Dimsyl Anion Enables Visible-Light-Promoted Charge Transfer in Cross-Coupling Reactions of Aryl Halides

Abstract

A methodology is reported for visible-light-promoted synthesis of unsymmetrical chalcogenides enabled by dimsyl anion in the absence of transition-metals or photoredox catalysts. The cross-coupling reaction between aryl halides and diaryl dichalcogenides proceeds with electron-rich, electron-poor, and heteroaromatic moieties. Mechanistic investigations using UV-Vis spectroscopy, time-dependent density functional theory (TD-DFT) calculations, and control reactions suggest that dimsyl anion forms an electron-donor-acceptor (EDA) complex capable of absorbing blue light, leading to a charge transfer responsible for generation of aryl radicals from aryl halides. This previously unreported mechanistic pathway may be applied to other light-induced transformations performed in DMSO in the presence of bases and aryl halides.

Introduction

Diaryl sulfides, selenides, and tellurides are valuable moieties across all chemical sectors from pharmaceuticals to material sciences.^[1] Consequently, the construction of these carbon-chalcogenide bonds has attracted significant interest.^[2] Most methods rely on transition-metal catalysis for the cross-coupling of aryl halides^[3,4] with aryl chalcogenides (Scheme 1A). High temperatures and expensive ligands are often required to avoid thiolate-induced catalyst deactivation and poisoning.^[5] Additionally, the thiols and selenols used in these reactions are unstable, have strong unpleasant odors,^[6] and some derivatives, such as stannyl selenides,^[7] have increased toxicity making them unsuitable for large-scale reactions. Milder reaction conditions using photoredox strategies have been developed (Scheme 1B),^[8,9] but these methods often require the use of expensive photocatalysts. A limited number of catalyst-free photo-induced methodologies for C–S bond formation from aryl halides exist,^[2d,10] but they were not shown to work across group 16 elements.

Scheme 1.

Current and proposed approaches to generate diaryl sulfides and selenides.

Over the past decade, there has been an emergence of transition-metal-free cross-coupling reactions of aryl halides utilizing tert-butoxides (⁻OtBu). Initial observations demonstrated that KOtBu in the presence of ligands^[11] at high temperatures enabled the direct C–H arylation of benzene with aryl halides. In the presence of visible-light, similar transformations of aryl halides have been achieved at room temperature,^[12] often using dimethyl sulfoxide (DMSO) as solvent.

Tuttle and Murphy contributed significantly in elucidating the mechanism of these transition-metal-free transformations,^[13] and it was determined that in most cases KOtBu is not involved in the single electron transfer (SET) step. Instead, they identified various electron-donating species, such as the dimsyl anion, that are capable of performing a SET to aryl halides at high temperatures (135 °C).^[14] However, this work did not provide an explanation for the photo-induced activation of aryl halides in DMSO in the absence of additives. Recently, the Rossi group proposed that the dimsyl anion can be photo-excited to initiate a SET to alkyl halides,^[15] but an electron-donor-acceptor (EDA) complex was not observed.

Our interest in photoinduced SET performed by KOtBu, recently led us to identify a halogen-bonded EDA complex capable of C–H aminations.^[16] Herein, an unprecedented EDA complex between dimsyl anion and aryl halides is reported that is responsible for a photo-induced activation of aryl halides and enables a visible-light-promoted C–S, C–Se, and C–Te cross-coupling with diaryl dichalcogenides (Scheme 1C). The reaction proceeds at room temperature in the absence of catalysts generating the desired products in good to excellent yields across a range of substrates. The involvement of this EDA complex is supported by UV-Vis spectroscopy, time-dependent density functional theory (DFT) calculations, and experimental controls. The observation of this EDA complex provides a compelling mechanistic explanation for the SET that initiates this reaction and possibly other photo-induced transformations of aryl halides that use DMSO in the presence of base.

Results and Discussion

We selected 4-iodoanisole (1 a) and diphenyl disulfide (2 a) as model reagents to optimize the reaction. No product was detected in the absence of base under inert atmosphere with DMSO as the solvent (Table 1, entry 1). In the presence of three equivalents of KOtBu the reaction provided the desired product in 88% isolated yield (entry 2). Using NaOtBu or LiOtBu as base resulted in lower yields (entries 3, 4) indicating the importance of solubility of the base. Other solvents such as acetonitrile (CH₃CN) or dichloromethane (CH₂Cl₂) (entries 5, 6) also gave lower yields (see full details in the Supporting Information, S4). Inorganic and organic bases (Cs₂CO₃, DABCO, and Et₃N) were explored to further understand the potential role of the base in this transformation (Table 1, entries 7–9). Unfortunately, the yields for these reactions were low. The poor performance with Cs₂CO₃ (entry 7) further emphasizes that a different mechanism is at play compared to previously proposed transformations.^[2d] Reducing the amount of 2 a to 1.5 equivalent (entry 11) did not affect the reaction. Further reducing 2 a to 1.1 equivalent (entry 12) lowered the yield. When the reaction was performed in dark (entry 13) no product was formed, which highlights the fundamental role of photons. Importantly, a decrease in yield was also observed when the reaction was carried out in air (entry 14). Finally, a time study showed that for the reaction was completed in 4 h (entry 15).

Table 1.

Optimization of reaction conditions.^[a]

entry	base (equiv.)	2a (equiv.)	solvent (mL)	yield (%)[b]
1	–	2.0	DMSO (1.0)	–
2[a]	KOtBu (3.0)	2.0	DMSO (1.0)	85 (88)[c]
3	NaOtBu (3.0)	2.0	DMSO (1.0)	76
4	LiOtBu (3.0)	2.0	DMSO (1.0)	63
5	KOtBu (3.0)	2.0	CH3CN (1.0)	12
6	KOtBu (3.0)	2.0	CH2Cl2 (1.0)	–
7	Cs2CO3 (3.0)	2.0	DMSO (1.0)	35
8	DABCO (3.0)	2.0	DMSO (1.0)	38
9	Et3N (3.0)	2.0	DMSO (1.0)	trace
10	KOtBu (2.0)	2.0	DMSO (1.0)	25
11[a]	KOtBu (3.0)	1.5	DMSO (1.0)	83
12	KOtBu (3.0)	1.1	DMSO (1.0)	59
13[d]	KOtBu (3.0)	1.5	DMSO (1.0)	–
14[e]	KOtBu (3.0)	1.5	DMSO (1.0)	53
15[f]	KOtBu (3.0)	1.5	DMSO (1.0)	82

^[a]Conditions: 1 a (0.2 mmol), 2 a, base, solvent, room temperature around reaction flask was 35 °C (heating caused by the LED lamp), under N₂, 24 h.

^[b]Yields are based on 1 a, determined by ¹H-NMR using dibromomethane as internal standard.

^[c]Isolated yield.

^[d]The reaction was performed in dark covered by aluminium foil.

^[e]The reaction was performed in the air.

^[f]4 h reaction.

We continued our study investigating the substrate scope for this transformation. The scope of aryl halides is quite broad (Scheme 2, products 1–14). Both electron-donating (OMe, Me) and electron-withdrawing (CF₃, F) groups at the para positions afforded the desired cross-coupled products 3–6 in good to excellent yields (77–89%). Highly electron-poor substrates such as 1-bromo-3,5-ditrifluoromethylbenzene also gave desired product 7 albeit in slightly lower yields (66%). The presence of an aryl nitrile afforded product 8 (33%), primarily due to hydrolysis of the cyano group. Acidic or easily hydrolysable functional groups were not well tolerated. Excellent yields were obtained when 1-iodonaphthalene and 2-bromopyridine were used as coupling partners affording products 9 and 10 (92% and 93%). This promising result led us to explore other heteroaromatic halides, particularly those with privileged frameworks in bioactive compounds. 3-Iodopyridine and 6-iodoquinoline provided desired products 11 and 12 in good yield (75%). 5-Bromoquinoline afforded product 13 in 53% yield. Finally, 2,6-dibromopyridine was efficiently coupled at both halogen sites giving product 14 in a single step (72%).

Scheme 2.

Aryl halide and aryl disulfide scope. Conditions: 1 (0.2 mmol), 2 (1.5 equiv.), KOtBu (3 equiv.), DMSO (1 mL), room temperature around reaction flask was 35 °C (heating caused by the LED lamp), under N₂, 24 h. ^[a] 2 (3 equiv.).

Various aryl disulfides were also investigated (Scheme 2, products 15–28). Product scope was extended to heteroaromatic moieties present on the disulfide, which complements those in the aryl halides. For example, 2,2′-dipyridyldisulfide reacted readily in good to excellent yields with both electron-rich aryl iodides (15, 16) and electron-poor aryl halides (17–19, 73–80%) even tolerating highly electron deficient aromatic rings containing nitro (NO₂) functionalities. Having heteroaromatic moieties on both coupling partners does not negatively affect the yield affording products 21 and 22 in 73% and 75% yields, respectively. Both electron-rich and electron-poor aromatic rings are well tolerated in the disulfide to give products 23, 26, and 27 (88%, 83%, and 89%). However, the presence of aniline affected product formation 24 (22%). Finally, disubstituted product 28 was generated in one step in excellent yields (91%) from 1,4-diiodobenzene, and 84% from 1-chloro-4-iodobenzene, further indicating that selected chlorinated substrates can be coupled (Scheme 2). Unfortunately, other aryl chlorides provided poor yields, presumably because they do not form EDA complexes with the dimsyl anion (See supporting information).

Other commercially available aryl dichalcogenides (Scheme 3) were investigated. Using diphenyl diselenide, electron-rich (29) and electron-poor (30) aryl iodides were coupled in moderate yields (57% and 52%). Amide functionality on the para position was tolerated to give 31 (42%). Importantly, heteroaromatic halides also cross-coupled in moderate to good yields. Indeed 2-bromopyridine and 3-iodopyridine afforded products 32 and 33 (43% and 51%). Quinolines and bipyridines were well-tolerated affording products 34–37 in moderate to good yields, emphasizing potential application of this method to drug discovery. One-step disubstitution of 2,6-dibromopyridine also provided product 39 in moderate yield (46%). Attempting to cross-couple aryl iodides with complex ester functionalities and natural product moieties also proved to be successful, generating isoborneol (40) and menthol (41) derivatives (60% and 49%). Finally, diphenyl ditelluride was also investigated as a coupling partner and afforded desired products 42–47 (32–79%), further demonstrating that the method presented herein works across group 16 elements. Importantly the synthesis of these diaryl tellurides tolerated heteroatoms, amide functionalities, and worked for both aryl bromides and iodides (Scheme 3).

Scheme 3.

Aryl diselenide and ditelluride scope. Conditions: 1 (0.2 mmol), 2 (1.5 equiv.), KOtBu (3 equiv.), DMSO (1 mL), room temperature around reaction flask was 35 °C (heating caused by the LED lamp), under N₂, 24 h. ^[a] 2 (3 equiv.).

To investigate the mechanism of this transformation we performed UV-Vis spectroscopy experiments on various samples of 4-iodoanisole (1 a) and tert-butoxides and NaH in DMSO (Scheme 4, see full details in the Supporting Information S24–S28). As we increased the equivalency of KOtBu in our solution of 1 a in DMSO (Scheme 4A), we observed the formation of a new peak (λ_max=329 nm; shoulder: λ=360 nm). Importantly, an identical peak was formed when KOtBu was replaced with NaH (λ_max=328 nm; shoulder:λ=360 nm) (Scheme 4B), indicating that the resulting EDA complex formed herein does not involve K⁺ or ⁻OtBu. Instead, we propose that this peak results from the absorption of an EDA complex between the dimsyl anion and the aryl halide 1 a.

Scheme 4.

Selected UV-Vis experiments (Full experimental data can be found in the supporting information S24–S28).

Also, as shown in Scheme 4D, not all aryl halides efficiently engage with the dimsyl anion to form an EDA complex. Indeed, 4-iodoanisol (1 a) and 4-bromoanisol (1 b) generate an intense absorption band when mixed with KOtBu in DMSO, while 4-chloroanisole (1 c) has a significantly smaller absorption band under identical conditions. This observation serves as a possible explanation for the lower reactivity of aryl chloride substrates.

We performed additional experiments to verify that aryl halide radical initiation can take place in the absence of thiolates^[2d] or KOtBu (Scheme 5A, Eq. 1 and Eq. 2). A mixture of 1 a and KOtBu in DMSO under blue-light irradiation generates aryl radicals trapped using 1,1-diphenylethylene (1,1-DPE) to form product 50. Replacing KOtBu by NaH under identical conditions also generated the desired aryl radicals. The use of radical quenchers, such as TEMPO and 1,1-DPE (Scheme 5B), lead to lower yields for the desired product 3. However, using NaH as a base did not significantly affect the C–S cross-coupling reaction and afforded the desired product in 71% yield. These results further suggest that an EDA complex between dimsyl anion and aryl halide is responsible for the observed reactivity.

Scheme 5.

Control experiments and calculated EDA complex.

Finally, to support these observations, we performed time dependent-density functional theory (TD-DFT) calculations involving a π-π interaction between the dimsyl anion and the aryl halide where the shortest distance between the moieties is 3.5 Å (Scheme 5C and D). Results show that the least energetic electronic transition appears at 392 nm and is responsible for the observed visible-light absorption. This transition has a charge-transfer excitation character originating from the dimsyl anion molecular orbital (MO) π_HOMO to the aryl halide MOs σ_LUMO, π_LUMO+1 and π_LUMO+2 with a 14%, 73%, and 8% contribution, respectively (supporting information S29–34). Based on previous reports of KOtBu performing SET,^[11,16] we performed various DFT calculations with analogous complexes involving ⁻OtBu and aryl iodide (see supporting information S31–S34). Unfortunately neither halogen-bonded intermediates nor anion-π complexes modeled in our study had charge-transfer character, further eroding a potential involvement of ⁻OtBu in the SET step.

Based on the calculations and mechanistic experiments presented above, we propose that in presence of base, DMSO generates dimsyl anions capable of forming an EDA complex with the aromatic halide (Scheme 6A). A charge-transfer from the dimsyl anion to the aryl halide forms aryl radical anion and dimsyl radical. Loss of iodide generates aryl radical that then couples with disulfide or thiyl radical to give the desired product. Decomposition of the dimsyl radical via reactions with tert-butanol, disulfide or by other processes lead to the formation of minor impurities.

Scheme 6.

A) Proposed Mechanism. B) Ongoing transformations utilizing same mechanism.

Finally, to further demonstrate the value of the proposed mechanism and open the door to future work and other possible transformations, we demonstrated that similar transition-metal-free reaction conditions that generate dimsyl anion can be used to generate new C–B and C–P bonds from (Scheme 6B). Indeed, starting from 1 a in presence of B₂pin₂ or P(OEt)₃ (Scheme 6B, Eq. 3 and Eq. 4) and NaH in CH₃CN with 3 equivalents of DMSO generated the desired aryl-Bpin and aryl-phosphate in good unoptimized yields (67% and 42%, respectively). These two additional examples suggest that our EDA complex strategy can be applied to other functional group interconversions and that DMSO can be used as a reagent and not as a solvent. These examples also further demonstrate that KOtBu is not required as a reagent and it is therefore not involved in the SET step. While more work and further optimization is required for these two reactions, we believe other valuable transformations that generate aryl radicals from aryl halides via EDA complexes are possible.

Conclusion

In summary, we have developed a visible-light-induced cross-coupling reaction between aryl halides and diaryl disulfide, diselenides, and ditellurides to form unsymmetrical diaryl chalcogenides without using transition-metal catalysts, a photocatalyst, or added ligands. The transformation proceeds under mild reaction conditions, exhibits tolerance for various functional groups, and can be applied to cross-couplings of heteroaromatic halides. We also have postulated a novel mechanism to account for these transformations based on control experiments, a UV-Vis spectroscopy investigation, and TD-DFT calculations. We surmise that an EDA complex between dimsyl anion and the aryl halide is formed during the reaction. Upon absorbance of visible-light, the EDA complex undergoes a charge transfer that leads to loss of halide and formation of an aryl radical. From this observation, we envision future dimsyl anion enabled cross-coupling reactions of aryl halides with different aryl radical trapping agents.

Experimental Section

General Procedure for the Synthesis of Unsymmetrical Diaryl Chalcogenides

A 10 mL microwave vial was charged with aryl halides (0.2 mmol), KOtBu (0.6 mmol), diaryl dichalcogenides (0.3 mmol), 1.0 ml DMSO and capped with 20 mm microwave crimp caps with septa. After using acetone-dry ice cooling bath to freeze the mixture, the vial was evacuated and filled with N₂ three times (freeze-pump-thaw). Remove the reaction mixture from cooling bath and let it warm to room temperature. Then put the vial approximately 4 cm away from the Blue LED lamp (427 nm or 456 nm) and stirred at room temperature for 24 h. After the reaction was completed, the reaction mixture was quenched with DI water (1 mL) and then extracted with EtOAc (3 × 10 mL). The organic layer was dried over Na₂SO₄ and filtered through a pad of Celite, and the filtrate was concentrated in vacuo. Finally, the residue was purified using flash chromatography (ethyl acetate/hexane) on silica gel to yield the desired product.