunately, transitioning to secondary potassium trifluoroborates provided improved yields of the products 3q–s, obtained in yields of 55, 53, and 66%, respectively. Furthermore, ester-bearing secondary borates offering a handle for further diversification provided 3t–v in moderate yields of 43, 52, and 62%, respectively. Finally, potassium naphthylic trifluoroborates were utilized to explore the effects of the resonance-stabilized radical in the synthesis of 3w and 3x, and resulted in a notable reduction in yield in comparison to 3a and 3k. To broaden the scope of the reaction, the nonstabilized cyclohexyl, cyclopentyl, and tert-butyl aliphatic potassium trifluoroborates were subjected to the standard reaction conditions; however, the reactions failed to yield the desired intermediates. These examples, in addition to the poor-yielding p-methoxy and p-CO2Me substrates, demonstrate the challenges in selecting coupling partners for radical–radical couplings, which are a less-explored class of reactions in comparison to radical addition reactions with proposed trends in radical reactivity.[29]

Zoom Image
Scheme 1 Scope of potassium trifluoroborate salts. Reaction conditions: azolium 2 (1 equiv, 0.5 mmol), potassium trifluoroborate 1 (1.1 equiv), TFA (1 equiv), 4CzIPN (2 mol%), MeCN (0.2 M), LED irradiation (λ = 456 nm), 24 or 48 h; then DBU (1 equiv).

Following the investigation of the potassium trifluoroborate substrates, we sought to explore a variety of acyl azolium coupling partners 1 (Scheme [2]). Initially, a variety of arene substitution patterns were employed in the reaction, and alkyl (4a), electron-deficient (4b–f), and electron-rich (4g) substituents were well tolerated. It is notable that the para-methoxy-substituted azolium performed well in the reaction, achieving 73% yield of 4g, albeit requiring 48 hours to reach completion. The extended π-system of a naphthalene-substituted azolium resulted in a reduced yield of 4h in comparison to 3k. The alkyl substituents of the secondary aliphatic azolium 1 were expanded to include cyclopentyl (4i) and 1-methyl-2-phenylethyl (4j), which gave appreciable yields. Finally, we employed the U.S. Food and Drug Administration-approved drug gemfibrozil to synthesize a tertiary acyl azolium 1n, which was coupled with the secondary potassium trifluoroborate 2h to furnish the ester-bearing analogue 4m.

Zoom Image
Scheme 2 Scope of acyl azolium triflate salts. Reaction conditions: azolium 2 (1 equiv, 0.5 mmol), potassium benzyl(trifluoro)borate (1a; 1.1 equiv), TFA (1 equiv), 4CzIPN (2 mol%), MeCN (0.2 M), LED irradiation (λ = 456 nm), 24 or 48 h; then DBU (1 equiv).

The proposed mechanism of this transformation proceeds through a reductive quenching process in which the photocatalyst 4CzIPN is excited by blue light to form 4CzIPN* (Figure [2]). The activated photocatalyst subsequently oxidizes the potassium benzyl(trifluoro)borate (2a) and liberates a transient benzyl radical (5a). The resulting photocatalyst radical anion, 4CzIPN∙–, performs a single-electron reduction of acyl azolium 2b to access the persistent azolium radical 5b. This reduction might occur by single-electron transfer (SET) followed by fast protonation from TFA, or by a potential proton-coupled electron transfer (PCET) process to yield 5b and TFA–. It is difficult to fully determine whether SET and protonation or PCET is the dominant pathway, as both processes might be operative.[30] The two resulting radical partners then undergo radical–radical coupling to form 5c, a tertiary alcohol intermediate that was isolated and characterized by means of single-crystal X-ray crystallography.[21b] [23] This stable intermediate is then transformed into the desired ketone product by treatment with DBU, liberating the neutral azolium 8. Importantly, when the experiment was run in the absence of a photocatalyst, no intermediate formation or subsequent product was observed. Addition of a Brønsted acid led to a divergence in the reaction profile, primarily the complete disappearance of 6, a major byproduct (based on HRMS analysis; see SI for details).[31] The formation of this byproduct is consistent with a reduction to the ketyl radical anion followed by transesterification and a hydrogen-atom abstraction or an electron-transfer/proton-transfer event. When a radical-trapping experiment was performed in the presence of three equivalents of TEMPO, the formation of the intermediate 5d and the corresponding product 3a was suppressed which suggests a radical mechanism. Additionally, trapping of the benzyl adduct 7 demonstrates the presence of a benzyl radical which was trapped and observed by high-resolution mass spectrometry.

Zoom Image
Figure 2 Proposed mechanism for the 4CzIPN-dependent radical–radical coupling of potassium benzylic trifluoroborate salts and aliphatic acyl azoliums. Proposed structure of ester byproduct 6 observed by LC-MS when the reaction was run in the absence of TFA. The TEMPO-trapped adduct 7 was observed by HRMS when TEMPO (3 equiv) was added to the standard reaction conditions.

In summary, this work describes the use of potassium trifluoroborate salts as radical precursors for use in radical–radical coupling with persistent azolium-derived ketyl radicals. This reaction utilizes a pair of readily available and bench-stable salts derived from organohalides and esters, an easily accessible organophotocatalyst, and visible light. By taking advantage of the persistent-radical character of azolium intermediates in ketone synthesis, we have employed a new coupling partner class with well-established potassium trifluoroborates. This catalytic method provides a metal-free route to a variety of substituted ketone products, including drug derivatives. Overall, the development of this new transformation of potassium trifluoroborate salts broadens their utility in organic chemistry and further substantiates the early fundamental studies by the pioneers of boron chemistry.


#

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

The authors thank Northwestern University for support of this work. We also thank Charlotte Stern (NU) for X-ray crystallographic analysis.

Supporting Information

    Supporting information for this article is available online at https://doi-org.accesdistant.sorbonne-universite.fr/10.1055/s-0041-1738448.
  • Supporting Information

Corresponding Authors

Karl A. Scheidt
Department of Chemistry, Northwestern University
2145 Sheridan Road, Evanston, IL 60208
USA   

Publication History

Received: 04 May 2023

Accepted after revision: 14 June 2023

Article published online:
16 August 2023

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Figure 1 Reaction plan overview: (A) boronic acids and potassium trifluoroborate as radical precursors in heteroarylation; (B) Application in photoredox catalysis; (C) Transition-metal/photoredox dual catalysis for ketone synthesis; (D) Radical–radical coupling of potassium trifluoroborates and isolated acyl azoliums.
Zoom Image
Scheme 1 Scope of potassium trifluoroborate salts. Reaction conditions: azolium 2 (1 equiv, 0.5 mmol), potassium trifluoroborate 1 (1.1 equiv), TFA (1 equiv), 4CzIPN (2 mol%), MeCN (0.2 M), LED irradiation (λ = 456 nm), 24 or 48 h; then DBU (1 equiv).
Zoom Image
Scheme 2 Scope of acyl azolium triflate salts. Reaction conditions: azolium 2 (1 equiv, 0.5 mmol), potassium benzyl(trifluoro)borate (1a; 1.1 equiv), TFA (1 equiv), 4CzIPN (2 mol%), MeCN (0.2 M), LED irradiation (λ = 456 nm), 24 or 48 h; then DBU (1 equiv).
Zoom Image
Figure 2 Proposed mechanism for the 4CzIPN-dependent radical–radical coupling of potassium benzylic trifluoroborate salts and aliphatic acyl azoliums. Proposed structure of ester byproduct 6 observed by LC-MS when the reaction was run in the absence of TFA. The TEMPO-trapped adduct 7 was observed by HRMS when TEMPO (3 equiv) was added to the standard reaction conditions.