Introduction

Phenolic compounds, abundantly present in natural and bioactive molecules, have attracted substantial interest in scientific exploration due to their vital role as synthetic intermediates. Among these, the para-aminophenol framework exhibits tremendous potential for drug development (Fig. 1A)1,2. Paracetamol, a widely utilized analgesic and antipyretic agent, stands out for its favorable side effect profile3,4. Similarly, Fenretinide and AM404 show promise as anti-cancer agents, with Fenretinide demonstrating the rapeutic efficacy against breast, ovarian, and prostate cancers5,6. Conventional industrial methods for para-aminophenol synthesis entail nitrification, hydrolysis, and reduction, yielding significant quantities of by-products and waste. Significant strides have been made in selective hydroxylation using hydroxylamines (Fig. 1B). Since Bamberger et al.’s pioneering work on para-aminophenol synthesis via hydroxylamine rearrangement in sulfuric acid solution7. Notably, Maulide et.al. and Gao et.al. have contributed valuable research on efficient hydroxylation using strontium catalysis and fluorosulfury imidazolium salt, respectively8,9. Obviously, using simple and unfunctionalized arenes via the oxidative C(sp2)-H hydroxylation is a straightforward approach to phenols, which effectively decreases the generation of waste and simplifies the reaction process. Transition metal catalysis is used to achieve this regioselective oxidative hydroxylation, in which directing groups are necessary to improve the reactivity and site selectivity10,11,12,13,14,15,16,17. Alternative approaches, including the utilization of enzymatic peptide ligands for iron catalysis or stoichiometric peroxides as oxygen sources, have been explored18,19,20,21. Nevertheless, external oxidants are often required in these processes, fueling the quest for greener oxidation strategies in organic synthesis.

Fig. 1: Applications and synthesis strategies of hydroxylated compounds.
figure 1

A Examples of bioactive compounds. B para-Aminophenol was synthesized by industrial methods and Bamberger Rearrangement. C Photochemical and electrophotochemical hydroxylation using H2O as a hydroxyl source. D Electrophotochemical oxidative hydroxylation of anilides.

Full size image

The direct C(sp2)-H hydroxylation of arenes using water (H2O) as a hydroxyl source proves enticing yet challenging, attributed to the weak nucleophilicity of H2O22. Noteworthy progress has been achieved by Tung and Wu et al. through the amalgamation of photoredox and cobalt catalysis for hydroxylation of aromatic hydrocarbons using water without external oxidants23. In this context, electrophotochemical synthesis emerges as a powerful protocol to address over-oxidation and over-reduction issues caused by high redox potentials, gaining prominence as an environmentally friendly and versatile tool in green synthesis24,25,26,27,28,29,30,31,32. Utilizing DDQ as a photocatalyst, Lambert et.al. achieved electrophotocatalytic C(sp2)-H hydroxylation of nondirectional arenes via arene radical cation intermediates. (Fig. 1C)33 Despite these photochemical and electrophotochemical advances, region-selectivity remained elusive, and challenges persisted with electron-rich arenes34.

Results

To tackle the above limitations, we direct our attention to anilide, a fundamental building block, wherein direct para-hydroxylation stands as an ideal approach for the molecules of interest in Fig. 1A. The crux of the challenge lies in the ease with which the hydroxylated product (Eox = 1.70 V for 4-N-Boc aminophenol) can be easily oxidized compared to the substrate (Eox = 2.28 V for N-Boc aniline), leading to undesired deposition of phenolic products on the electrode surface (Fig. 1D). The interaction between eosin-like photocatalysts and amine molecules, forming ion-association complexes, has been documented in previous studies, influencing the optical properties of photocatalyst35,36,37,38. Building on this insight, we propose that manipulating the chemistry of substrate through its interaction with the photocatalyst presents a promising avenue for achieving para-hydroxylation of arylamines through electrochemical means. This approach holds substantial potential for controlled synthesis and enhanced selectivity in the production of para-hydroxylated arylamines.

To investigate the potential influence of the interaction between eosin-like photocatalysts and anilide substrates on the electrochemical properties of the substrates, cyclic voltammetry (CV) experiments were conducted. After subjecting the eosin-like photocatalysts to various substrates under Blue LEDs for 30 min, substantial alterations in the oxidative potential of the substrates were observed, high-lighting the efficient capability of eosin photocatalyst to modify the oxidative potential of anilides. Surprisingly, when different substrates were combined with 4′,5′-dichlorofluorescein (DCFS) as the photocatalyst (red line in Fig. 2A), the oxidative potential significantly decreased. For instance, the oxidative potential of N-Boc aniline decreased to 1.84 V (Ep/2 vs Ag/AgCl) when mixed with DCFS (Fig. 2B), already surpassing the inherent oxidation potential of N-Boc aniline about 2.28 V (Ep/2 vs Ag/AgCl, blue line in Fig. 2B). Furthermore, upon exposure to DCFS and illuminated for 30 min, the oxidative potential was further decreased to 1.60 V (Ep/2 vs Ag/AgCl, wine red line in Fig. 2B). However, the oxidative potential of the hydroxylated product (3a) exhibited minimal change or a potential increase, both in the presence of the photocatalyst and the excited state photocatalyst (Fig. 2D). These results accentuate the effectiveness of the activated photocatalyst in diminishing the oxidative capacity of anilide.

Fig. 2: Cyclic voltammetry experiment and optimization of reaction conditions.
figure 2

A Oxidation potential of photocatalyst mixed with the substrate under light. red line: used DCFS as photocatalyst. blue line: used Fluorescein as photocatalyst. yellow line: used Eosin Y as photocatalyst. green line: used Acid Red 92 as photocatalyst. grey line: no photocatalyst. B Cyclic voltammetry experiment of N-Boc aniline and DCFS. C Optimization of reaction conditions. D Cyclic voltammetry experiment of 4-N-Boc aminophenol and DCFS. [a] standard conditions: N-Boc aniline 1 (0.6 mmol), 2a (0.3 mmol), DCFS (10 mol%), nBu4NBF4 (1.0 equiv.), TMBA (4.0 equiv.), DMAc (6 mL). CR as an electrode, Blue LEDs, 10 mA, 10 h. [b] Yield of NMR, with 1,3,5-trimethylbenzene as internal standard. [c] Yields of isolated products.

Full size image

Building upon the insights from the CV experiments, DCFS was selected as the photocatalyst, and N-Boc aniline (1a) as the substrate. Employing constant current and Blue LEDs irradiation (Fig. 2C), we conducted a comprehensive screening of reaction conditions to identify the optimal parameters. After extensive experimentation, it was determined that using 4′,5′-dichlorofluorescein (DCFS) as the photocatalyst, 2,4,6-trimethylbenzioc acid (TMBA) as the acid additive, tetra-butyl tetrafluoroborate (nBu4NBF4) as the electrolyte, and N, N-dimethylacetamide (DMAc) as the solvent, coupled with a constant current of 10 mA and Blue LEDs light irradiation for 10 h, resulted in the para-hydroxylation product 3a with an 80% yield and a para/ortho product ratio exceeding 25/1 (Fig. 2C, Entry 1). Notably, the removal of Blue LEDs during the reaction led to a decreased yield of only 42% and a para/ortho ratio of approximately 10/1 (Fig. 2C, Entry 2). In the absence of both DCFS and Blue LEDs, the yield decreased further to 38%, and the para/ortho ratio dropped to less than 8/1 (Fig. 2C, Entry 3). Furthermore, in the absence of current, no desired product was formed (Fig. 2C, Entry 5). Additionally, it was observed that the reaction was sensitive to air, with the yield significantly decreasing under ambient atmosphere (Fig. 2C, Entry 6). Detailed screening results can be found in the Supplementary Information (Supplementary Table 1).

Scope of substrates

With the optimal conditions established, we proceeded to explore the substrate scope for this electrophotochemical para-selective oxidative hydroxylation (Fig. 3). Generally, ortho- or meta-substituted anilides exhibited good compatibility and provided satisfactory yields. ortho-substituted anilides with Me, Et, iPr, CH2OH, Ph, OMe, OEt, and F substituents were competent reaction partners, affording the corresponding products in moderate to good yields (3a-3i). Electron withdrawing groups, such as ketone groups, were well-tolerated, resulting in the desired product with a 64% yield (3j). Similarly, meta-substituted anilides were investigated under standard conditions (3k-3o). Notably, both terminal and internal alkenes were compatible (3p, 3q). Furthermore, anilides with alkyne substituents were examined, demonstrating tolerance towards terminal alkyne and Ph, cyclopropyl, and nBu on the internal alkyne substituents, leading to products in yields ranging from 54% to 89% (3r-3u). Anilide derivatives with multiple substituents were also tested, and para-selective hydroxylation products were obtained in good yields of 60% to 87% (3v-3x). The method proved amenable to various amide protective groups, including carbomethoxy (4a), carbethoxy (4b), N-butoxycarbonyl (4c), carbobenzoxy (4d), and allyloxycarbonyl (4e), yielding para-selective products in moderate to good yields. Both primary and tertiary amides were well-suited (4f-4h). For cyclic amides, such as those with adamantane, three-membered, four-membered, and six-membered rings, the products were formed with singular para-selectivity.

Fig. 3: Scope of substrates.
figure 3

All values indicate the yield of the isolated product. Reaction conditions: 1 (2.0 equiv.), 2a (0.3 mmol), DCFS (10 mol%), nBu4NBF4 (1.0 equiv.), TMBA (4.0 equiv.), DMAc (6 mL). CR as electrode, 10 mA, Blue LEDs, 10 h. [a] 1a (2.0 equiv.), 2a (6 mmol), DCFS (10 mol%), nBu4NBF4 (1.0 equiv.), TMBA (4.0 equiv.), DMAc (100 mL). CR as the anode, CF as the cathode 10 mA, and Blue LEDs. [b] Derivatization by 6c, refer to the Supplementary Information for details.

Full size image

With the phenol group holding significant importance, the direct installation of hydroxyl groups on biologically active molecules and their derivatives carried valuable significance. Our method proved to be an alternative means for late-stage selective hydroxylation of biological molecules. Late-stage functionalizations of Flurbiprofen, Ketoprofen, and Zatoprofen derivatives were tested, yielding hydroxylation products with moderate to high yields (5a-5c). Additionally, this protocol found utility in Cinnamic acid derivatives commonly found in flavors (5d). Furthermore, we applied this method to important molecules such as amino acids, achieving yields of 35% to 70% (5e-5g). Hydroxylation of the pesticide Propham was also accomplished with a 92% yield (5 h). In a subsequent gram-scale reaction on Propham, the desired product was obtained with a 75% yield, showcasing the practical applicability of this method. The inhibitor of endocannabinoid reuptake, AM404 (6a), could be successfully synthesized under standard conditions with a 71% yield. The method was also suitable for the synthesis of Fenretinide (6b), an excellent antitumor drug, achieving a yield of 62%. Paracetamol (6c) could be prepared in one step using this electrochemical method. Furthermore, we conducted derivative experiments using Paracetamol as a raw material and successfully synthesized Acetaminosalol (6d) and Practolol (6e) in excellent yields. Moreover, we successfully removed the protecting group from para-aminophenols, obtaining para-aminophenol derivatives with good yields (4-amino-3-fluorophenol 6 f, 4-amino-3-ethylphenol 6 g, Supplementary Fig. 5).

Mechanistic studies

The isotope labeling experiment using H218O confirmed the role of water as the hydroxyl source, with the 18O-labeled product 7 being isolated in a 68% yield under standard conditions, with only a trace amount of 16O-labeled product detected (Fig. 4A). A series of kinetic investigations were conducted with the objective of elucidating the intricate relationship between 1a, water, and electricity within the reaction (Fig. 4B and Supplementary Fig. 6). The kinetic analysis of 1a revealed a first-order kinetics pattern. Notably, as the concentration of water increased, the initial rates exhibited minimal fluctuation, indicating a zero-order kinetic behavior concerning H2O. Furthermore, an initial first-order kinetic trend was observed at lower electricity levels. However, a continued increase in electricity resulted in potential overoxidation of the product (Supplementary Fig. 6C). To further explore the C(sp2)-H activation of anilide, intermolecular kinetic isotope effect (KIE) experiments were conducted utilizing N-Boc aniline (1a) and N-Boc aniline-D5 (1a-[D]) (Fig. 4C). Two parallel reactions were carried out under standard conditions, yielding a calculated value of KH/KD of 1.3. Additionally, an intermolecular competition reaction was executed within a single electrochemical cell, and the isotope effect of KH/KD was determined to be 1.25 (Supplementary Fig. 7B). These KIE results indicate that the C(sp2)-H bond cleavage of anilide does not determine the rate of the reaction. Consequently, based on the KIE and kinetic studies, it can be inferred that the rate-determining step is the electrochemical oxidation reaction of anilide.

Fig. 4: Mechanistic experiments.
figure 4

A Isotope labeling experiment. B Kinetic experiments. C Kinetic Isotope Effect (KIE) experiments. D Radical trapping experiments. E Electron Paramagnetic Resonance (EPR) experiments. F Electrochemical ESI-MS experiments.

Full size image

To investigate the presence of radical intermediates in the reaction, 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was employed as the trapping reagent (Fig. 4D). High-resolution mass spectrometry (HRMS) identified the presence of product 8 or 8′, wherein free radicals were effectively trapped. Subsequently, electric paramagnetic response (EPR) experiments were conducted, revealing a signal under standard conditions (Fig. 4E). The analysis indicated the presence of both a hydrogen radical (g = 2.0075, AN = 14.6 G, AH = 19.6 G) and a carbon-centered radical (g = 2.0073, AN = 14.5 G, AH = 19.9 G). Based on fitting, these radicals are proposed to be carbon radical intermediates 10. Furthermore, electrochemical ESI-MS experiments yielded direct evidence concerning the active intermediates involved in this electrochemical reaction (Fig. 4F). Following the cessation of electrolysis, solely the molecular ionization peak of 1a ([M+Na]+, m/z 216.0995, detected at 216.1000) was observable (Supplementary Fig. 10). However, upon subjecting the system to electric current for a duration of two minutes, a notable decline in the molecular ion peak intensity of 1a was observed. This reduction was concomitant with the emergence of an aromatic radical intermediate 13/14 ([M+Na]•+, m/z 215.0917, detected at 215.0922), an aromatic carbocation intermediate 15/16 ([M]+, m/z 192.1019, detected at 192.1028), and the formation of product 3a/3a′ ([M+Na]•+, m/z 232.0945, detected at 232.0951)39,40.

In NMR titration experiments, we explored the interaction of acetanilide with DCFS in the use of DMSO-d6 as solvent (Fig. 5A). We observed a prominent downfield shift in the N-H resonance of acetanilide, along with a smaller downfield shift in the hydrogen signal of the β-carbon. Other hydrogen signals in acetanilide remained largely unchanged. Importantly, in the combined spectra, the DCFS peak between 6.7 ppm and 6.8 ppm exhibited an upfield shift (Supplementary Fig. 3A). To further probe this interaction, we conducted fluorescence assays (Fig. 5B), where an enhancement in fluorescence intensity confirmed the interaction upon adding acetanilide to DCFS. Conversely, the addition of acetaminophen to DCFS resulted in only a slight fluorescence quenching. UV-visible spectroscopy further substantiated the interaction between activated DCFS and compound 1a (Supplementary Fig. 9), marked by an increased absorption peak at 531 nm and a subtle redshift.

Fig. 5: Role of photocatalyst and proposed mechanism.
figure 5

A NMR titration experiments. NMR frequency: 400 MHz, NMR solvent: DMSO-d6. B Fluorescence quenching experiment. C The proposed mechanism.

Full size image

Based on the mechanistic insights obtained from the above results, a plausible mechanism for the electrophotochemical para-hydroxylation of anilides is proposed (Fig. 5C). The coordination of substrate 1a (N-Boc aniline) with photocatalyst DCFS generates intermediate 11, which is then excited to state 12 under Blue LEDs irradiation. Anodic oxidation and deprotonation of 12 lead to the formation of radical intermediate 13/14. Secondary oxidation at the anode produces cationic intermediates 15/16. Since the effect of DCFS may lead to the enrichment of 16 relative to 15, and the position of the aromatic ring is increased due to the steric effect of DCFS, favoring the nucleophilic attack of H2O to 16, resulting in the formation of 17. But the specific mechanism of action needs further41. Subsequent dehydro-aromatization of 17 leads to the formation of the target product 3a and regeneration of the photocatalyst. Additionally, there is also a process in which 1a is anodically oxidized to form the 15/16 intermediate, which is then attacked by water nucleophilic to form our target product 3a/3a’ (Supplementary Fig. 11B). In the cathode, protons are reduced to H2, as detected by gas chromatography.

Discussion

In conclusion, we have successfully developed an electrophotochemical para-selective C(sp2)-H bond hydroxylation through precise tuning the oxidative potential of anilides. This approach has demonstrated high efficacy, enabling the transformation of a diverse array of anilide substrates into para-hydroxylation products with good regional selectivity and wide functional group applicability. Moreover, the protocol proves amenable to late-stage functionalization of biologically active molecules, further emphasizing its versatility and practical utility. Mechanistic studies elucidate the role of the excited state photocatalyst in decreasing the oxidative potential of anilides and enhancing the regional selectivity of the reaction. Overall, our electrophotochemical route offers a potent and efficient strategy for achieving selective anilide hydroxylation, illuminating valuable insights into the manipulation of oxidative potentials in organic synthesis and drug discovery.

Methods

In an oven-dried undivided three-necked bottle (10 mL) equipped with a stir bar. The bottle was equipped with two carbon rods (ϕ 6 mm). A solution of aniline (0.6 mmol), H2O (0.3 mmol), DCFS (0.03 mmol, 10.0 mol%), nBu4NBF4 (0.3 mmol) and TMBA (1.2 mmol) in DMAc (6 mL) stirring under nitrogen atmosphere was electrolyzed at a constant current of 10.0 mA and irradiated by blue LEDs at room temperature for 10 h. After completion of the reaction, it was quenched by saturated NaHCO3(aq.). The aqueous solution was extracted with EA (3 × 15 mL) and the combined extracts were dried with anhydrous Na2SO4. The solvent was removed under reduced pressure by rotary evaporation. Then, the pure product was obtained by flash column chromatography on silica gel (eluent: EA / PE = 1 / 5).