Solvent Anions Enable Photoinduced Borylation and Phosphonation of Aryl Halides via EDA Complexes

Abstract

We report the synthesis of aryl boronic esters and aryl phosphonate esters promoted by visible-light in the absence of transition-metals or photoredox catalysts. The transformation proceeds at room temperature using sodium hydride, as a non-nucleophilic base, and exhibits functional group tolerance for anilines, amides, and esters. UV–vis spectroscopy, radical trapping experiments, and computational (TD-DFT) calculations suggest an electron-donor–acceptor (EDA) complex between solvent anions and aryl halides as the species responsible for this reactivity.

Graphical Abstract

Aryl boronic esters are valuable moieties in organic synthesis and drug manufacturing as they are essential reagents for the Suzuki–Miyaura cross-coupling reaction.¹ Similarly, aryl phosphonate esters are found in a wide variety of compounds including pharmaceuticals, agrochemicals, and in ligands for metal catalysis.² Aryl halides are the most commonly used precursors to access these moieties given their commercial availability and stability. Traditional methods use highly reactive Grignard and organolithium reagents to activate the aryl halide into coupling with trialkyl boranes³ or trialkyl phosphites.⁴ Transition-metal-catalyzed methods have been successful at increasing functional group compatibility with the help of metals such as Pd, Co, Zn, Ni, Cu, and Fe (Scheme 1A).^5,6 However, these methods often require high temperatures and expensive ligands. Additionally, the cost associated with metal residue removal, in selective industrial applications, further reduces the cost-efficiency of some of these reactions.⁷ While metal-free borylations have been reported,⁸ they often require the use of uncommon silylborane reagents.

Scheme 1. Current and Proposed Approaches to Generate Arylboronates and Arylphosphonates^a.

^a(A) Transition-metal approaches. (B) Photoredox approaches. (C) Precedent for solvent anion EDA complex. (D) Proposed method1

Recently, photoredox pathways have been reported to generate aryl boronates⁹ and aryl phosphonates¹⁰ from aryl halides using rare metal-based photocatalysts or organic dyes (Scheme 1B). Catalyst-free photoinduced borylations¹¹ and phosphonations¹² have been gaining attention over the past decades due to their low environmental footprint and mild reaction conditions, but these methods often require the use of dangerous UV radiation, which limits scalability and broad adoption.

Our interest in catalyst-free photoinduced transformations,¹³ recently led us to discover the ability of aryl halides to form electron-donor–acceptor (EDA) complexes with the dimsyl anion, which enabled the synthesis of chalcogenides (Scheme 1C).^13a The involvement of the N,N-dimethylformamide (DMF) anion as a possible electron donor using heat was previously proposed by Taillefer and Yan,^14a,b although K⁺ cation and ⁻O^tBu were also proposed to be involved in the single electron transfer (SET). Inspired by these discoveries, herein, we report the first catalyst-free visible-light-induced borylation and phosphonation of aryl iodides that is proposed to involve an EDA complex between aryl halides and solvent anions generated in situ (Scheme 1D), further eroding mechanistic proposals that involve tert-butoxides in the SET step.^13a,15

We selected 4-iodoanisole (1a), bis(pinacolato)diboron (2a) as model reagents to optimize the reactions (Table 1). Based on our previous work,^13a we showed that KO^tBu is not involved in the single electron activation of aryl halides. Therefore, we initiated our screens using a strong non-nucleophilic base like sodium hydride (NaH) that is better able to form the desired solvent anions responsible for the activation of aryl halides. While DMSO and N,N-dimethylacetamide (DMA) did not afford the desired product in significant yield (Table 1, entries 1 and 2), DMF afforded the borylated product in 12% yield (entry 3). To improve solubility, the use of a cosolvent system was investigated (entry 4) further improving the yield to 20%. Replacing the base with DIPEA, DBU, or NaO^tBu improved the yield (entries 5, 6, and 7) affording the borylated product in yields ranging from 34–75%. Reducing the amount of NaH from 3 to 1.5 equiv and reducing DMF from 0.1 mL to 1.5 equiv (entry 8) improved the yield significantly (86%). However, reducing these ratios further to 1.1 equiv (entry 9) lowered yields (78%). In the absence of DMF, the reaction still provides the desired product (entry 10) albeit in lower yield (40%), indicating that both acetonitrile and DMF promote the reaction and are possibly involved in the mechanism. Control experiments show that in the absence of base using 390 nm light (entry 11) some background reaction gave 25% of product. However, using a 427 nm light in the absence of base (entry 12) does not afford the desired product, showing the importance of the base in the formation of the borylated product. Comparing light sources between 390 nm (entry 8), 427 nm (entry 13), and 440 nm (entry 14) shows that increasing the wavelength decreases the yield of the reaction from 86% to 79% and 72%, respectively. Finally, performing the reaction in the dark at 40 °C (entry 15) decreased the yield to 25%, showing that this reaction requires photons to be efficient (additional screens see SI, pages S6–S7).

Table 1.

Optimization of the Reaction and Its Conditions^a

entry	base (equiv)	additives	solvent (mL)	yield (%)b
1	NaH (3.0)	–	DMSO (1.0)	trace
2	NaH (3.0)	–	DMA (1.0)	2
3	NaH (3.0)	–	DMF (1.0)	12
4	NaH (3.0)	DMF (0.1 mL)	CH3CN (0.9)	20
5	DIPEA (3.0)	DMF (0.1 mL)	CH3CN (0.9)	34
6	DBU (3.0)	DMF (0.1 mL)	CH3CN (0.9)	75
7	NaOtBu (3.0)	DMF (0.1 mL)	CH3CN (0.9)	59
8	NaH (1.5)	DMF (1.5 equiv)	CH3CN (1.0)	86
9	NaH (1.1)	DMF (1.1 equiv)	CH3CN (1.0)	78
10	NaH (1.5)	–	CH3CN (1.0)	40
11	–	DMF (2.0 equiv)	CH3CN (1.0)	25
12 c	–	DMF (2.0 equiv)	CH3CN (1.0)	Trace
13 c	NaH (1.5)	DMF (1.5 equiv)	CH3CN (1.0)	79
14 d	NaH (1.5)	DMF (1.5 equiv)	CH3CN (1.0)	72
15 e	NaH (2.0)	DMF (2.0 equiv)	CH3CN (1.0)	25

^aReaction conditions: 1a (0.2 mmol, 1 equiv), 2a (3 equiv), base, solvent, room temperature around reaction flask was 35 °C (heating caused by the 390 nm LED lamp), under argon, 24h.

^bYields are based on 1a, determined by ¹H NMR using dibromomethane as internal standard.

^c427 nm.

^d440 nm.

^eThe reaction was performed in dark covered by aluminum foil at 40 °C.

Following reaction optimization, we continued our study investigating the substrate scope for this transformation. We began exploring the scope of aryl iodides compatible with B₂pin₂ under the optimal reaction conditions (Scheme 2). Electron-donating aryl iodides (OMe, Me, ^tBu) afforded products 3–5, in good to excellent yields (67–81%). It is worth noting that some isolated yields are lower than ¹H NMR due to decomposition during purification via column chromatography. Highly polar and electron-deficient products are particularly prone to deborylation during purification. 4-Iodoaniline afforded borylated product 6 in lower but respectable yield (53%), showing that this methodology is compatible with amine functional groups. Electron-neutral iodobenzene also reacted efficiently to give product 7 in good yield (90%). Electron-withdrawing iodoarenes (OCF₃, F, CF₃) gave the desired products 8–10 in moderate NMR yields (54–57%). The method was well tolerated in the presence of tertiary amide (11, 63%). The reaction also tolerated phenol and TMS-protected phenol affording products 12 and 13 in good yields (56% and 73%). Steric hindrance in meta position was also tolerated affording products 14–17 in moderate to good yields (51–86%) in the presence of electron-rich groups, anilines, and secondary amides. Ortho substituents were also tolerated giving borylated products 18 and 19 in 34% and 46% NMR yields, respectively. 1-Iodonaphthalene afforded product 20 in good yield (72%). Heteroaromatic iodides such as quinoline and thiophene gave products 21 and 22 in low to moderate NMR yields (39–58%). Bioactive 1,3-benzodioxole and 2,3-dihydrobenzofuran afforded product 23 and 24 in good yields (81–82%). Finally, natural product containing substrate 25, which contains an ester moiety, was obtained in 59% yield.

Scheme 2. Aryl Iodide Substrate Scope^†.

†Reaction conditions: 1 (0.2 mmol, 1 equiv), 2a (0.6 mmol, 3 equiv), NaH (1.5 equiv), DMF (1.5 equiv), CH₃CN (1 mL), room temperature around reaction flask was 35 °C (heating caused by the 390 nm LED lamp), under Ar, 24 h. ^aIsolated yields. ^b1H NMR yields. ^cIsolated yield for 1 mmol scale reaction

We continued our study using triethyl phosphite as a phosphonation reagent (Scheme 3). Optimal conditions for the phosphonation reaction were found to be slightly different than for borylation. The use of DMF as sole solvent in the presence of 2 equiv of NaH under 427 nm light irradiation afforded the best transformation. Aryl iodides with electron-donating functional groups (OMe, Me, ^tBu) afforded phosphonated products 26–28 in good yields (69–75%). 4-Bromoanisole afforded product 26 in low yield (22%), but 64% of the starting material remained unreacted. Unfortunately, aryl chlorides only afforded products in very low yield. Tertiary amide was tolerated and gave product 29 in moderate yield (52–58%) using both aryl iodide and bromide. Similarly to the borylation reaction, electron-withdrawing groups (CF₃, F) generated the desired products 30 and 31 in moderate yields (31–56%). Finally, several heterocyclic iodides were also tested (products 32–34) but gave the phosphonated products in low to moderate yields (11–44%). These results make our method complementary to to current catalyst-free phosphinylations from the Che group, which only proceeds with electron-poor heteroaromatic substrates.^12b

Scheme 3. Scope of the Phosphonation Reaction^‡.

‡Reaction conditions: 1 (0.2 mmol, 1 equiv), 2b (3 equiv), NaH (2 equiv), DMF (1 mL), room temperature around reaction flask was 35 °C (heating caused by the 427 nm LED lamp), under Ar, 24 h. Isolated yields unless otherwise noted. ^aYields are based on 1, determined by ¹H NMR using dibromomethane as internal standard. ^bYields of unreacted starting materials, determined by ¹H NMR using dibromomethane as internal standard. ^c48 h reaction.

To investigate the mechanism of these two transformations, we performed several control experiments using different radical trapping agents (Scheme 4). Addition of 2,2,6,6-tetramethylpiperidinyloxy (TEMPO), 1,1-diphenylethylene (1,1-DPE), and 1,4-dinitrobenzene (1,4-DNB) to both the borylation reaction (Scheme 4A) and the phosphonation reaction (Scheme 4B) reduced product formation, which suggests that the reaction probably goes through a radical pathway. It is worth noting that for the borylation reaction (Scheme 4A) four different radical intermediates were trapped and observed via GC-MS (see Supporting Information for GC-MS, S27–29). The aryl radical intermediate was trapped as a 1,1-DPE adduct (35). Additionally, DMF (36) and CH₃CN radicals (37) were also trapped as 1,1-DPE adducts.

Scheme 4. Radical Trapping Experiments^a.

^a(A) Radical trapping experiments under borylation conditions. (B) Radical trapping experiments under phosphonation conditions. (C) DFT calculated structure of the EDA complexes and their representation.

Similarly, the phosphonation reaction in the presence of radical trapping agents saw a dramatic decrease in yield, and the aryl radical intermediate was observed via GC-MS as a 1,1-DPE adduct (35). A possible DMF radical was also trapped as a 1,1-DPE adduct (36) (Scheme 4B). It is worth noting that the fragmentation patterns in GC-MS confirm the structure of the observed DMF adduct 36 as an acyl adduct, which suggests the presence of an acyl radical of DMF (see SI, page S27).

Given our previous finding regarding the formation of an EDA complex between aryl halides and dimsyl anion,^13a we postulate that a similar mechanism may be involved in these transformations, but via a DMF or acetonitrile anion in lieu of the dimsyl anion (Scheme 4C). Time dependent-density functional theory (TD-DFT) calculations suggest that the acetonitrile anion and the DMF anion (carbamoyl anion)¹⁴ could form an EDA complex with aryl iodides where the shortest distance between the moieties is 3.7 and 3.9 Å, respectively. The calculations also show that the least energetic electronic transition between in these two complexes appears at 326 and 341 nm, respectively, and they may be responsible for the observed light absorption (see Supporting Information for all calculations, molecular orbital representations, and corresponding contributions, S21–S24). Finally, UV–vis experiments that replicate reaction conditions were also performed (see Supporting Information, S19) and show an increase in absorption when NaH, DMF, and aryl iodides are combined together after 2 h of light irradiation, while in the absence of base of aryl iodide, there is no absorption in the desired range.

Based on these results and literature precedents,^{10a,11,13a,14a,16} we propose that this reaction may be proceeding via an EDA complex between solvent anions of DMF and acetonitrile with aryl halides (Scheme 5). After the light absorption, the EDA complex could deactivate from the excited state through a single electron transfer (SET) yielding the aryl radical anion A and the solvent radical (DMF or acetonitrile radicals). Species A can then generate aryl radical B through the loss of iodide. In the case of the phosphonation, intermediate B reacts with P(OEt)₃ to form phosphoranyl radical C, which are known to fragment across C–O bonds^16b,c and generate ethyl radical D and the desired final product. Alkyl radical D can activate another molecule of aryl iodide in a chain propagation step (Scheme 5). Similarly, the borylation mechanism is proposed to proceed through the trapping of aryl radical B by DMF–B₂pin₂ species E proposed by Studer,^16a and affords the desire product and forms borate radical anion F. Borated species F may also be able to react with a molecule of aryl iodide and carry a chain SET as proposed by Studer (Scheme 5).^16a

Scheme 5. Proposed Mechanism.

In summary, we have developed a visible-light photoinduced cross-coupling reaction between aryl halides and bis(pinacolato)diboron or triethyl phosphite to synthesize aryl bonoric esters and aryl phosphonate esters at room temperature with good functional group compatibility. These reactions are complementary to current catalyst-free borylation and phosphonations since they enable functionalization of electron-rich, electron-poor, and heteroaromatic aryl halides. Based on control experiments, the mechanism of this work involves a radical pathway, and it is presumed to involve solvent anions generated in situ. Further studies to determine mechanistic details, to fully elucidate the role of these anions in solution, and to expand the applications of this synthetic strategy are ongoing in our laboratory. Despite these current unknown, we believe this work opens significant synthetic opportunities in the study of anion-promoted reactions of aryl halides and other reagents, which have been exclusively explored using Lewis basic KO^tBu and may have limited the discovery of novel transformations.