Tamagna Mandal†
a,
Anupam Dey†b,
Dipayan Mandalb and
Tapas Kumar Maji*ab
aNew Chemistry Unit (NCU), School of Advanced Materials (SAMat), Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India
bChemistry and Physics of Materials Unit (CPMU), School of Advanced Materials (SAMat), Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India. E-mail: tmaji@jncasr.ac.in; Web: https://www.jncasr.ac.in/faculty/
First published on 25th September 2025
Heterogeneous photoredox catalysis, a powerful approach for activating small molecules, is experiencing a notable resurgence due to the availability and eco-friendly benefits of natural sunlight. Most of the photoredox organic transformations are either electron or hole mediated pathways to drive the reaction with the assistance of sacrificial agents. Recently, the simultaneous use of photogenerated electrons and holes for reductive and oxidative reactions has emerged as an intriguing approach in organic photoredox reactions. In this study, we report the design and synthesis of a donor–acceptor imine based Py-Tz COF, composed of pyrene and tetrazine based building units and explored oxidative thiocyanation and reductive hydroxylation reactions. The efficient photogenerated electron–hole separation, driven by pyrene's strong electron-donating and tetrazine's electron-accepting properties, along with a low exciton binding energy (61.4 meV), facilitated both reactions efficiently, as realized in the transformation of substrates having different functional groups. Py-Tz COF was further employed as a photocatalyst for CO2 reduction, coupled with a thiocyanation reaction within a single redox cycle. By leveraging the distinct oxidation and reduction energy levels of the Py-Tz COF, oxidative thiocyanation and reductive hydroxylation were achieved in a one-pot transformation. The mechanism of each reaction was evaluated by identifying the reaction intermediates through in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), supported by the different photophysical and electron paramagnetic resonance (EPR) studies. Post-catalysis characterizations confirmed that the COF retained its crystallinity and photostability after the reactions. This report highlights the unique application of a metal-free, heterogeneous photocatalyst for two different photoredox reactions conducted in a single pot.
Covalent organic frameworks (COFs), a class of porous crystalline organic polymers, emerge as excellent catalysts for metal-free heterogeneous organic transformation reactions. The development of COFs in photocatalytic organic transformations is stimulated by several advantages related to the robust π-conjugated covalent framework, large surface area, and tunable band gap based on a variety of organic chromophoric building units.14 Various types of heterogeneous photocatalytic organic transformations, like C–H functionalization,15 sulfoxidation,16 amine oxidation,17 C–H borylation,18 thioamide cyclization,19 and alcohol oxidation20 have been reported earlier. However, creating a robust metal-free photocatalyst that can concurrently generate and utilize electron–hole pairs while minimizing coulombic recombination remains a significant challenge. The majority of COF-based photoredox studies have reported dual-catalysis where photoexcited electrons are employed for H2O2 generation, while photogenerated holes participate in oxidative organic transformations. These systems are typically focused on single organic transformations mediated by reactive oxygen species (ROS), with O2 to H2O2 conversion occurring as a side or supportive process. Thus, metal-free polymer-based photocatalysts reported so far have demonstrated either oxidative or reductive processes individually, but not concurrently for two distinct, synthetically valuable organic transformations.21–24 In this regard, the construction of donor–acceptor (D–A) based COFs with low exciton binding energy is very crucial to establish the dual-functionalized photocatalytic system.25,26 The donor–acceptor dyad in COFs facilitates the separation and migration of charge carriers via the intramolecular charge transfer (ICT) process, which can enhance the scope to utilize both the e− and h+ separately for two different redox reactions.27
In this context, we have synthesized a D–A based, Py-Tz COF with low exciton binding energy (61.5 meV) from a donor-pyrene-aldehyde [Py(CHO)4] and tetrazine-amine [Tz(NH2)2] as an acceptor moiety. The well-separated highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) in Py-Tz COF, leading to an intramolecular charge transfer (ICT) band in the visible light, led to efficient separation of electrons (e−) and holes (h+) that was employed for e− and h+ mediated hydroxylation of boronic acids and thiocyanation reaction, respectively (Scheme 1). Both reactions investigated in this study hold significant value in the domain of organic synthesis. Phenols are widely recognized for their importance in natural products, pharmaceuticals, and polymer chemistry.28,29 Similarly, thiocyanation, involving the formation of C–SCN bonds, provides access to organo-thiocyanates an important class of sulphur-containing compounds.30–32 Thus, we developed a green, efficient, and atom-economical photocatalytic method for both thiocyanation and hydroxylation reactions, using a D–A type Py-Tz COF as the photocatalyst. To monitor the real-time progression of both reactions, we performed in situ DRIFTS to track various key intermediates during the catalytic process. Furthermore, to identify the reactive species generated under light irradiation, we employed EPR and steady-state photoluminescence (PL) spectroscopy, which provided deeper insights into the mechanism and reactive species involved in both the photocatalytic transformations. This photocatalysis approach works well for the quantitative transformation of various substrates, including even bulky organic molecules, into the desired products. Py-Tz COF was also utilized to perform dual-functional photocatalytic thiocyanation and hydroxylation reactions in a single pot. Additionally, efficient dual-catalytic transformation was also achieved during simultaneous photocatalytic CO2 to CO conversion (690.2 μmol g−1) and thiocyanation reaction (≈74%). This approach presents a promising strategy for utilizing metal-free COFs in photocatalytic organic transformations while minimizing waste product formation.
Solid-state UV-vis diffuse reflectance spectroscopy of Py-Tz COF revealed a broad absorbance in the visible range with a prominent band at ∼500 nm (Fig. 2a), and the corresponding optical band gap was calculated to be 2.4 eV from the Tauc plot (inset, Fig. 2a). Time-dependent density functional theory (TDDFT) calculations considering the smallest repeating unit of the COF framework suggested that the broad absorption near 500 nm can be associated with the HOMO to LUMO transition at 483.23 nm. Further, the position of the frontier molecular orbitals of the Py-Tz COF was checked by the DFT method, indicating that the HOMO and LUMO were localized on the electron-rich pyrene moiety and the electron-deficient tetrazine moiety, respectively, constructing a perfect donor–acceptor framework (Fig. 2b). Thus, the absorption band corresponds to the HOMO to LUMO transition in Py-Tz COF, which denotes the intramolecular charge transfer (ICT) phenomenon. The additional broad band in the region 550–600 nm in the solid-state UV-vis absorption spectrum of Py-Tz COF could be attributed to the n–π* transition of the tetrazene moiety, as confirmed from the similar experimental band present in the solid-state UV-vis absorption spectrum of Tz(NH2)2 monomeric unit (Fig. S11).37 TDDFT calculation also revealed a theoretical absorption band at 587.5 nm, which can be attributed to a HOMO-2 to LUMO electronic transition. Both the HOMO-2 and LUMO orbitals are localized on the tetrazine moiety, indicating that this band arises from a local excitation within the tetrazine unit (Fig. S12).
A metal-free donor–acceptor-based polymeric catalyst's efficiency strongly depends on a key parameter called Frenkel exciton binding energy (Eb),38 which determines the Coulombic interaction and the chance of electron–hole (excitons) separation during the photocatalysis process. The Py-Tz COF exhibited a low exciton binding energy of 61.4 meV (Fig. S13).39–43 The conduction band (CB) position was calculated to be −0.63 V vs. NHE as calculated from the Mott–Schottky measurement (Fig. S14a). Further, by adding the band gap, the valence band (VB) position was calculated to be 1.79 V vs. NHE (Fig. 2c). The relative band diagram of Py-Tz COF depicts that the CB is more negative than the reduction potential of oxygen (O2) to the superoxide radical anion (O2˙−) (E0 = −0.33 V vs. NHE), while the VB is more positive than the potential of SCN− to SCN˙− (E0 = 0.87 V vs. NHE) (Fig. 2c).44,45 This indicates Py-Tz COF can efficaciously reduce O2 to O2˙− and oxidize SCN− to SCN˙− under visible light irradiation. Photogenerated separation and recombination of charges was assessed through transient photocurrent measurements, revealing minimal current in the absence of light, which significantly increased upon illumination (Fig. S14c). Additionally, electrochemical impedance spectroscopy (EIS) was conducted under both illuminated and dark conditions revealing a reduced semicircle radius in the Nyquist plot under light indicating a lower charge transfer resistance (Fig. S14b). Together, these results demonstrate efficient electron–hole separation in the Py-Tz COF upon light exposure. Moreover, electron paramagnetic resonance was also measured in the presence of light in order to evaluate the ability of Py-Tz COF to generate unpaired electrons. The gradual increase in EPR signal over time strongly indicated the generation and accumulation of photoexcited electron–hole pairs on the COF framework (Fig. S15).46–49
To get insight into the distinct contributions of photogenerated electrons and holes to drive photocatalytic reactions, two reactions were performed separately: electron-driven hydroxylation and hole-driven thiocyanation reaction. At first, the photocatalytic hydroxylation reaction was conducted with 4-formyl-phenylboronic acid (1a) (0.125 mmol) as the model substrate and acetonitrile (5 mL) as the solvent. The reaction was carried out in the presence of oxygen (O2) as the oxidant and triethylamine (TEA) (0.43 M) as the sacrificial electron donor (Table 1; Entry 1). The successful formation of the product was confirmed by the 1H NMR study (Fig. S16). The yield of the product was also obtained from the NMR, which showed an impressive ≈99% yield of 1a after 12 h of continuous visible light irradiation. Control experiments were performed without the photocatalyst, light, or TEA, which resulted in negligible product formation, highlighting the essential role of each component during the photocatalytic process (Table 1; Entries 2–4). Additionally, only trace amounts of product were observed under argon atmosphere, confirming that O2 served as the primary oxidant in this catalysis (Table 1; Entry 5). The reaction yield was further quenched in the presence of AgNO3 and p-benzoquinone as scavengers for electrons and superoxide radical anions (O2˙−), respectively, indicating active participation of these species in the photocatalytic process (Table 1; Entry 6 and 7). The nearly unchanged yield in the presence of L-histidine, a well-known singlet oxygen (1O2) scavenger, indicates negligible formation of 1O2 in the reaction medium (Table 1; Entry 8). Additionally, various boronic acid derivatives with different functionalities (–CHO, –CO2H, –CO2CH3, –Br, and –CN) were tested under optimized reaction conditions to demonstrate the broad applicability of this synthetic protocol. All substrates, regardless of the functional group, were effectively transformed into the intended products with excellent yields of ≈95–99%, demonstrating the effectiveness of Py-Tz COF as a photocatalyst for hydroxylation reactions (Table 2). Next, we performed a hole-driven thiocyanation reaction by Py-Tz COF on aromatic compounds. To optimize the reaction conditions, N,N-dimethylaniline was selected as the model substrate (0.5 mmol), ammonium thiocyanate (NH4SCN) (1.5 mmol) was used as the thiocyanate source, and THF was chosen as the solvent (5 mL) (Table 3; Entry 1). The formation of the desired product was confirmed by 1H NMR study (Fig. S17). Further, several control studies suggested the indispensable role of photocatalyst, light, and O2 during the catalysis (Table 3; Entry 2–4). Moreover, in the presence of a radical scavenger, such as TEMPO, only trace amounts of the product were detected, indicating that radicals are the primary active species, driving the photoredox reaction (Table 3; Entry 5). The addition of p-benzoquinone led to a significant decrease in the reaction yield to ≈14%, indicating that superoxide radicals (O2˙−) are the predominant reactive species (Table 3; Entry 6). In contrast, the thiocyanation reaction proceeded efficiently in the presence of L-histidine, suggesting that singlet oxygen (1O2) does not play a significant role in the reaction pathway (Table 3; Entry 7).50 After establishing the feasibility of the reaction under standard conditions, various aniline derivatives were investigated as substrates. Notably, the substrates containing tertiary amine nitrogen atoms (Table 4) were successfully transformed into the 4-thiocyanated products (Table 4, 2a–2d) with high chemoselectivity and yield (≈91–98%), regardless of the substituents on the nitrogen atom. Further, secondary amines (2e and 2f) were also explored, including heterocycle indole, resulting in the formation of C-3 thiocyanation products with good yield (≈89%). To assess the benefits of the heterogeneous nature of COF as a photocatalyst, we conducted a recyclability test over four consecutive cycles (Fig. S18). After each catalytic reaction, COF samples were recovered through centrifugation, washed with THF and ethanol, dried at 60 °C, and then used for the next cycle. The Py-Tz COF exhibited outstanding recyclability, maintaining consistent yields in each cycle, without notable changes in PXRD, FTIR, FESEM, and TEM analyses (Fig. S19–S21).
Entry | Variation of reaction condition | Yielda (%) |
---|---|---|
a Yield was calculated from 1H NMR using 1,3,5-trimethoxybenzene as the internal standard. (1) Reaction condition: 0.125 mmol of boronic acid derivatives, 5 mL of CH3CN solvent, triethylamine (TEA) (0.43 M), 2 mg catalyst, O2 as oxidant, 12 h under irradiation of 300 W Xe lamp. | ||
1 | No variation | 99% |
2 | Without photocatalyst | Not observed |
3 | Without light | Not observed |
4 | Without TEA | Not observed |
5 | Ar instead of O2 | Trace |
6 | Presence of AgNO3 | 10% |
7 | Presence of p-benzoquinone | 10% |
8 | Presence of L-histidine | 97% |
Entry | Variation of reaction condition | Yielda (%) |
---|---|---|
a Yield was calculated from 1H NMR using 1,3,5-trimethoxybenzene as internal standard. (1) Reaction condition: 1.5 mmol of NH4SCN, 0.5 mmol of the substrate, 5 mL of THF solvent, 2 mg catalyst, O2 as oxidant, 12 h under irradiation of 300 W Xe lamp. | ||
1 | No variation | 96% |
2 | Without photocatalyst | Not observed |
3 | Without light | Not observed |
4 | Ar instead of O2 | Trace |
5 | Presence of TEMPO | Trace |
6 | Presence of p-benzoquinone | 14% |
7 | Presence of L-histidine | 93% |
8 | CO2 instead of O2 | 74% |
To investigate the underlying reaction mechanism, a steady-state photoluminescence quenching experiment was conducted. The broad emission of Py-Tz COF, centered at 480 nm in acetonitrile dispersion, was significantly quenched in an O2-rich atmosphere (Fig. S22a). Time-correlated single photon counting (TCSPC) studies further revealed a reduced lifetime under O2-rich conditions of 1.75 ns compared to the argon atmosphere (2.56 ns) (Fig. S22b). These further demonstrate that, following photoexcitation, rapid electron transfer occurs from Py-Tz COF to O2, leading to the generation of superoxide (O2˙−).51 To detect key reactive oxygen species, a UV-vis experiment was conducted using N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) in the presence of Py-Tz COF, O2, and light. A noticeable increase in absorbance with prominent bands at 516 nm, 560 nm, and 615 nm indicated the formation of the cation–anion adduct between (O2˙−) and TMPD˙+ cationic radical (Fig. 3a). Additionally, EPR spectroscopy was performed with 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) for both hydroxylation and thiocyanation reactions. The strong EPR signal observed under light confirmed the formation of the DMPO-O2˙− adduct, reinforcing that superoxide (O2˙−) was generated through a single electron transfer (SET) process (Fig. 3b and S23).52 In addition, steady-state photoluminescence (PL) quenching was performed with varying concentrations of N,N-dimethylaniline. It was observed that as the concentration of N,N-dimethylaniline increased, the PL intensity was significantly quenched (Fig. 3c).53 The Stern–Volmer constant and electron transfer rate were determined to be 54.32 M−1 and 2.12 × 1010 M−1 S−1, respectively (Fig. S24a, b and Table S1). These data indicate that in the presence of amines, the photoexcited Py-Tz COF undergoes reductive quenching, forming the radical cation of the respective amines. To further reveal the progress of the reaction process and trap the reactive intermediates during the hydroxylation reaction in situ DRIFTS spectroscopy was performed under visible light irradiation and continuous O2 purging conditions. After considering the substrate, triethyl amine, Py-Tz COF, and O2, one background spectrum was collected before visible light irradiation. After visible light irradiation, various peaks were observed in the spectrum to identify the reactive species involved in this reaction. The peaks at 862 cm−1 and 952 cm−1 were attributed to O–O stretching frequencies.54 Notably, the peaks at 1335 cm−1 and 1141 cm−1 were assigned to the O–O vibration of adsorbed oxygen species (O–Oads) and also suggesting initial adsorption of O2 on the catalyst surface in mainly Yeager type.55–57 The peak at 1143 cm−1 suggested adsorbed O2 was reduced to O2˙−.58 Notably, the stretching frequencies for C–O, CC–O, and C–O–H were observed at 1427 cm−1, 1261 cm−1, and 1090 cm−1, respectively (Fig. 3d), indicating the advancement of the hydroxylation reaction.59–61 Similarly, in situ DRIFT experiment was also performed for the thiocyanation reaction in the presence of NH4SCN, Py-Tz COF, O2 and substrate. Similarly, the peaks associated with previously identified intermediates, such as O–Oads, O2˙−, and O–O, were observed, along with a newly appeared peak at 711 cm−1, suggesting the formation of the C–S bond (Fig. 3e).62,63
Furthermore, we conducted the photocatalytic reduction of CO2 in the presence of N,N-dimethylaniline and NH4SCN, aiming to elucidate the mechanistic pathway wherein the photoexcited Py-Tz COF facilitates electron transfer from its conduction band to CO2, leading to CO generation. Simultaneously, the photogenerated holes are expected to drive the thiocyanation of N,N-dimethylaniline (Fig. S25a). The experiment was conducted in a THF medium under a CO2 atmosphere, employing N,N-dimethylaniline (0.5 mmol) and NH4SCN (1.5 mmol) as the sacrificial electron donor and substrates for the thiocyanation reaction (Table 3, Entry 8).33 The photoreduced gaseous product was quantified using gas chromatography-mass spectrometry (GC-MS) analysis of the gas collected from the reactor headspace. After 12 hours of continuous irradiation, 690.2 μmol g−1 of CO was produced, while N,N-dimethyl-4-thiocyanatoaniline was obtained with a yield of ≈74% (Fig. S25b). Along with CO, a negligible amount of H2 was also produced. An isotope labelling experiment conducted with 13CO2 led to the formation of 13CO, providing definitive evidence that the generated CO originates from CO2 (Fig. S26c). Notably, in situ DRIFTS measurements of the thiocyanation reaction coupled with photocatalytic CO2 reduction revealed a gradual emergence of multiple peaks within the 1000–2100 cm−1 range, indicating the sequential formation of various reaction intermediates responsible for CO2 reduction process (Fig. S27). The infrared peak observed at 1526 cm−1 was attributed to the COOH* intermediate, a pivotal species in the transformation of CO2 to CO.25,33,64 Additionally, the observed peaks at 1707 cm−1 (CO bending mode), 1645 cm−1 (CO2˙−), and 2042 cm−1 (*CO), along with the characteristic vibrational signatures at 1378 cm−1 attributed to m-CO32− species, indicate the successful formation of CO, providing strong evidence for the progression of the CO2 reduction process.65,66 Additionally, the peak observed at 731 cm−1 further confirmed the formation of the C–S bond during the thiocyanation reaction (Fig. S23). Thus, the simultaneous utilization of the electron–hole pair was demonstrated through the CO2 reduction reaction (CO2RR) and the thiocyanation reaction.
Further, in order to fully harness the photogenerated charge carriers (e− and h+) within a single photocatalytic system, we carried out both the hydroxylation of boronic acid and the thiocyanation of aniline derivatives in a one-pot reaction, in presence of 4-formylphenylboronic acid (0.125 mmol), NH4SCN (1.5 mmol), N,N-diethylaniline (0.5 mmol), Py-Tz COF, O2 and the reaction was performed in THF medium (Fig. 4a). The successful conversion of both substrates into the desired product was confirmed through NMR analysis (Yieldhydroxylation ≈ 99% and Yieldthiocyanation ≈46%) (Fig. S28). This dual-substrate transformation was enabled by the spatial separation of the HOMO on the pyrene unit and the LUMO on the tetrazine moiety within the COF framework (Fig. 2b). It is proposed that, upon visible light irradiation, intramolecular charge transfer (ICT) from the pyrene to the tetrazine results in the transfer of photoexcited electrons to the tetrazine moiety. These electrons activate the oxygen molecule via a single electron transfer (SET) process, leading to the formation of reactive superoxide radical anion (O2˙−). This O2˙− radical anion, characterized by its Lewis basic nature, was captured by the boron atom of the boronic acids (A), leading to the formation of intermediate B. Subsequently, intermediate B is subjected to intermolecular proton abstraction, resulting in intermediate C. Following this, intermediate C participated in intramolecular aryl group migration, during which the aryl group attacked the oxygen atom and broke the weak O–O bond with the release of the OH− group, leading to the formation of intermediate D. Finally, product-2 was produced from intermediate D through the hydrolysis process. At the same time, the photogenerated hole oxidized the aniline substrate (A′) and SCN− leading to the formation of the corresponding anilinium radical cation (B′) and SCN˙− radical anion. The generated aniline radical cation (B′) transformed into its resonance structure (C′), which was attacked by SCN˙− radical to form an intermediate (D′). Subsequently, the intermediate D′ lost one proton to rearomatize and give the final product-1 (Fig. 4b).67,68
![]() | ||
Fig. 4 (a) Reaction conditions and (b) plausible mechanism of dual-functional photocatalysis mediated by Py-Tz COF. |
Footnote |
† Contributed equally. |
This journal is © The Royal Society of Chemistry 2025 |