A Bifunctional Organic Photocatalyst for Efficient Single‐Electron and Energy Transfer Activation

Abstract

Bifunctional photocatalysts capable of mediating both single‐electron transfer (SET) and energy transfer (EnT) processes are rare and typically metal based. Here, we present 3‐thioaryl‐4‐hydroxycoumarins, a new family of cost‐effective organic photocatalysts that leverage a stabilized charge‐transfer (CT) excited state to achieve both strong reducing power and efficient energy transfer. The spatial separation of the HOMO and LUMO stabilizes the CT state, enhancing SET reactivity (E*_red = −3.08 V vs. SCE) while maintaining a sufficiently high triplet energy (E _T = 67 kcal mol⁻¹) for EnT‐driven transformations. This dual reactivity enables the activation of redox‐inert substrates (E _red < −2.8 V vs. SCE) via SET reduction, generating radicals suitable for diverse C─S, C‐─P, C─B, and C─C bond‐forming transformations, alongside EnT‐based processes such as E/Z olefin isomerization and [2 + 2] photocycloadditions. Mechanistic studies, supported by photophysical and theoretical analyses, confirmed the catalyst's bifunctionality.

Thioaryl hydroxycoumarins are versatile, metal‐free photocatalysts for both single‐electron (SET), and energy transfer (EnT) pathways, enabling radical and triplet energy transfer reactions with a single organic scaffold.

Introduction

Photocatalysis has advanced modern synthesis significantly,^{[ 1 ]} enabling highly selective transformations under mild conditions. Photocatalysts typically activate substrates through single‐electron transfer (SET)^{[ 2 ]} or energy transfer (EnT)^{[ 3} , ^{4 ]} mechanisms. However, achieving both high EnT efficiency and strong SET reducing power within a single photocatalyst remains rare and holds significant potential for streamlining catalyst selection and reaction development. Metal‐based systems, such as cyclometalated Ir(III) complexes, have shown potential as bifunctional photocatalysts (Figure 1a).^{[ 5} , ^{6 ]} These complexes are effective EnT‐based photocatalysts due to their high triplet energy levels (E _T > 62 kcal mol⁻¹) and long lived excited states.^{[ 7} , ⁸ , ^{9 ]} However, their relatively weak reduction potentials (E _red < −1.5 V vs. SCE)^{[ 10 ]} limit their ability to activate substrates via an SET mechanism. To address this limitation, modified Ir(III) complexes incorporating isocyanoborato ligands were designed (Figure 1b),^{[ 11 ]} which exhibited improved properties, including a higher triplet‐state energy (E _T = 68.9 kcal mol⁻¹) and a stronger reducing power (E _red = −2.42 V vs. SCE), enabling effective SET activation of simple organic substrates.^{[ 12 ]} An alternative strategy for bifunctional photocatalysis involves Coulombic dyads,^{[ 13 ]} where electrostatic interactions between a cationic ruthenium complex and an anionic pyrene derivative enable efficient energy and electron transfer.

Figure 1.

a) Iridium‐based bifunctional photocatalysts with high energy transfer (EnT) activity but moderate redox potential. b) Iridium(III) isocyanoborato complex, a strong EnT and SET photocatalyst. c) Thioxanthone, an example of an organic bifunctional photocatalyst with limited redox power. d) Our newly developed class of bifunctional photocatalysts, 3‐thioaryl‐4‐hydroxycoumarins, exhibiting both versatile SET and EnT catalytic functions. Redox potentials (E _red) are reported vs. SCE in CH₃CN.

While transition‐metal systems have proven valuable, organic photocatalysts^{[ 14} , ^{15 ]} offer a compelling alternative due to their cost‐effectiveness and accessibility. However, the development of bifunctional organic photocatalysts capable of efficiently combining strong SET and EnT functions has lagged behind. Aromatic ketones, exemplified by thioxanthone^{[ 16} , ^{17 ]} (Figure 1c), are the main class of organic photocatalysts investigated for dual reactivity.^{[ 18} , ^{19 ]} While thioxanthone is effective in EnT activation, with a triplet energy of 63.4 kcal mol⁻¹, its limited photoreductive power (E*_red = −1.12 V vs. SCE) underscores the challenge of designing organic photocatalysts that can efficiently mediate both SET and EnT processes.

Herein, we disclose 3‐thiophenyl‐4‐hydroxycoumarins (Figure 2d), a family of organic bifunctional photocatalysts that, upon excitation, exhibit strong SET reducing power and high energy transfer capabilities. This dual functionality enables a wide range of transformations, including the activation of redox‐inert substrates containing strong carbon‐fluorine and carbon─chlorine bonds, the Birch reductions of arenes, as well as EnT‐mediated processes like E/Z olefin photoisomerization and [2 + 2] photocycloadditions.

Figure 2.

a) Proposed SET‐based mechanism using the 4‐hydroxycoumarin‐derived catalysts C for the hydrodefunctionalization of redox‐inert aryl halides 1; R‐O⁻ indicates the deprotonated catalyst C⁻ while R‐O ^∙ the resulting radical. b) Catalyst screening and control experiments; reactions performed on a 0.2 mmol scale using 1 equiv. of chloroanisole 1a or fluoroanisole 1b under illumination by a purple LED (Kessil lamp) at 390 nm. c) Synthesis of catalyst C2 and preparation of the active catalytic salt C2⁻ . d) UV–vis absorption spectra of catalyst C2 and the corresponding anion C2⁻ (formed in situ treating C2 with 3 equiv. of Cs₂CO₃) in CH₃CN (10⁻⁴ M). e) Emission of the excited anion C2^−* in CH₃CN (formed in situ treating C2 with Cs₂CO₃) upon irradiation at 354 nm and its intercept at 370 nm with the normalized absorption spectrum, with a 0–0 transition energy (E_0,0) of 3.35 eV. f) Cyclic voltammetry measurements of the deprotonated catalyst anion C2⁻ (oxidation onset) carried out in CH₃CN vs. Ag/AgCl at a scan rate of 100 mV s⁻¹. Estimated redox potential vs. SCE in CH₃CN; NMP: N‐methyl‐2‐pyrrolidone. ^aYields of 2a measured by ¹H NMR analysis using dimethylbromide as the internal standard. ^bPrices of reagents estimated based on the highest quantity and purity available in the Sigma–Aldrich catalogue (February 2025).

Results and Discussion

Background and Photocatalyst Design

Our catalyst design was informed by insights from our recent study on the direct photoexcitation of 4‐hydroxycoumarin derivatives.^{[ 20 ]} Specifically, we discovered that in situ deprotonated 3‐benzyl‐4‐hydroxycoumarin C1 (Figure 2b) could be excited by purple light (390 nm) to then serve as SET reductant, enabling activation of alkyl halides to generate carbon radicals. Additionally, sparse literature reports suggest the capacity of excited 4‐hydroxycoumarins to mediate EnT processes and undergo nonradiative phosphorescence decay.^{[ 21} , ²² , ²³ , ^{24 ]} Building on these findings, we hypothesized that suitably modified 4‐hydroxycoumarins could effectively combine strong SET potential with efficient EnT activity.

We first examined the ability of C1 and the newly designed hydroxycoumarin‐based catalysts C2–C5 as strong SET reductants upon excitation. Mechanistically, we hypothesized that the electron‐rich anion C⁻ , formed in situ by deprotonation of catalyst C, would reach a highly reducing excited state (C⁻*) under light irradiation (Figure 2a). SET activation of a difficult‐to‐reduce electron‐rich C(sp² )—X substrate 1 would generate an aryl radical. γ‐Terpinene, serving as a hydrogen atom donor, was expected to quench this aryl radical via hydrogen‐atom transfer (HAT), yielding the reduced product 2. The resulting cyclohexadienyl radical II would then undergo either SET or HAT^{[ 25 ]} with the radical intermediate I, thereby regenerating the catalyst and completing the catalytic cycle.

Developing Efficient Single‐Electron Transfer Catalysts

To evaluate the reducing power of our organic catalysts, we tested the hydrodechlorination of 4‐chloroanisole (1a, E _red = −2.9 V vs. SCE) and the more challenging defluorination of 4‐fluoroanisole (1b, E _red < −3.0 V vs. SCE), both requiring highly negative redox potentials. Reactions were conducted with 10 mol% of catalysts C, Cs₂CO₃ as the base (3 equiv.), γ‐terpinene, and under irradiation by 390 nm light in N‐methyl‐2‐pyrrolidone (NMP, Figure 2b). The 3‐benzyl‐4‐hydroxycoumarin catalyst C1, with an excited‐state reduction potential (E*_red) = −2.85 V vs. SCE,^{[ 20 ]} yielded 20% of product 2a from 1a but was ineffective for reducing 1b. To enhance the photocatalyst's redox power, we replaced the benzyl fragment on the hydroxycoumarin framework in C1 with a thioaryl group (C2–C5). This modification was designed to leverage the sulfur lone pair to enhance electronic conjugation within the catalyst,^{[ 26 ]} thus improving electron delocalization and stabilizing a charge‐transfer (CT) state for more efficient redox activity.^{[ 27} , ²⁸ , ^{29 ]} The resulting 3‐thiophenyl‐4‐hydroxycoumarin catalyst C2 exhibited remarkable SET activity, affording the target product 2a in 98% yield from 4‐chloroanisole 1a and 50% yield from the fluoro substrate 1b, successfully activating both C(sp² )─Cl and C(sp² )─F bonds. Control experiments showed that the reaction performed worse in the absence of γ‐terpinene (Figure 2b, entry 2), while alternative hydrogen donors, such as i‐Pr₂NEt and sodium formate, led to moderate results (entry 3). Importantly, the absence of Cs₂CO₃ completely inhibited the reactivity (entry 4). Other carbonate bases, such as Na₂CO₃ and K₂CO₃, remained effective but gave slightly reduced yields (80% and 90%, respectively), whereas organic bases (e.g., pyridine, N,N‐diisopropylethylamine, and 1,1,3,3‐tetramethylguanidine) failed to promote the reaction (see Table S1 in the Supporting Information. We also confirmed that the presence of the catalyst and light were essential for the reaction to occur. Notably, catalyst C2 is an air‐stable solid that can be synthesized on a gram scale in a single step from inexpensive, commercially available 4‐hydroxycoumarin and thiophenol (Figure 2c).

We then explored further catalyst modifications (catalysts C3–C5) to evaluate whether better results could be obtained in the reduction of fluorine substrate 1b. Catalyst C5, featuring an electron‐donating methoxy group on the thiophenyl scaffold and an electron‐withdrawing fluorine atom on the coumarin core, exhibited enhanced efficiency in the photoreduction of 1b compared to the simpler catalyst C2 (Figure 2b). This observation suggests that a push‐pull effect, by stabilizing charge‐separated states, may enhance redox activity of our photocatalysts.^{[ 30 ]}

Mechanistic Insights Into SET Activity and Charge‐Transfer State Formation

To understand the photocatalytic SET performance of our catalysts, we conducted both experimental and computational studies. Treatment of catalyst C2 with Cs₂CO₃ (3 equiv.) in CD₃CN rapidly and quantitatively generated the electron‐rich anion C2⁻ , as confirmed by ¹H NMR analysis (see Section F4 in the Supporting Information (SI) for details). Absorption spectroscopic analysis (Figure 2d, see also Section F2.1 in the Supporting Information) revealed that catalyst C2 lacks significant absorption in the visible region. Deprotonation with Cs₂CO₃ to form C2⁻ induced a minimal bathochromic shift, extending absorption to ∼390 nm (the excitation wavelength used in this study) though the extinction coefficient remained low (ε ₃₉₀ = 81 M⁻¹ cm⁻¹ at 1 mM). Upon excitation at 354 nm of anion C2⁻ in CH₃CN (in situ‐generated by mixing C2 with Cs₂CO₃), an emission maximum at 445 nm was detected, confirming that C2⁻ could access an electronically excited state (Figure 2e). From the overlap of normalized UV–visible absorption and emission spectra, the 0−0 transition energy (E₀,₀) was determined to be 3.35 eV at 370 nm. Next, electrochemical studies using cyclic voltammetry revealed that anion C2⁻ exhibited an irreversible oxidation peak at +0.65 V vs. Ag/AgCl in CH₃CN (Figure 2f). Using the Rehm–Weller formalism,^{[ 31 ]} the redox potential of the excited state of anion C2⁻ (E(I/C2⁻*)) was estimated to be −2.78 V vs. SCE, confirming its strong reducing capability upon excitation. Further experimental evidence from Stern–Volmer quenching experiments (see Section F2.6 and Figure S21 for details) demonstrated that the fluorescence of excited C2⁻ was quenched by 4‐chloroanisole 1a (K_SV  = 1.5 × 10⁻³ M⁻¹; k_q  = 5.0 x 10⁵ M⁻¹ s⁻¹), corroborating its potential for productive SET reactivity. Similar experiments were performed for photocatalysts C3–C5, and the results are reported in Section F of the Supporting Information.

Our experiments demonstrated that replacing the benzyl substituent in catalyst C1 with the thiophenyl fragment in catalysts C2 and C5 significantly enhanced photoreductive performance. We therefore used a combination of photophysical data and computational investigations to rationalize the enhanced SET function of C2 and C5 (Figure 3). Photophysical characterization revealed that catalysts C2 and C5 exhibited similar singlet excited‐state lifetimes (3.0 ns for deprotonated C2 and 3.4 ns for deprotonated C5) as catalyst C1 (5.1 ns). However, catalysts C2 and C5 displayed a more pronounced charge‐transfer (CT) state,^{[ 30} , ^{32 ]} as indicated by the increased Stokes shifts compared to C1. Given that the singlet excited‐state lifetimes and redox potentials of the excited states are comparable (Figure 3a, redox potential of the excited anion C5⁻* estimated as −3.08 V vs. SCE), this suggests that the enhanced performance of C2 and C5 is not primarily due to thermodynamic differences but rather kinetic factors. The greater stabilization of the CT state in C2 and C5 may contribute to a more efficient SET process due to more stable excited‐state dynamics and electron‐transfer kinetics.^{[ 27} , ²⁸ , ²⁹ , ^{30 ]}

Figure 3.

DFT‐calculated HOMO–LUMO orbitals for the deprotonated catalysts C1, C2, and C5, along with their experimentally measured photophysical properties; calculations were carried out at B3LYP/6‐31 + G** level of theory. τ _S1 denotes the singlet excited‐state lifetime, ε ₃₉₀ is the molar extinction coefficient at 390 nm (1 mM concentration), the Stokes shift refers to the difference in wavelength between the maxima of absorption and emission spectra of the deprotonated catalysts, and E*_red indicates the excited‐state reduction potentials. b) Absorption and emission spectra of the deprotonated catalysts C⁻ upon laser excitation at 354 nm (10⁻⁴ M solutions in CH₃CN obtained by mixing C1, C2, or C5 in the presence of 3 equiv. of Cs₂CO₃).

To better understand the influence of the substituents on the photocatalyst scaffold, we performed DFT calculations at the B3LYP/6‐31 + G** level of theory, using the PCM solvent model to account for solvation effects. Time‐dependent DFT (TDDFT) calculations revealed that the primary electronic transition (S₀ → S₁) for all deprotonated catalysts C1–C5 was governed by the HOMO–LUMO transition, with notable differences across the series (see Figure 3a and Section F5 in Supporting Information for details). Computational studies showed that in C1, the HOMO was predominantly localized on the carbonyl group of the hydroxycoumarin core, limiting electronic communication within the catalyst. In contrast, the incorporation of a thiophenyl substituent in C2 shifted the HOMO to the thiophenyl ring, lowering its energy by 0.33 eV and enhancing electronic delocalization. This structural modification facilitated the formation of a more stabilized CT state, as corroborated by the observed Stoke shift differences of C2 and C5. Further, computational studies revealed that the introduction of additional substituents in C5, specifically a fluorine electron‐withdrawing group on the hydroxycoumarin core and a methoxy electron‐donating group on the thiophenyl fragment, lowered the HOMO‐LUMO gap by 0.48 eV and further increased the Stokes shift. Overall, these studies, combined with the experimental data, support the conclusion that the enhanced CT state stabilization in C2 and C5 is key to their superior SET properties and improved catalytic performance. To further validate our design strategy, we also computationally screened all possible permutations of MeO and F substituents across the X and Y positions, which confirmed that the substitution pattern in C5 provides the most favorable redshift and push–pull characteristics (see Section F5.1 in the Supporting Information for details).

Developing Efficient Energy Transfer Catalysts

To evaluate the potential of our newly developed 3‐thiophenyl‐4‐hydroxycoumarin‐based compounds C2–C5 to also serve as effective energy transfer catalysts, we first examined their ability to promote the E–Z photoisomerization^{[ 33 ]} of methyl fumarate E‐3a (E _T = 66.2 kcal mol⁻¹) to methyl maleate Z‐3a (E_T = 71.0 kcal mol⁻¹, Figure 4a). The high triplet energy of methyl fumarate E‐3a was selected to challenge our catalysts, pushing them to operate under conditions where a high triplet energy and long‐lived excited states are crucial for driving the transformation. The reactions were performed in CH₃CN using 10 mol% of the EnT catalyst C and under 370 nm light irradiation. Conducting the reaction in the presence of 3 equiv. of Cs₂CO₃, previously used for SET activation, did not provide any E/Z isomerization (entry 1). In contrast, performing the reaction without any base afforded high selectivity for the desired product Z‐3a, yielding an E:Z ratio of 8:92 (entry 2). Given this observation, the neutral forms of other catalysts were tested; however, catalysts C3–C5 proved much less effective (entries 3–5). Finally, control experiments, conducted in the absence of either catalyst or light (entry 6), confirmed the necessity of both. In general, the neutral EnT catalysts performed worse under 390 nm irradiation—used to trigger SET processes— likely due to the poorer absorption at this wavelength (see Scheme S1 in the Supporting Information for details).

Figure 4.

Testing the energy transfer activity. a) Photoisomerization of E‐methyl fumarate 3a; the E/Z ratio was determined by ¹H NMR analysis of the crude mixture. b) Relevant photophysical properties of neutral catalysts C2–C5. c) Normalized emission spectra of catalyst C2 (10⁻⁴ M solution in CH₃CN): blue line represents the spectrum in degassed, deaerated CH₃CN at room temperature; dashed blue line represents the spectrum in aerated CH₃CN at room temperature; red line represents the spectrum in a 77 K solid matrix of degassed CH₃CN. d) Exploring the substrate scope of the energy transfer‐mediated E/Z photoisomerization; reactions performed on a 0.2 mmol scale with 10 mol % of catalyst C2 in dry, degassed CH₃CN (1 mL). Yields refer to the isolated Z isomers of products 3, with the Z/E ratio indicated in parentheses. e) Intermolecular and f) intramolecular EnT‐mediated [2 + 2] photocycloadditions; reactions performed on a 0.2 mmol scale with 10 mol % of catalyst C2 or C3 in dry, degassed solvent (1 mL); yields refer to the isolated products 4–6, while the yield of 7 was measured by ¹H NMR analysis using dibromomethane as the internal standard.

To rationalize these observations, we investigated the photophysical properties of catalysts C2–C5 in their neutral forms, conducting studies in the absence of any base (Figure 4b). The triplet‐state lifetimes and triplet energies were determined from emission studies in a 77 K frozen matrix (Figure 4c shows data for catalyst C2; see Figure S20 in Supporting Information for other studies). Catalyst C2 exhibited a triplet energy E _T of 67.0 kcal mol⁻¹ and a triplet‐state lifetime of 9 µs. The presence of a triplet excited state was further confirmed by comparing the emission spectra of C2 under degassed and aerated conditions at room temperature (Figure 4c, blue vs. dashed blue line), where oxygen quenching indicated the involvement of a long‐lived triplet excited state. As for catalysts C3–C5, which were not effective in promoting E/Z isomerization, this was likely due to their lower triplet energies (E _T ranging from 62.4 to 63.7 kcal mol⁻¹), which may be insufficient to effectively activate E‐3a via EnT sensitization.

To rationalize the difference in EnT activity between the neutral and deprotonated forms of catalyst C2 in the photoisomerization of dimethyl fumarate E‐3a (entries 1 & 2 in Figure 4a), we performed Stern–Volmer fluorescence quenching studies using E‐3a as the quencher (see section F2.6 for details). The deprotonated catalyst C2⁻ displayed linear quenching behavior (K_SV  = 5.3 × 10⁻² M⁻¹), while GC‐MS analysis confirmed the formation of dimethyl succinate, the reduced product derived from 3a (see Scheme S2 in Supporting Information for details), suggesting that a photoinduced SET mechanism is operative in the presence of a base. This is in line with the reduction potential of E‐3a (E _red = −1.55 V vs. Ag/AgCl),^{[ 34 ]} which renders SET thermodynamically feasible for C2⁻. In contrast, the neutral catalyst C2 exhibited a nonlinear, upward‐curved and complex quenching profile (see Figure S23 in Supporting Information), which can be attributed to triplet–triplet EnT processes.^{[ 35 ]} These studies therefore suggest that a mechanistic shift from EnT to SET takes place upon deprotonation of the catalyst.

Next, we explored the reactivity of various E‐olefins 3 in the E/Z photoisomerization using the neutral catalyst C2 (Figure 4d). An unsymmetrical trisubstituted E‐keto ester was isomerized to the corresponding Z‐3b in good yield with high selectivity (97:3 Z/E ratio). The low 53:47 Z/E ratio observed for fumaronitrile 3c is consistent with previous literature reports.^{[ 36 ]} This result can be explained by the fact that both E and Z isomers of 3c have a similar triplet energy of ≈60 kcal mol⁻¹, thus resulting in an approximately equilibrated ratio of isomers at the photostationary state.^{[ 36 ]} We then evaluated the EnT‐based isomerization of the oxime derivative 3d.^{[ 37 ]} The recovered high selectivity (91:9 Z/E ratio) is congruent with the large triplet energy difference between the E‐3d (59 kcal mol⁻¹) and Z‐3d (72 kcal mol⁻¹)^{[ 38 ]} isomers, which aligns well with the catalyst's triplet energies, enabling selective isomerization. To further extend the scope of photoisomerization, we investigated styrene derivatives. While trans‐β‐methylstyrene did not undergo efficient E/Z isomerization (see Figure S1 in Supporting Information for a full list of unreactive or moderately reactive substrates), the introduction of a phosphonate group at the β‐position (substrate 3e) resulted in a Z/E ratio of 91:9 for Z‐3e. This outcome is consistent with the reported strategy of leveraging destabilizing A_1,3‐strain to deconjugate the Z‐alkene chromophore, thereby increasing its triplet energy.^{[ 33} , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ^{43 ]} Also diethyl (E)‐(2‐phenylprop‐1‐en‐1‐yl)phosphonate was efficiently isomerized to Z‐3f in a 93:7 Z/E ratio.^{[ 44 ]}

Beyond E→Z isomerization, our new organic photocatalysts were suitable for mediating another hallmark transformation of energy transfer catalysis: [2 + 2] photocycloadditions.^{[ 45 ]} Specifically, we focused on reactions previously reported in the literature and known to proceed via triplet‐sensitization. In the absence of base, catalyst C2 efficiently catalyzed the EnT‐mediated intermolecular [2 + 2] cycloaddition of coumarin (E _T = 63.2 kcal mol⁻¹) and cyclohexene, yielding cyclobutene product 4 (Figure 4e, left panel).^{[ 46 ]} In this reaction, we also detected a minor amount (∼15%) of coumarin homodimerization, which was the only case where such byproducts were observed in our study. Interestingly, as depicted in Figure 4e, right panel, catalyst C3 outperformed C2 in the EnT‐driven [2 + 2] reaction between styrene and ethyl cinnamate (E_T ≈ 49 kcal mo⁻¹l for methyl cinnamate),^{[ 47 ]} delivering product 5 in 50% yield (C2 offered 25% yield).^{[ 48 ]} These results demonstrate how structurally related catalyst variants can offer a versatile platform to fine‐tune reactivity across different substrates and reaction types. Additionally, C2 smoothly promoted the intramolecular [2 + 2] photocycloaddition of a styrene derivative (Figure 4f, left panel),^{[ 49 ]} delivering product 6 in 58% yield. C2 proved also effective in the intramolecular [2 + 2] cycloaddition of norbornadiene (E _T = 62.3 kcal mol⁻¹)^{[ 11 ]} to quadricyclane 7, affording the product in moderate yield (Figure 4f, right panel). Overall, these results highlight that our catalyst platform is useful for promoting diverse classical triplet‐sensitized photochemical transformations.

Further Synthetic Applications via SET Activation

To fully explore the applicability of our organophotocatalytic system, we focused on the SET activation of redox‐inert aryl halides characterized by highly negative reduction potentials (E _red < −3.0 V vs. SCE) and high bond dissociation energies (∼125 ± 2 kcal mol⁻¹). To enable the SET reducing power of our system, we used catalyst C2 in the presence of Cs₂CO₃ (3 equiv.). C2 was chosen due to its low cost and easy accessibility; however, when its performance was insufficient, we evaluated catalyst C5 as an alternative. We first examined a series of fluoroanisole derivatives, which underwent efficient reductive dehalogenation to afford the corresponding anisoles 2a–c in good yields (Figure 5a). To further assess the catalytic efficiency of C2, we extended the study to both electron‐rich and electron‐deficient chlorides. These substrates, known for their resistance to SET reduction,^{[ 50 ]} were successfully converted into the desired hydrochlorination products 2a, 2d–h, achieving high yields. In cases where lower yields were observed, a simple switch to catalyst C5 led to significantly improved outcomes (e.g., 2c and 2h). Importantly, the high yield of products 2g and 2h demonstrates the compatibility of our catalyst with heteroaromatic chlorides. The versatility of our catalytic system was further demonstrated by the successful reductive detosylation of N,N‐diphenyl‐p‐toluenesulfonamide and N‐(4‐methoxy‐phenyl)‐4‐methyl‐benzenesulfonamide, affording products 2i and 2j in high yield (Figure 5b).^{[ 51 ]} The synthetic utility of our platform was further demonstrated by intercepting the aryl radicals generated upon catalyst‐mediated reductive cleavage of aryl chlorides (Figure 5c). Specifically, the use of bis(pinacolato)diboron (B₂pin₂) and trimethyl phosphite (P(OMe)₃) enabled efficient formation of C(sp² )─B and C(sp² )─P bonds within products 8a–b ^{[ 52 ]} and 8c–d,^{[ 53 ]} respectively. C(sp² )─S bond formation was also achieved by reacting aryl chlorides and aryl thiols, which afforded the target products 8e–f. The experiments in Figure 5c were performed using electron‐rich (para‐MeO) and electron‐poor (para‐CN) aryl chlorides, chosen as model substrates to demonstrate the feasibility of the process regardless of the electronic nature of the radical precursors. In another application, catalyst C2 was effective in the selective mono‐defluorination of trifluorotoluene, achieved via SET reduction (Figure 5d).^{[ 54 ]} The resulting aryl‐CF₂· radical was efficiently trapped with thiophenol, providing a streamlined approach to CF₂─S bond formation (product 8g). Our catalyst's ability to efficiently reduce PhCF₃ (E _red = −3.04 V vs. SCE) was also applied in radical‐mediated C─C bond formation via addition to unactivated olefins (Figure 5e). In contrast to previous methods that required elevated temperatures,^{[ 55 ]} our system operates smoothly at room temperature, delivering product 8h. Also 1,3‐bis(trifluoromethyl)benzene (E _red = −2.07 V vs. SCE) could be readily activated via SET to yield product 8i.

Figure 5.

Synthetic Applications via SET Activation. a) Reductive dehalogenation of aryl fluorides and chlorides. b) N‐Detosylation of selected amine substrates. c) Photocatalytic borylation, phosphorylation, and thiolation of aryl chlorides. d) Selective mono‐defluorination of trifluorotoluene, followed by thiophenol trapping, yielding CF₂─S bond formation. e) Radical‐mediated C─C bond formation via aryl‐CF₂· addition to unactivated olefins. f) Light‐induced Birch reductions of unactivated arenes. g) Radical cyclizations. h) Three‐component radical difunctionalization. Reactions performed on a 0.2 mmol scale using photocatalyst C2 or C5 (10 mol%) in a PhotoRedOx Box TC™ reactor under illumination by a Kessil lamp (max 52 W, λ_max = 390 nm). Yields of products 2 and 9 were determined by ¹H NMR analysis using dimethylbromide as an internal standard, unless otherwise noted. Yields of products 8, 11, and 12 refer to isolated materials after purification. ^a) Yield of the isolated purified product. ^b) Performed using HCO₂Na (1 equiv.) instead of γ‐terpinene. ^c) tBuONa used instead of C₂CO₃. ^d) Anthracene‐9‐carbonitrile was used. NMP: N‐methyl‐2‐pyrrolidone.

We also evaluated the capability of catalyst C2 to drive light‐induced Birch reductions of unactivated arenes (Figure 5f).^{[ 50} , ^{56 ]} This transformation is particularly difficult due to the absence of leaving groups, which often leads to unproductive back‐electron transfer. Our organic catalyst enabled the activation of naphthalene, affording the corresponding dihydro product in good yield (9a). In the case of anthracene‐9‐carbonitrile, the reaction proceeded via decyanation followed by Birch reduction (9b). As for more challenging substrates, such as N‐methyl indole and benzofurane, the use of catalyst C5 proved effective to achieve appreciable yields (adducts 9c–d). We then showcased the versatility of our catalyst in driving radical cyclizations (Figure 5g). The chloro‐substrate 10a was successfully activated to afford 2,3‐dehydrogenated benzofuran 11a,^{[ 57 ]} while the reductive imino cyclization from 10b smoothly led to product 11b. Finally, we used catalyst C2 to promote a three‐component difunctionalization of para‐methoxy styrene with 4‐cyano‐chlorobenzene and 2,2,6,6‐tetramethylpiperidin‐1‐ol (TEMPO‐H), delivering product 12 (Figure 5h).^{[ 58 ]} Overall, these results highlight the potential of our catalytic system for promoting diverse and complex SET‐mediated radical transformations.

Conclusions

In summary, we have introduced 3‐thioaryl‐4‐hydroxycoumarins as a new class of inexpensive, organic bifunctional photocatalysts capable of mediating both SET and EnT processes. These metal‐free catalysts combine strong reducing power—enabling the activation of redox‐inert substrates—with high triplet energies, allowing for efficient energy transfer. Their ability to promote a wide array of C─C, C─S, C─P, and C─B bond‐forming reactions, as well as photochemical isomerizations and cycloadditions, highlights their synthetic utility. Comprehensive mechanistic investigations, supported by photophysical and theoretical analyses, shed light on the basis of this dual activity, which is uncommon among small organic scaffolds. These mechanistic insights may provide a foundation for expanding the scope of organic photocatalysis as a sustainable alternative to transition‐metal‐based systems, while the concept of engineering charge‐transfer excited states to unify SET and EnT pathways may prove useful in guiding future advances in photocatalyst design.

Supporting Information

Details of experimental procedures and full characterization data and copies of NMR spectra (PDF).

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Mallik S., Wang H., Matera N., Li Bi‐X., Stagni S., Melchiorre P., Angew. Chem. Int. Ed. 2025, 64, e202509770. 10.1002/anie.202509770

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.