Open Access Article
Constantinos T. Constantinou
a,
Petros L. Gkizis
a,
Dragos F. Flesariub,
Stamatis K. Servioua,
Georgia A. Zissimou
b,
Panayiotis A. Koutentis
b and
Christoforos G. Kokotos
*a
aLaboratory of Organic Chemistry, Department of Chemistry, National and Kapodistrian University of Athens, Panepistimiopolis 15771, Athens, Greece. E-mail: ckokotos@chem.uoa.gr
bDepartment of Chemistry, University of Cyprus, P.O. Box 20537, 1678 Nicosia, Cyprus
First published on 9th March 2026
3H-Phenothiazin-3-one is a small organic molecule with established biological properties, yet its potential as an organic photocatalyst remains largely unexplored. Herein, we report a green and efficient aerobic oxidation of boronic acids employing 3H-phenothiazin-3-one as the photocatalyst. The transformation proceeds under mild conditions, using green LED irradiation and atmospheric oxygen as the terminal oxidant. A broad range of boronic acids were successfully converted to the corresponding hydroxy derivatives, showcasing the excellent functional-group tolerance and synthetic utility of the developed photochemical protocol. Preliminary mechanistic investigations were undertaken to elucidate the reaction pathway and to identify the reactive oxygen species involved.
The direct conversion of boronic acids or esters into hydroxy compounds is well-studied.6 Several photochemical protocols for the aerobic oxidation of boronic acids have been reported, using various photocatalysts. Jorgensen and co-workers employed a ruthenium-based catalyst in DMF (Scheme 1A).7 From a pharmaceutical industry perspective, the use of DMF and prolonged reaction time limits the practicality of this approach and raises environmental concerns, due to DMF's toxicity. In 2013, Scaiano and co-workers investigated methylene blue as an organic photocatalyst for the same transformation (Scheme 1B).8 Compared to the ruthenium complex [Ru(bpy)3Cl2]·6H2O, methylene blue offered superior mechanistic performance, although the substrate scope remained limited. Similarly, Zhang and co-workers utilized Rose Bengal as the photocatalyst (Scheme 1C),9 but this protocol required high catalyst loading (5 mol%). Our group has focused on developing eco-friendly, aerobic photochemical protocols.10 Using CFL lamps as the irradiation source, we identified 2,2-dimethoxyphenylacetophenone as a photoinitiator with N,N′-diisopropylethylamine (DIPEA) (7.5 equiv.) as sacrificial donor (Scheme 1D).11 Replacing 2,2-dimethoxyphenylacetophenone with anthraquinone allowed a 10-fold reduction in DIPEA loading (Scheme 1D),12 and further studies led to a catalyst-free UV-light protocol (370 nm) combined with a low amount of DIPEA under mild reaction conditions, further expanding the substrate scope.13
Despite these advances, milder reaction conditions using lower energy irradiation sources, such as green LEDs remain largely unexplored.14 A literature example of green LED-mediated oxidation of boronic acids employed a Zr-based MOF as the photocatalyst. Building on our prior studies of quinonimines as photocatalysts in aerobic oxidation,15 we investigated their application to the aerobic oxidation of boronic acids to the corresponding hydroxy derivatives. Quinonimines exhibit multiple biological and antioxidant activities,16 and molecules such as 3H-phenothiazin-3-one have demonstrated diverse biological effects, including antiparasitic activity,17 enzyme inhibition (e.g., succinoxidase, cytochrome oxidase, and cytochromes P450), and psychopharmacological effects.17–20 While phenothiazine derivatives, such as methylene blue21 or methylene violet,22 have been widely employed as photocatalysts, the structurally simpler 3H-phenothiazin-3-one remains underexplored in photochemical applications. Previous studies from the Koutentis group, described the synthesis of the quinonimines investigated in this work.16a,23,24 Building on these studies and our earlier investigations of the effect of irradiation wavelength on the aerobic oxidation process of boronic acids,13 we developed a mild, green and sustainable aerobic photooxidation protocol that converts boronic acids to hydroxy derivatives using 3H-phenothiazin-3-one as the photocatalyst under green light irradiation (Scheme 1E).
![]() | ||
| Scheme 2 Quinonimine-based photocatalysts for the aerobic oxidation of phenylboronic acid (1a) under green LED irradiation (525 nm, 40 W). | ||
Upon identifying phenothiazinone 3j as the most efficient photocatalyst, we investigated solvent effects (Table 1) (see SI, Sect. S2). Polar aprotic solvents, such as dichloromethane, chloroform or ethyl acetate, afforded moderate yields (entries 2–4). Polar protic solvents (EtOH, i-PrOH) gave variable results, with i-PrOH completely suppressing reactivity (entries 5 and 6), while ethereal solvents (dioxane, THF) afforded moderate yields (entries 7 and 8). Reactions in non-polar solvents were sluggish (petroleum ether, cyclohexane; entries 9 and 10), likely due to poor solubility of the polar substrate and catalyst. Aromatic solvents, particularly toluene and xylene, performed best among the alternatives (entries 11–13), whereas benzene gave only moderate yield. The highly polar aprotic solvent dimethylsulfoxide (DMSO) completely suppressed reactivity (entry 14), likely due to quenching of the photocatalyst excited state or interference with electron-transfer processes. Although DMF and DMAc showed excellent performance by 1H NMR analysis, they were not pursued, due to their poor green credentials (entries 15 and 16). Water was also evaluated as a greener alternative, but low solubility of the organic substrate and catalyst resulted in only trace product (entry 17).
| Entry | Solvent | 2a a(%) |
|---|---|---|
| a The reaction was performed with phenylboronic acid (1a) (24 mg, 0.20 mmol), DIPEA (26 mg, 0.20 mmol) and catalyst 3j (1.0 mol%) in solvent (1 mL), under green LED (Kessil PR160L, 525 nm, 40 W) irradiation for 18 h. Conversion was determined by 1H NMR. Yield of isolated product after column chromatography in parenthesis. | ||
| 1 | MeCN | 100 (90) |
| 2 | CH2Cl2 | 57 |
| 3 | CHCl3 | 29 |
| 4 | EtOAc | 60 |
| 5 | EtOH | 65 |
| 6 | i-PrOH | 0 |
| 7 | Dioxane | 55 |
| 8 | THF | 32 |
| 9 | Pet. Ether (40–60 °C) | 7 |
| 10 | Cyclohexane | 14 |
| 11 | Benzene | 65 |
| 12 | Toluene | 87 (58) |
| 13 | Xylene | 84 (79) |
| 14 | DMSO | 0 |
| 15 | DMF | 90 (61) |
| 16 | DMAc | 82 |
| 17 | H2O | 2 |
| 18 | H2O–MeCN (1 : 1) |
85 (80) |
We next investigated the effect of the additive on the reaction outcome (Table 2). Having established DIPEA as the base, we first examined the effect of lowering its equivalents (see SI, Sect. S2). Decreasing the amount of DIPEA led to a gradual decline in yield, with 0.75 and 0.5 equivalents affording 63% and 50% of phenol 2a, respectively (entries 2 and 3). Use of primary amines, such as propylamine, were ineffective (5% yield, entry 4), while secondary amines, including morpholine or N-methylmorpholine, afforded poor yields (18% and 27%, entries 6 and 7). Tertiary amines, such as triethylamine, provided a moderate yield (76%, entry 5). The biomimetic reducing agent Hantzsch ester was also tested as a hydrogen donor, affording phenol 2a in 46% yield (entry 8). Finally, inorganic base K2CO3 proved completely ineffective, with no product observed (entry 9).
| Entry | Additive (equiv.) | 2a a (%) |
|---|---|---|
| a The reaction was performed with phenylboronic acid (1a) (24 mg, 0.20 mmol) and catalyst 3j (1 mol%) in MeCN (1 mL), under green LED (Kessil PR160L, 525 nm, 40 W) irradiation for 18 h.b Yield of isolated product after column chromatography. | ||
| 1b | DIPEA (1) | 90 |
| 2 | DIPEA (0.75) | 63 |
| 3 | DIPEA (0.5) | 50 |
| 4 | Propylamine (1) | 5 |
| 5 | Et3N (1) | 76 |
| 6 | Morpholine (1) | 18 |
| 7 | N-Methylmorpholine (1) | 27 |
| 8 | Hantzsch ester (1) | 46 |
| 9 | K2CO3 (1) | 0 |
Having established the optimal reaction conditions, we next turned our attention to the substrate scope and functional group tolerance of the photooxidative protocol. A variety of boronic acids were subjected to the optimised conditions (Scheme 3). The reaction tolerates both electron-withdrawing and electron-releasing substituents at ortho-, meta- and para-positions of the phenyl boronic acid, affording the corresponding phenols in good to excellent yields (2a–p), indicating only a subtle electronic influence on the reaction efficiency. Minor variations in yields may be attributed to the possible sublimation during their isolation. The position of the substituent on the aryl ring did not impede the reaction, although slight differences in yield were observed (2h–r). Notably, in all cases, functional groups including methyl (2b, 2h, 2i), nitro (2c), trifluoromethyl (2d), cyano (2e, 2p), halogens (2f, 2j–o), and methoxy (2g) remained intact under the reaction conditions. Fused arylboronic acids were also well tolerated, giving the corresponding phenols in good to excellent yields (2q–s). More challenging substrates were then examined. 1,1′-Diphenyl-4,4′-bisboronic acid (1t), was converted to the corresponding 1,1′-diphenol (2t) in 60% yield. Heteroarylboronic acids (2w–x) were also tested: pyrid-3-ylboronic acid (1u) gave pyridin-3-ol (2u) in 53% yield without any N-oxidised side product. In contrast, fur-3-ylboronic acid (1w) and thien-3-ylboronic acid (1x) produced complex mixtures that were not further investigated. Finally, an alkylboronic acid, cyclohexylboronic acid (1v) gave cyclohexanol (2v) in 71% yield, with no overoxidized products, such as cyclohexanone, observed.
With phenothiazinone 3j established as an efficient photocatalyst for the aerobic photochemical oxidation of boronic acids, we next investigated the reaction mechanism. The involvement of radical intermediates was probed by conducting the reaction in the presence of TEMPO or BHT as radical scavengers (see SI, Table S4). In both cases, formation of the desired product was significantly suppressed, consistent with a radical pathway. Additional control experiments further demonstrated the necessity of all reaction components (Table 3).
UV-Vis studies were conducted to probe the possible formation of an electron donor–acceptor (EDA) complex (see SI, Sect. S5), as the pivotal role of EDA complexes in photochemical reactions is well established.25 Mixing phenylboronic acid (1a) with DIPEA did not produce a bathochromic shift in the green-light region, and the addition of phenothiazinone 3j did not induce further spectral changes (SI, Fig. S6). However, upon irradiation of phenothiazinone 3j at 525 nm in the presence of DIPEA, a significant change in the UV-Vis spectrum was observed, suggesting that the excited catalyst interacts with DIPEA (SI, Fig. S5). Prior to irradiation, the mixture of catalyst 3j and DIPEA showed absorption bands at 367 and 500 nm, corresponding to the catalyst. After irradiation at 525 nm for 120 s, the original bands disappeared, and a single band at 323 nm appeared, characteristic of phenothiazines (SI, Fig. S5).26
To further probe the role of DIPEA and its interaction with the excited phenothiazinone catalyst 3j, a series of NMR experiments were conducted (see SI, Sect. S6). Irradiation of catalyst 3j or DIPEA individually produced no observable changes. However, upon irradiation of a mixture of catalyst 3j with DIPEA at 525 nm in CDCl3 for 3 h, consumption of catalyst 3j was observed, accompanied by the formation of 3-hydroxyphenothiazine salt 4 and 3-hydroxy-phenothiazine sulfoxide 5 (Scheme 4A). The formation of sulfoxide 5 can be rationalized by a photochemical aerobic sulfide oxidation. The aerobic oxidation of salt 4 to the correspondingsulfoxide 5 is well-precedented. Such oxidations of diaryl sulfides typically proceed via SET between the sulfide and molecular oxygen, generating superoxide radical anion.10c,27,28 Therefore, the formation of sulfoxide 5 provides evidence that superoxide anion is generated under our photochemical conditions.20 When the same irradiation experiment was carried out in the presence of phenylboronic acid (1a), no formation of by-product 5 was detected. Further evidence for the generation of superoxide anion came from the H2O2 test (see SI, Sect. S8). H2O2 was detected upon irradiation of phenothiazinone 3j with DIPEA at 525 nm, whereas irradiation of 3j in the absence of DIPEA did not produce H2O2. Additional experiments were conducted to gain insight into the fate of the 3-hydroxyphenothiazine salt 4. Irradiation of a mixture of phenothiazinone 3j with DIPEA at 525 nm under an oxygen-free atmosphere (argon) resulted exclusively in the formation of salt 4 (Scheme 4B). When the reaction mixture was subsequently stirred under air and treated with 1.0 equivalent of acetic acid to quench the salt, regeneration of catalyst 3j was observed (Scheme 4C). Furthermore, when phenylboronic acid (1a) was added instead of acetic acid, phenol (2a) was formed (Scheme 4D) (see SI, Sect. S6).
Cyclic voltammetry (CV) measurements of phenothiazinone 3j showed a reduction potential of Ered = −0.66 V vs. SCE (see SI, Sect. S9). The excited-state reduction potential of catalyst 3j was estimated at
vs. SCE, indicating that the excited state is a strong oxidant capable of accepting an electron from DIPEA (Ered = 0.64 V vs. SCE).15 To evaluate possible side pathways, we assessed whether DIPEA could be oxidized by 1O2 to generate superoxide. This process is thermodynamically disfavoured: the reduction potential of 1O2 (Ered = −0.87 V vs. SCE)27 is more negative than the oxidation potential of DIPEA (Eox = 0.64 V vs. SCE). Consequently, oxidation of DIPEA by 1O2, to form superoxide is not supported.8 Additionally, the quantum yield (Φ) of the photochemical reaction was determined following the procedure by Yoon and co-workers.29 The value of Φ = 0.28 is consistent with a closed catalytic cycle rather than a radical chain mechanism.
Summarizing our findings, a plausible mechanistic pathway is depicted in Scheme 5. Upon irradiation at 525 nm, phenothiazinone 3j is excited to its singlet state and via intersystem crossing is led to its triplet state 3j*. The excited photocatalyst then interacts with DIPEA via a SET event to produce the radical anion of photocatalyst 3j, which proceeds to be quenched by molecular oxygen and provides superoxide radical anion that oxidizes boronic acid 1a, leading to intermediate II (Scheme 5, Path A top). A HAT event leads to III, which upon rearrangement and hydrolysis leads to the desired product. Our mechanistic (NMR and UV-Vis) studies further support an alternative possible interaction between the excited state catalyst 3j* and DIPEA, leading to species V. Under aerobic conditions, V is unstable and undergoes oxidation, generating hydrogen peroxide, which can contribute to the oxidation of boronic acids (Scheme 5, Path B).
![]() | ||
| Scheme 5 Proposed mechanistic pathway for the photochemical aerobic oxidation of phenylboronic acid (1a). | ||
| This journal is © The Royal Society of Chemistry 2026 |