Photocatalytic conversion of aryl diazonium salts to sulfonyl fluorides

Abstract

Sulfonyl fluorides have emerged as powerful tools in chemical biology for the selective labelling of proteins. A photocatalytic method is described for the conversion of aryl diazonium salts to aryl sulfonyl fluorides. The diazonium substrates are easily obtained in one step from functionalized anilines. We present the optimization of this mild method for the synthesis of sulfonyl fluorides, the scope of the transformation with a series of functionalized diazonium salts, and we discuss photophysical measurements that provide detailed information about the mechanism of the photochemical process.

1. Introduction

In 2014, Sharpless and co-workers introduced the concept of Sulfur(VI) Fluoride Exchange (SuFEx) chemistry of sulfonyl fluorides as a second generation click reaction [1]. This transformation allows for the selective and efficient reactivity of a sulfur(VI) center with a nucleophilic amino side chain under physiological conditions, exchanging the fluoride for a nucleophile [2]. The selective reactivity of sulfonyl fluorides is attributed to the abnormally strong sulfur–fluorine bond and sluggish reactivity at the sulfur(VI) center, and it requires activation (H⁺, R₃Si⁺, etc.) to cleave the sulfur-fluorine bond [3]. Since Sharpless’ seminal study, sulfonyl fluorides have emerged as useful chemical warheads for the site-selective modification of proteins in biological systems, with broad applications in drug discovery, biochemistry, and target identification [4].

As the use of sulfonyl fluorides has increased, so has the demand for novel methods for their synthesis (Fig. 1). Aryl sulfonyl fluorides (1) can be accessed through the activation and substitution of sulfonyl chlorides, sulfonamides, and sulfonic acids [5] as well as the transition metal catalyzed activation of aryl halides, triflates, and boronic acids [6]. Recent advances in electrochemistry have enabled the conversion of thiophenols, disulfides, and aryl halides into sulfonyl fluorides [7].

Fig. 1.

Previous methods for the synthesis of aryl sulfonyl fluorides and our strategy for the conversion of aryl diazonium salts to sulfonyl fluorides.

With an on-going interest in developing mild and general methods to synthesize functionalized sulfonyl fluorides from readily available starting materials, we recently reported a photocatalytic conversion of bench-stable alkyl organoboron substrates into sulfonyl fluorides [8]. Based on this chemistry, we were interested in developing a strategy to access sulfonyl fluorides from aryl diazonium salts (2). Herein, we describe the photocatalytic transformation of aryl diazonium salts (2) into sulfonyl fluorides (1) via a visible light-induced electron transfer pathway.

Diazonium salts are versatile substrates that are easily synthesized from anilines in a single step with high conversion [9,10]. When formed with tetrafluoroborate as the counterion, these compounds are easily isolated and recrystallized as stable solids [11]. Diazonium salts have been traditionally used to convert anilines into competent electrophiles for nucleophilic aromatic substitution [12]. They have also become common substrates for transition metal catalyzed cross-couplings [13]. More recently, methods have been developed for the reductive activation of diazonium salts for homolytic cleavage to generate aryl radicals (3) [14]. Hence, these substrates are of great synthetic utility, as they would enable direct late-stage modifications of aniline-containing drug-like molecules [15].

Our approach proceeds through a three-component coupling between the transiently formed aryl radical (3), a sulfur dioxide reagent, and a fluorine reagent. Recently, Tlili and co-workers reported a similar strategy for the generation of aryl sulfonyl fluorides from diazonium salts [16]. In addition, Chen, Weng, and Von Wangelin have developed transition metal mediated or thermal activations of diazonium salts to achieve similar products [17]. We demonstrate a complementary method that relies on a different photoredox catalyst and a distinct fluorine reagent. We describe detailed photophysical studies that shed light on the mechanism of both methodologies. We demonstrate the compatibility of our method with several aryl diazonium salts, with the goal of utilizing this strategy on medicinally relevant anilines that can be conveniently converted to their corresponding diazonium salts. We anticipate this approach will facilitate the generation of libraries of aryl sulfonyl fluorides from chemically diverse anilines in two steps.

2. Results and discussion

We began our studies by converting para-methoxybenzenediazonium tetrafluoroborate 2a to sulfonyl fluoride 1a (Table 1). Under our standard optimized conditions, diazonium salt 2a was subjected to Ru(bpy)₃Cl₂$6H₂O, 467 nm LED irradiation, diazabicyclo[2.2.2]octane bis(sulfur dioxide) adduct (DABSO) as a source of sulfur dioxide [18] and N-fluorobenzenesulfonimide (NFSI) as a fluorinating reagent in acetonitrile. The desired sulfonyl fluoride product 1a was formed in 93% NMR yield, 73% isolated yield (entry 1). When the photocatalyst was replaced with Eosin Y (entry 2) or [Ir(ppy)₂(dtbbpy)]PF₆ (entry 3), the NMR yield of desired product diminished to 85% and 71%, respectively. With a different sulfur dioxide source (aqueous sulfurous acid, entry 4) or fluorine source (Selectfluor^™, entry 5), the yields dropped significantly to 6% and 39%, respectively. In the presence of a different light source at a higher energy wavelength, the yield dropped slightly to 83% (entry 6). In 1,2-dichloroethane, the yield diminished slightly to 80% (entry 7) and when toluene was used, the yield of product decreased significantly to less than 5% (entry 8). Control experiments revealed that both the photocatalyst and light source are required for reactivity, as in the absence of both there was no product formation (entries 9 and 10).

Table 1.

Development of sulfonyl fluorination of aryl diazonium salts.

Entry	Variation frorm Standard Conditions	Yield (%)a
1	None	93 (73)
2	Eosin Y as photocatalyst	85
3	[Ir(ppy)2(dtbbpy)]PF6 as photocatalyst	71
4	SO2H as SO2 Source	6
5	Selectfluor as Fluorine Source	39
6	427 nm light source	83
7	DCE as Solvent	80
8	PhMe as solvent	< 5
9	No photocatalyst	< 5
10	No light source	< 5

Reaction conditions: Diazonium salt 2a (0.23 mmol), photocatalyst (5 mol%), SO₂ source (1.5 equiv), fluorine source (1.5 equiv), Kessil lamp, solvent (0.1 M), 23 °C, 21 h.

^(a)NMR yield based on dimethoxybenzene (DMB) as internal standard. Isolated yield in parenthesis.

With optimal reaction conditions, we synthesized a small library of aryl sulfonyl fluorides to display the robustness of the method (Table 2). The reaction was productive on a wide range of substrates. Good isolated yields of products were observed with electron donating substituents such as methoxy (entry 1), as well as electron withdrawing groups such as nitro (entry 2) and trifluoromethyl (entry 3). Overall, the transformation was tolerant of carbonyl groups, including a ketone (entry 4) and ester (entry 5). We were also pleased to see product formation with sterically hindered diazonium salts with an ortho-methoxy substituent (entry 6), meta-methoxy substituent (entry 7), meta-nitro substituent (entry 8), and multi-substituted aromatic rings (entries 9–10).

Table 2.

Development of sulfonyl fluorination of aryl diazonium salts.

Entry	Substrate	Product	Yield (%)a
1			73
2			48
3			26
4			43
5			38
6			49
7			36
8			45
9			33
10			34

Reaction conditions: Diazonium salt 2 (0.23 mmol), Ru(bpy)₃Cl₂·6H₂O (5 mol%), DABSO (1.5 equiv), NFSI (12 equiv), 467 nm Kessil lamp, MeCN (0.1 M), 23 °C, 21 h.

^(a)Isolated yield.

3. Mechanistic experiments

Tables 1 and 2 inform that the choice of diazonium salt, solvent, fluorine reagent, and source of sulfur dioxide all play a role in the overall yields. To gain mechanistic information, we first turned to excited-state quenching experiments using the Stern-Volmer approach (Fig. 2A–C). The experiments were carried out in argon-purged acetonitrile containing 0.1 M TBABF₄ to maintain a constant ionic strength throughout the experiment and avoid salt effects that could originate when the concentration of diazonium tetrafluoroborate salts is increased [19]. When para-methoxybenzenediazonium tetrafluoroborate 2a was used, intense excited-state quenching was observed (Fig. 2A). Stern-Volmer analysis yielded a linear relationship, and from the slope a quenching rate constant, k_q, of 1.5 × 10⁹ M⁻¹s⁻¹ was determined (Fig. 2C). A much larger quenching rate constant (k_q = 10.2 × 10⁹ M⁻¹s⁻¹) was obtained when para-nitrobenzenediazonium tetrafluoroborate 2b was used (Fig. 2B and C). Low excited-state quenching that quickly reached a plateau was observed when NFSI was used as the quencher (Figure S3). DABSO led to excited-state quenching that also appeared to decrease/plateau when the concentration reached 5 mM (Figure S3). Furthermore, when DABSO was used as quencher, the formation of an unidentified white precipitate occurred over time, which hindered rigorous excited-state quenching experiments. That observation, coupled to the lack of photoproducts by photolysis (vide infra) point towards an artifact in excited-state quenching measurements with DABSO or an excited-state electron transfer coupled to very fast back-electron transfer with extremely small cage-escape yields [20]. Although electron transfer cannot be certified via these steady-state photoluminescence quenching experiments, the increase in quenching rate constant observed between MeO– and NO₂– substituted benzenediazonium salts was consistent with an increased driving force for electron transfer from the excited photosensitizer to the para-nitrobenzenediazonium tetrafluoroborate 2b. Excited-state quenching was also observed with Selectfluor^™ (Figure S3), which represents a competitive quenching pathway that could explain the decreased product yields observed when this reagent was used as the fluorine source (Table 1, entry 5).

Fig. 2.

Steady-state photoluminescence quenching of [Ru(bpy)₃]²⁺ recorded in argon purged CH₃CN containing 0.1 M TBABF₄ electrolyte with increasing amounts of para-methoxybenzenediazonium 2a (A) and para-nitrobenzenediazonium 2b (B), and the corresponding Stern-Volmer plots (C). Panel D shows the ground-state (dashed), excited-state (red) and transient absorption with 30 mM para-methoxybenzenediazonium (blue) spectra recorded in argon-purged acetonitrile. Both excited-state and transient absorption spectra are recorded 50 ns after pulsed 420 nm light excitation. The inset shows the single wavelength absorption changes recorded at 450 nm after pulsed 420 nm light excitation.

The excited-state quenching experiments were also corroborated by steady-state photolysis experiments (Figure S1–S2) and nanosecond transient absorption spectroscopy (Fig. 2D). Photolysis of [Ru(bpy)₃]²⁺ in argon-purged acetonitrile in the presence of 15 mM of para-methoxybenzenediazonium tetrafluoroborate 2a led to the gradual decrease of the metal-to-ligand charge transfer (MLCT) transition at 450 nm, indicative of oxidation of Ru(II) to Ru(III) [21]. Similar results were obtained in 1,2-dichloroethane, although the oxidation proceeded slower than in acetonitrile, presumably due to the lower solubility of para-methoxybenzenediazonium 2a (Figures S1–S2). Photolysis experiments could not be carried out in toluene due to the extremely low solubility of [Ru(bpy)₃]²⁺, explaining the low sulfonyl fluorination yields. When similar experiments were carried out with DABSO, the overall absorbance increased drastically, indicative of scattering (Figure S4). This observation was corroborated by the appearance of an unidentified white solid, as noted during the excited-state quenching experiments. The authentic spectra of [Ru(bpy)₃]²⁺ was recovered when the solution was syringe-filtered, indicating the absence of excited-electron transfer from [Ru(bpy)₃]²⁺ to DABSO (Figure S4).

Nanosecond transient absorption was then carried out to provide unequivocal evidence that oxidative excited-state electron transfer was indeed occurring (Fig. 2D). First, the excited-state absorption spectra of [Ru(bpy)₃]²⁺ was recorded in argon-purged acetonitrile. Spectral data exhibited the prototypical changes associated with MLCT transitions, which include a ground-state bleach around 450 nm corresponding to the depopulation of Ru(II) states to form the corresponding Ru(III) and the 2,2′-bipyridine radical, as evidenced by the positive absorption features below 410 nm and above 690 nm (Fig. 2D, red trace) [22]. When similar experiments were carried out in the presence of 30 mM of para-methoxybenzenediazonium tetrafluoroborate 2a, bleached signals were observed in almost the entire spectral window (Fig. 2D, blue trace). The absence of positive absorption features coupled with the greater than 150 μs lifetime of the transient species (Fig. 2D, inset) are clear indicators of excited-state electron transfer and formation of the corresponding Ru(III) center [20].

Our proposed mechanism is depicted in Fig. 3. Steady-state and time-resolved experiments are in line with light-induced population of a [Ru(bpy)₃]²⁺* excited-state that is able to efficiently transfer an electron to aryl-diazonium derivatives 2, with quenching rate constants in the 10⁹–10¹⁰ M⁻¹s⁻¹ range. This electron transfer generates the corresponding aryl radical 3, N₂, and the corresponding oxidized [Ru(bpy)₃]³⁺. Subsequent coupling of aryl radical 3 with DABSO results in the formation of sulfonyl radical 4, which reacts with NFSI to form sulfonyl fluoride product 1 [18]. Attempts to determine the regeneration pathway of [Ru(bpy)₃]²⁺ were performed through photolysis experiments using DABSO or NFSI. Experiments were frustrated once again by the appearance of a white precipitate when DABSO was used, but the recovery of Ru(II) was somewhat more efficient when DABSO was used instead of NFSI (Figures S5–S6). Therefore, we propose that the reaction of oxidized Ru(III) with diazabicyclo[2.2.2]octane mono(–sulfur dioxide) adduct 5 or DABCO 7 regenerates Ru(II).

Fig. 3.

Proposed mechanism based on photolysis and experimental results.

4. Conclusion

We have reported a mild photocatalytic procedure that allows for the conversion of readily available aryl diazonium tetrafluoroborate substrates 2 into the corresponding sulfonyl fluoride products 1. [Ru(bpy)₃]²⁺ outperformed the two other photosensitizers investigated herein and exhibited good photostability in acetonitrile and 1,2-dichloroethane (Figures S7–S8). Excited-state quenching experiments and nanosecond transient absorption spectroscopy confirmed efficient excited-state electron transfer from the photosensitizer to the aryl diazonium derivatives with quenching rate constants in the 10⁹–10¹⁰ M⁻¹s⁻¹ range. Overall, the method will enable the mild and safe installation of sulfonyl fluorides into functionalized aryl-diazonium tetrafluoroborate substrates.

5. Experimental section

5.1. General procedure

Aryl diazonium tetrafluoroborate salt 2 (0.23 mmol), DABSO (0.345 mmol, 1.5 equiv, 81 mg), NFSI (0.345 mmol, 1.5 equiv, 107 mg), and Ru(bpy)₃Cl₂·6H₂O (0.0115 mmol, 5 mol%, 8 mg) were added to a flame-dried cylindrical septum-sealed reaction vial under argon. The vial was degassed and purged three times with argon, and anhydrous acetonitrile (0.1 M, 2.3 mL) was added under argon. The solution was stirred while irradiated with a 467 nm Kessil lamp (100% intensity, ~3 cm from the vial) for 21 h. The vial was kept at room temperature with a cooling fan. Upon completion of the reaction, the mixture was quenched with a solution of saturated aqueous ammonium chloride (30 mL), and extracted with dichloromethane (3 × 25 mL). The combined organic layers were dried with anhydrous sodium sulfate, filtered, and concentrated in vacuo. The resulting aryl sulfonyl fluoride 1 was purified by silica gel flash column chromatography with hexanes/EtOAc as the eluent.

5.2. Characterization data

4-methoxybenzenesulfonyl fluoride (1a):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (73% yield): ¹H NMR (600 MHz, CDCl₃): 7.94 (d, J = 8.8 Hz, 2H) 7.06 (d, J = 9.11 Hz, 2H), 3.92 (s, 3H). ¹³C NMR (150 MHz, CDCl₃): 165.12, 130.78, 124.04, 114.79, 55.84. ¹⁹F NMR (564.5 MHz, CDCl₃): 67.31. ESI-MS calculated for [C₇H₈FO₃S, M + H]⁺: 191.0178, Found 191.0178.

4-nitrobenzenesulfonyl fluoride (1b):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (48% yield): ¹H NMR (600 MHz, CDCl₃): 8.50 (d, J = 9.0 Hz, 2H), 8.25 (d, J = 9.0 Hz, 2H). ¹³C NMR (150 MHz, CDCl₃): 151.72, 138.37, 129.95, 124.82. ¹⁹F NMR (564.5 MHz, CDCl₃): 66.27 ESI-MS calculated for [C₆H₄FNO₄S, M]: 204.9845, Found 204.9861.

4-(trifluoromethyl)benzenesulfonyl fluoride (1c):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (26% yield): ¹H NMR (600 MHz, CDCl₃): 8.18 (d, J = 8.4 Hz, 2H), 7.93 (d, J = 8.4 Hz, 2H). ¹³C NMR (150 MHz, CDCl₃): 137.45, 136.74, 127.80 (t, J = 3.3 Hz, 1C), 123.77, 121.96. ¹⁹F NMR (564.5 MHz, CDCl₃): 65.96, −63.49 ESI-MS calculated for [C₇H₃F₄O₂S, M – H]⁻: 226.9795, Found 226.9804.

4-acetylbenzenesulfonyl fluoride (1d):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (43% yield): ¹H NMR (600 MHz, CDCl₃): 8.17 (d, J = 8.4 Hz, 2H), 8.12 (d, J = 8.6 Hz, 2H), 2.69 (s, 3H). ¹³C NMR (150 MHz, CDCl₃): 196.29, 142.31, 136.01, 129.40, 129.01, 27.10. ¹⁹F NMR (564.5 MHz, CDCl₃): 65.87. ESI-MS calculated for [C₈H₈FO₃S, M + H]⁺: 203.0178, Found 203.0178.

Ethyl 4-(fluorosulfonyl)benzoate (1e):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (38% yield): ¹H NMR (600 MHz, CDCl₃): 8.22 (d, J = 8.6 Hz, 2H), 8.02 (d, J = 8.7 Hz, 2H), 4.38 (q, J = 7.3 Hz, 2H), 1.36 (t, J = 7.3 Hz, 3H). ¹³C NMR (150 MHz, CDCl₃): 164.58, 136.99, 130.81, 128.62, 127.16, 62.30, 14.35. ¹⁹F NMR (564.5 MHz, CDCl₃): 65.84 ESI-MS calculated for [C₉H₁₀FO₄S, M + H]⁺: 233.0284 Found 233.0302.

2-methoxybenzenesulfonyl fluoride (1f):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (49% yield): ¹H NMR (600 MHz, CDCl₃): 7.93, (d, J = 7.7 Hz, 1H), 7.70 (t, J = 7.8 Hz, 1H), 7.11 (s, 1H) 7.10 (s, 1H), 4.01 (s, 3H). ¹³C NMR (150 MHz, CDCl₃): 158.19, 137.48, 131.36, 121.43, 120.60, 112.84, 56.66. ¹⁹F NMR (376.6 MHz, CDCl₃): 58.57 ESI-MS calculated for [C₇H₇FO₃S, M]: 190.0100, Found 190.0106.

3-methoxybenzenesulfonyl fluoride (1g):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (36% yield): ¹H NMR (400 MHz, CDCl₃): 7.60 (d, J = 8.2 Hz, 1H), 7.53 (t, J = 8.2 Hz, 1H), 7.47 (m, 1H), 7.28 (d, J = 8.5 Hz, 1H), 3.89 (s, 3H). ¹³C NMR (150 MHz, CDCl₃): 160.29, 130.84, 129.81, 122.37, 120.72, 112.76, 56.00. ¹⁹F NMR (376.6 MHz, CDCl₃): 65.63 ESI-MS calculated for [C₇H₈FO₃S, M + H]⁺:191.0178, Found 191.0181.

3-nitrobenzenesulfonyl fluoride (1h):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (45% yield): ¹H NMR (400 MHz, CDCl₃): 8.88 (m, 1H), 8.65 (d, J = 7.7 Hz, 1H),8.37 (d, J = 7.7 Hz, 1H), 7.91 (t, J = 7.7 Hz, 1H) ¹³C NMR (150 MHz, CDCl₃): 148.58, 135.18, 133.91, 131.49, 130.14, 124.05. ¹⁹F NMR (376.6 MHz, CDCl₃):66.43 ESI-MS calculated for [C₆H₄FNO₄S, M]:204.9845, Found 204.9822.

3,5-dichlorobenzenesulfonyl fluoride (1i):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (33% yield): ¹H NMR (600 MHz, CDCl₃): 7.90 (s, 2H), 7.76 (s, 1H). ¹³C NMR (150 MHz, CDCl₃): 136.88, 135.70, 129.54, 126.77. ¹⁹F NMR (564.5 MHz, CDCl₃): 66.43. ESI-MS calculated for [C₆H₄Cl₂FO₂S, M + H]⁺: 228.9293, Found 228.9291.

2,4,6-trimethylbenzenesulfonyl fluoride (1j):

Prepared according to the general procedure and purified by silica gel column chromatography using hexanes/EtOAc (10:1) as eluent (34% yield): ¹H NMR (600 MHz, CDCl₃): 7.03 (s, 2H), 2.64 (s, 6H), 2.35 (s, 3H). ¹³C NMR (150 MHz, CDCl₃): 145.18, 140.21, 131.96, 129.26, 22.51, 21.34. ¹⁹F NMR (564.5 MHz, CDCl₃): 68.19. ESI-MS calculated for [C₉H₁₂FO₂S, M + H]⁺: 203.0542, Found 203.05430.

Data availability

Data will be made available on request.