Transition-Metal-Free Photoredox Phosphonation of Aryl C–N and C–X Bonds in Aqueous Solvent Mixtures

Abstract

Herein, we present an efficient and mild methodology for the synthesis of aromatic phosphonate esters in good to excellent yields using 10H-phenothiazine, an inexpensive commodity chemical, as a photoredox catalyst. The reaction exhibits wide functional group compatibility enabling the transformation in the presence of ketone, amide, ester, amine, and alcohol moieties. Importantly, the reaction proceeds using a green solvent mixture primarily composed of water, thus lowering the environmental footprint of this transformation compared to current methods. The transformation also proceeds under atmospheric conditions, which further differentiates it from current methods that require inert atmosphere. Mechanistic work using fluorescence quenching experiments and radical trapping approaches support the proposed mechanism.

Graphical Abstract

INTRODUCTION

The aromatic phosphonate ester motif is found in a wide variety of pharmaceuticals, photoelectric materials, agrochemical compounds, and in ligands for metal catalysis.^1–5 For these reasons, the synthesis of C–P bonds has attracted significant interest in the past decade.^6–10 Traditional ways to access aryl phosphonates requires the activation of P–H bonds in reagents such as diethyl phosphite (H(O)P(OEt)₂) using transition-metal catalysis (Scheme 1A). Coupling partners for these reactions have been widely explored and include aryl halides, aryl triflates, and aryl boronic acids.^11–14 However, the expensive and sometimes toxic nature of the catalysts and reagents required for these transformations, alongside substrate scope limitations because of harsh reaction conditions, have encouraged the development of milder methodologies. In the past decade, photoredox-catalyzed methods to generate aryl phosphonates has gained traction as an attractive alternative to traditional methods because of their milder conditions and broader functional group tolerance (Scheme 1B).^15–22 Pioneering work by Toste achieved phosphonation of aryldiazonium salts using H(O)P(OEt)₂ in the presence of a gold(I) catalyst and a Ru photocatalyst.¹⁵ Later work by König achieved these transformations using a Ru photocatalyst and a rhodamine dye.^16,17 More recently, the combination of photoredox catalysis and electrochemistry enabled Wickens to achieve the phosphonation of aryl chlorides.¹⁸

Scheme 1.

Current and Proposed Approaches to Aryl C–P Bond Formation: (A) Traditional Metal-Catalyzed Phosphorylations, (B) Photoredox-Catalyzed Approaches, and (C) These Studies

Despite all the progress achieved by these methods, their environmental footprint is far from desirable. The uses of rare metals, problematic solvents, highly toxic organic dyes,²³ or anhydrous and inert atmosphere remain challenges to the green synthesis of these valuable and industrially relevant compounds. Importantly, improving the sustainability of these transformations has to be achieved without compromising substrate scope and functional group compatibility.

Our interest in the development of metal-free green synthetic methodologies,^24–26 recently led us to explore new approaches for the radical activation of aryl halides.²⁵ Here, we report the first metal-free photoredox-catalyzed phosphonation of aryl trimethylammonium salts and aryl halides (iodides and bromides) using affordable phosphite reagents in green aqueous solvent mixtures. This reaction uses 10H-phenothiazine (PTZ), a common and inexpensive commodity chemical, as a photoredox catalyst. In addition to lowering the environmental footprint provided by the possible use of water as a solvent, this transformation proceeds unaffected under atmospheric conditions without the need for degassing or distilling of the solvents, or placing the reaction under inert atmosphere. These mild reaction conditions also provide broad functional group tolerance in the presence of reactive ketone, amide, ester, amine, and alcohol moieties.

RESULTS AND DISCUSSION

Our initial studies aimed at the phosphonation of 4-iodoanisole (1a) using triethyl phosphite (2a) as our phosphorus containing reagent (Table 1). To achieve our objectives of performing this transformation under the most environmentally friendly and cost-effective conditions possible, we focused our attention to the screening of bases, solvents, and atmospheric conditions using PTZ as our photocatalyst (< $90/kg).²⁷

Table 1.

Optimization of the Reaction Conditions^a

entry	base (equiv)	2a (equiv)	solvent	3 yield (%)b	1a recovered (%)b
1	DBU (3.0)	3.0	CH3CN	89 (A)	-
2	Et3N (3.0)	3.0	CH3CN	10 (A)	90
3	DIPEA (3.0)	3.0	CH3CN	12 (A)	87
5	K2CO3 (3.0)	3.0	CH3CN	34 (B)	60
6	LiOtBu (3.0)	3.0	CH3CN	58 (B)	39
7	-	3.0	CH3CN	9 (B)	90
8c	DBU (3.0)	3.0	CH3CN	6 (B)	83
9d	DBU (3.0)	3.0	CH3CN	- (B)	97
10e	DBU (3.0)	3.0	CH3CN	85 (B)	trace
11	DBU (2.0)	3.0	CH3CN	92 (B)	trace
12	DBU (3.0)	2.0	CH3CN	66 (B)	23
13e	DBU (2.0)	3.0	CH3CN/H2O (1:3)	90 (B)	4
14e	DBU (2.0)	3.0	CH3CN/H2O (1:9)	87 (B)	6
15e	DBU (2.0)	3.0	H2O	72 (B)	15

^aReaction conditions: 1a (0.2 mmol), 2a (0.6 mmol), DBU (0.6 mmol), solvent (1 mL), room temperature around the reaction flask is 35 °C (heating caused by the 456 (A) or 427 nm (B) LED lamp), under argon, 24 h.

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

^cThe reaction was performed in the absence of PTZ.

^dThe reaction was performed in the dark, covered by aluminum foil.

^eThe reaction was performed in air.

Gratifyingly, the reaction proceeded in excellent yield (89%) in acetonitrile using DBU (1,8-diazabicyclo(5.4.0)undec-7-ene) as a base (entry 1). Four other organic and inorganic bases were screened providing moderate yields for the desired product and poor conversions of the starting material 1a (Table 1, entries 2–6). The main byproduct of these reactions is the proto-dehalogenation of the aryl halide 1a. In the absence of base (entry 7) or photocatalyst (entry 8), the yields decreased dramatically (<10%), with an almost complete recovery of the starting material. Similarly, no desired product was obtained when the reaction was performed in the dark (entry 9). Importantly, the transformation did not seem to be negatively affected by the use of atmospheric conditions (entry 10). While reducing the amount of DBU to 2.0 equiv improved the yield (entry 11), reducing the amount of phosphite 2a had a deleterious effect on yield (entry 12). The reaction was also minimally affected by the addition of water (entries 13–14). Still, the use of water as the sole solvent (entry 15) afforded the desired product in good yield with minimal decomposition of 1a (72% product yield and 15% recovered starting material). However, the used of other solvents and bases resulted in lower yields (see Supporting Information for full table of optimization, S6).

With a wide range of possible reaction conditions available, we chose to explore the substrate scope using the aqueous mixture of CH₃CN/H₂O (1:3) as it provides a good overview on how efficient the transformation can be, and allows for future possible substrate optimization in acetonitrile or water as sole solvents. A variety of aryl iodides and bromides (1), as well as different phosphite reagents (2) were investigated (Scheme 2). Various aryl iodides and bromides were first reacted with triethyl phosphite (2a) as the coupling partner. The reaction proceeds smoothly in the presence of electrondonating groups (3–5, 91–93%) and electron-withdrawing groups (6–9, 50–88%). A large-scale (5 mmol, 1.17 g) reaction was also performed under optimal reaction conditions and afforded product 3 in 82% yield (see SI, S5 for large scale procedure). Acidic and protic functional groups such as phenol and alcohol (10, 12) were well tolerated (63–75%). Ketone (13) and amide (14) moieties afforded the product in moderate yield (40–53%). Of interest to synthetic chemist, silyl protected alcohols (15, 24) afforded the desired phosphonated products in good yields (52–62%). It is important to highlight that all the yields can be further enhanced by improving the solubility of the starting substrates through the use of acetonitrile as main solvent. Sterics at both ortho (16–19) and meta (20–25) positions were well tolerated (31–87%) even in the presence of aniline group; the only exception was substrate 23. 2-Iodonaphthalene afforded the desired product is excellent yield (26, 92%), and the method tolerated heteroaromatic pyridine (27) and quinoline (28) moieties often found in bioactive compounds (46–57%). The reaction of 1,4-diiodobenzene and 1-bromo-2-iodobenzene afforded the disubstituted products 29 and 30 in moderate to good yield (62–73%). Moreover, various medicinal and natural product frameworks containing esters were investigated (31–33) and generated the desired phosphonated products in moderate to good yield (43–73%), further showing functional group compatibility and the possibility for late-stage functionalization of complex molecules. It is important to highlight that the reactivity of the aryl halide substrates increases as the halide becomes a better leaving group (ArI > ArBr > ArCl ≫ ArF), with aryl chlorides providing only trace amounts of products and aryl fluorides remaining completely unreactive (see SI for additional substrates and experiments; S23). Finally, different trialkyl and triaryl phosphite reagents were tested (34–37) affording the desired products in low to excellent yields (21–92%).

Scheme 2.

Aryl Halides Substrate Scope and Functional Group Compatibility^a

^aReaction conditions: aryl halides 1 (0.2 mmol, 1 equiv), trialkyl phosphites 2 (0.6 mmol, 3 equiv), DBU (0.4 mmol, 2 equiv), PTZ (0.02 mmol, 10 mol %), CH₃CN/H₂O (0.25/0.75 mL), room temperature around reaction flask is 35 °C (heating caused by the 427 nm LED lamp), air, 24 h. All yields are isolated. [a] CH₃CN/H₂O (0.9/0.1 mL). [b] Trialkyl phosphites 2 (1.2 mmol, 6 equiv). [c] Gram scale reaction: 1 (5 mmol), 2 (10 mmol), DBU (10 mmol), PTZ (0.5 mmol), CH₃CN/H₂O (12/4 mL), 48 h.

Subsequently, we decided to apply this procedure to the phosphonation of aryl trimethylammonium salts. To this day, there are no examples of photoredox-catalyzed phosphonations of aryl ammonium salts. Current C–N phosphonation examples are limited to the use of aryl diazonium salts and hydrazine derivatives using transition-metal catalyzed approaches.^15,28–31 To our delight, the reaction successfully proceeds in comparable yields as with aryl halides, although a change in solvent mixture and light source was required. Indeed, quaternary arylammonium salts requires the use of CH₃CN as the major solvent and a 390 nm purple light source (Scheme 3).

Scheme 3.

Aryl Trimethylammonium Salt Substrate Scope^a

^aReaction conditions: aryl ammonium salt 1 (0.2 mmol, 1 equiv), trialkyl phosphites 2 (0.6 mmol, 3 equiv), DBU (0.4 mmol, 2 equiv), PTZ (0.02 mmol, 10 mol %), CH₃CN/H₂O (0.75/0.25 mL), room temperature around flask is 35 °C (heating caused by the 390 nm LED lamp), air, 24 h. All yields are isolated unless noted otherwise. [a] Yields are based on 1, determined by ¹H-NMR using dibromomethane as internal standard. [b] CH₃CN/H₂O (0.25/0.75 mL) with 427 nm LED lamp.

To explore the mechanism for this transformation, we carried out a series of radical trapping and fluorescence quenching experiments (Scheme 4). The addition of four different radical trapping reagents (TEMPO, 1,1-DPE, 1,4-DNB, and BHT) hampered the reactions and aryl radical intermediates were successfully trapped and observed via GC-MS (Scheme 4A). Also, the complete quenching of the reaction by 1,4-DNB strongly suggests the formation of a radical anion intermediate.

Scheme 4.

Mechanistic Experiments^a

^a(A) Radical trapping experiments were performed aryl halides 1 (0.2 mmol, 1 equiv), trialkyl phosphites 2 (0.6 mmol, 3 equiv), DBU (0.4 mmol, 2 equiv), PTZ (0.02 mmol, 10 mol %), and radical trapping agents ((0.4 mmol, 2 equiv), in CH₃CN (1 mL) using 427 nm LED. (B-1) Addition of aryl halide 1a to a mixture of PTZ + DBU. (B-2) Addition of phosphite 2a to a mixture of PTZ + DBU.

Then, fluorescence quenching experiments (Scheme 4B, see Supporting Information for full experiments and Stern–Volmer plots; S21) show that the addition of aryl halide 1a quenches fluorescence of the excited state of the photocatalyst (Scheme 4B-1), while addition of phosphite 2a does not (Scheme 4B-2). This indicates that the excited state of the catalyst initially reacts with the aryl halides 1a via a single electron transfer. It is important to highlight that, under our reaction conditions, we did not observe a significant difference in quenching by 1a of a solution of PTZ alone and a solution of PTZ that contained DBU (see Supporting Information). These results differs from previous observations by the Larionov group,³² which observed enhanced quenching when a mixture of PTZ + Cs₂CO₃ was used. These results suggest that, under our reaction conditions, PTZ alone is capable of reducing aryl iodides and that DBU only marginally increases PTZ’s reactivity. DBU is, however, an essential base in the final steps of the reaction mechanism.

On the basis of the above obtained mechanistic results and previous literature,³² a plausible reaction mechanism is proposed (Scheme 5). First, photoexcited PTZ reduces aryl halide via a single electron transfer to generate aryl radical anion A. Loss of halogen produces aryl radical that then couples with triethyl phosphite and forms phosphoranyl radical B. Facile oxidation of B ($E_{\text{ox}}^0=-1.87\text{ V}$)³³ by PTZ^•+ radical cation ($E_{\text{ox}}^0=0.71\text{ V}$) produces phosphonium species C and regenerates the active photocatalyst. Further reaction of phosphonium intermediate C with DBU affords the desired final product. While other groups have proposed that the base behaves as a terminal oxidant to close the catalytic cycle,^34,35 we do not believe that to be the case in this reaction. Indeed, DBU has an oxidation potential of $E_{\text{ox}}^0=1.21\text{ V}$, which makes the redox reaction with PTZ^•+ ($E_{\text{ox}}^0=0.71\text{ V}$) not spontaneous.¹⁰

Scheme 5.

Proposed Reaction Mechanism

CONCLUSION

In summary, we have developed a simple and efficient photoredox cross-coupling reaction between aryl halides or aryl ammonium salts with trialkyl phosphites to form aromatic phosphonates. This transformation is the first to achieve the photoredox phosphonation of aryl ammonium salts. Additionally, this transformation is the first example of a photoredox-catalyzed phosphonation reaction that can be performed under atmospheric conditions in a wide range of aqueous solvent mixtures and in the absence of transition-metals. The method exhibits broad tolerance of functional groups enabling the late stage phosphonation of complex compounds in the presence of ester, ketone, amide, amine, alcohol, and silyl functional groups.