Ultrathin triphenylamine–perylene diimide polymer with D–A structure for photocatalytic oxidation of N‐heterocycles using ambient air

Abstract

An ultrathin porous triphenylamine–perylene diimide (PDI) polymer photocatalyst with photooxidative ability was prepared. Triphenylamine–PDI oxidizes 1,2,3,4‐tetrahydroisoquinoline to 3,4‐dihydroisoquinoline with a yield of 84% and selectivity of 92% under the following conditions: room temperature, ambient air, aqueous solution and 420 nm irradiation. The yield was 62% higher than that of mesoporous g‐C₃N₄ and 72% higher than that of self‐assembled PDI. The high yield and enhanced catalytic activity of the triphenylamine–PDI were attributed to the introduction of the electron‐donating triphenylamine motif into PDI, thereby forming a donor–acceptor structure, which greatly promotes the separation of photogenerated charges. In addition, the triphenylamine–PDI with high crystallinity of was conducive to fast charge transport. Triphenylamine–PDI was also applicable to other N‐heterocyclic substrates. Moreover, using aqueous media and air as oxidant are consistent with green chemistry, and the photocatalyst can be recycled after the reaction.

INTRODUCTION

As an important synthetic intermediate,¹ 3,4-dihydroisoquinoline (DHIQ) plays a key role in the preparation of morphinanes and isoquinoline alkaloids.² DHIQ is a representative of a class of N-heterocyclic small molecule compounds widely distributed in nature and synthetic materials. The traditional method for the synthesis of dihydroisoquinoline is stemmed from 1,2,3,4-tetrahydroisoquinoline (THIQ), which is dehydrogenated under the induction of hydrogen acceptors,³ or oxidizing agents are used under an argon atmosphere.⁴ The former method has the advantages of simple operation without deoxygenation and the use of air as the oxidant,⁵ which is cheap and easy to obtain. However, there are challenges in the photocatalytic oxidation of such compounds in aqueous media. The immiscibility of substrate and water causes that heterogeneous photocatalysts can only function in pure aqueous or pure organic phases. If the reaction is carried out in the water phase, then it is necessary to rely on a water-soluble small molecule dye as the photocatalyst, resulting in it being difficult to recycle after the reaction.

Semiconductor photocatalysts feature mild reaction conditions⁶ and the use of ambient air as a green oxidant,⁷ providing new options for selective oxidation reactions. In addition, the separation and recovery of heterogeneous semiconductor photocatalysts are very simple. Compared with inorganic photocatalysts, organic polymer photocatalysts have the advantages of transition metals free,⁸ rich chemical modification sites,⁹ and extensive sources of raw materials.¹⁰ Conjugated polymer photocatalysts have a porous structure and powerful photoelectric conversion capabilities,¹¹ which enable them to have wide applications, including in alcohol oxidation,¹² indole thiocyanation,¹³ deuteration,¹⁴ C-C coupling,¹⁵ olefin oxidation¹⁶ and other reactions. Perylene diimide (PDI) is well-known as a stable n-type organic semiconductor. The PDI supramolecular self-assembled structure was first reported by our research group.¹⁷ Its in-depth valence band (VB) potential is beneficial to the efficient degradation of organic pollutants such as phenol, and it is also capable of evolving oxygen¹⁸ by oxidizing H₂O. Additionally, fast transport channels of photogenerated charges can be constructed through π-π stacking between PDI molecules. Furthermore, the combination of PDI and other conductive materials,¹⁹ such as graphene, can further improve its performance owing to the enhancements in photogenerated charge separation and transfer.

Because PDI has both strong oxidizing ability and hydro-lipophilicity, it can be used for the photocatalytic oxidation of various organic compounds. However, pristine PDI has massive crystalline defects and lacks a strong electron donor–acceptor structure, which greatly restricts the further improvement of its photocatalytic activity. Because of these issues, it is necessary to introduce a strong driving force for charge separation into pristine PDI photocatalysts. One method to enhance the charge transfer of PDI is to combine it with other materials to form a heterojunction structure,²⁰ thereby enhancing the interfacial electric field. The other method is to introduce electron donors,²¹ i.e., to drive the donor to inject electrons into the PDI, then broaden and force the intrinsic absorption redshift of the catalyst, providing the intrinsic driving force for photogenerated charge transport. Here, we adopted the latter method and introduced the triphenylamine motif, an electron donor²² frequently used in organic optoelectronic materials, to link PDI monomers. After the introduction, a donor–acceptor (D–A) structure was constructed, which strengthened the built-in electric field and accelerated charge separation,²³ thereby improving photocatalytic performance. Here, the designed triphenylamine–PDI photocatalyst was able to oxidize N-heterocycles resulting in a considerable yield, which was greater than those of several other organic photocatalysts, and it was applicable to diverse substrates. In addition, the catalyst exhibited strong reusability and stability owing to the PDI structural unit being covalently connected. This work provides the feasible use of semiconductor photocatalysts in selective oxidation, with practical application potential.

RESULTS AND DISCUSSION Construction of a high crystalline ultrathin porous triphenylamine–PDI nanopolymer photocatalyst

Hydrazine and tris(4-aminophenyl)amine were used as linkers to prepare hydrazine–PDI²⁴ and triphenylamine–PDI,²⁵ respectively. The solid-state ¹³C-NMR spectra (Figure 1A and S1) revealed that both possessed multiple continuous peaks at 120–135 ppm that overlapped and were difficult to accurately identify²⁶ owing to the large conjugation effect of electronic delocalization. In each, the peak at 160 ppm corresponded to the carbonyl group of PDI. The peak of the phenylene appeared at 148 ppm, indicating that the triphenylamine motif was introduced into the product, and the corresponding hydrazine–PDI did not have this peak. In the XRD pattern (Figure 1B), two characteristic sharp peaks at 12° and 27° corresponded to the PDI motif and π-π stacking,²⁷ respectively, indicating good crystallinity. Also, peaks at 25° and 31° were attributed to distances of π-π stacking interaction between PDI molecules adopting twisted arrangements.¹⁷ As shown in Figure 1C, the triphenylamine–PDI formed nanoplates. Obvious lattice fringes in the HRTEM image (Figure 1D) were observed, implying a highly ordered structure. The fringe width of 0.34 nm was also consistent with π-π stacking. The thickness of the flakes was measured using AFM (Figure 1E) to be ~4.5 nm, which was the thickness of the ultrathin layer structure.

FIGURE 1. Characterization of the morphology and structure of triphenylamine–PDI. (A) Solid-state 13C-NMR spectra; (B) XRD pattern; (C) TEM image; (D) HRTEM image with selected area electronic diffraction image; (E) AFM image and thickness

FT-IR spectra (Figure S2) showed that there were no obvious peaks at 2800–3600 cm⁻¹, indicating that both hydrazine and tris(4-aminophenyl)amine completely reacted, without carboxyl residues of 3,4,9,10-perylene tetracarboxylic dianhydride. The peaks of the carboxyl group and phenylene of PDI were visible. Namely, 3,4,9,10-perylene tetracarboxylic dianhydride showed two broad IR vibrational bands at 1774 and 1720 cm⁻¹, whereas triphenylamine–PDI exhibited sharp bands centered at 1700 and 1663 cm⁻¹, which were associated with CO asymmetric and symmetric stretching frequencies,²⁸ respectively. Similar IR shifts were obtained for hydrazine–PDI. In addition, for the isothermal nitrogen adsorption (Figure S3), triphenylamine–PDI showed a large specific surface area, reaching 385 m²/g (Table S1). This was caused by the formation of a porous structure (Table S2) by the linker that supported the PDI structural unit. The above characteristic results proved the successful preparation of hydrazine–PDI and triphenylamine–PDI polymer photocatalysts.

Donor–acceptor structure of triphenylamine–PDI

To investigate the band structures of PDI polymer photocatalysts, theoretical calculations were used to obtain the HOMO and LUMO of PDI-based photocatalysts. The results are shown in Figure 2A,B. The HOMO of triphenylamine–PDI was contributed by the triphenylamine moiety, whereas the LUMO was contributed by the PDI moiety,²⁹ thereby forming a D–A structure that was different from the molecular orbitals of hydrazine–PDI. The D–A structure was verified using powder ultraviolet–visible-near-infrared diffuse reflectance spectroscopy. (Figure 2C) Triphenylamine–PDI exhibited a broad absorption peak at 700–800 nm, corresponding to the generation of newly formed electron transfer, whereas no such peak was observed in hydrazine–PDI. This D–A structure was favorable for electron transfer.

FIGURE 2. D–A structure of triphenylamine–PDI. HOMO and LUMO diagrams of (A) Hydrazine–PDI and (B) Triphenylamine–PDI; (C) Ultraviolet–visible-near-infrared diffuse reflectance spectroscopy of hydrazine–PDI and triphenylamine–PDI

Furthermore, the positions of VB and conduction band (CB)³⁰ in the PDI-based photocatalysts were investigated using ultraviolet–visible diffuse reflectance spectroscopy (Figure 2C and S4) and the Mott–Schottky method (Figure S5). Flat potentials of hydrazine–PDI and triphenylamine–PDI were measured to be −0.44 V and − 0.60 V (vs Ag/AgCl), respectively. According to the literature,³¹ the conduction band potential is 0.1–0.3 V above the flat band potential. Here, 0.2 V was uniformly taken as the benchmark,³² thus the CB potentials of hydrazine–PDI and triphenylamine–PDI were estimated to be −0.64 V and − 0.80 V (vs Ag/AgCl), respectively. Then, the bandgap was calculated using the Kubelka–Munk formula. The results are listed in Table S3. The VB and CB of triphenylamine–PDI were − 0.60 V and 1.10 V (vs NHE), respectively, and the VB and CB of hydrazine–PDI were − 0.44 V and 1.43 V (vs NHE), respectively. These redox potentials are reasonable and expected to oxidize various substrates.

Visible-light-driven selective oxidation over triphenylamine–PDI

To investigate the photocatalytic performance of designed PDI polymer photocatalysts, oxidation of tetrahydroisoquinoline to dihydroisoquinoline was selected as the model reaction. After irradiation for 15 h, 90% conversion and 92% selectivity were reached, thus the yield was 84% over triphenylamine–PDI (Figure 3A). This yield was 49% higher than that of PDI under the same conditions, and it was also significantly higher than several commonly used organic photocatalysts, such as mpg-C₃N₄, SA–PDI, and SA–TCPP. It should be noted that the yield over SA–PDI was quite low, probably owing to the disintegration of the supramolecular structure. Triphenylamine–PDI showed good catalytic stability and reusability (Figure 3B). After four catalytic cycles, the performance barely decreased, and the structure of triphenylamine remained nearly unchanged (Figure S6), indicating the stability of the covalent linkage. The yields of other N-heterocyclic substrates (Figure 3C), such as indoline and tetrahydroquinoxaline, were all greater than 90%. The yield of tetrahydroquinoline was lower (65%) than that of THIQ, due to the higher potential required (Figure S7b) for its oxidation reaction. However, for substrates without an amino group, the photocatalytic oxidation reaction is more difficult, causing yields to be less than 60%. This is due to the strong electronegativity of O atoms. It is difficult for the outer electrons to leave, and the direct single-electron transfer process is less likely to generate. For aniline, isophorone and toluene, oxidation reactions require the assistance of strong oxidizing reagents and more severe conditions. This was attributed to the fact that the valence electrons of the terminal N or C atoms are stabilized by p-π conjugation.

FIGURE 3. Photocatalytic oxidation reaction using triphenylamine–PDI. (A) Photocatalytic selective oxidation of 1,2,3,4-tetrahydroisoquinoline to 3,4-dihydroisoquinoline over different photocatalysts (mpg-C3N4 = microporous g-C3N4, SA–PDI = self-assembled supramolecular PDI, SA–TCPP = self-assembled tetracarboxylic phenyleneporphyrin, Hyd–PDI = hydrazine–PDI, TPA–PDI = triphenylamine–PDI; (B) Continuous catalytic runs for the oxidation of 1,2,3,4-tetrahydroisoquinoline to 3,4-dihydroisoquinoline over triphenylamine–PDI; (C) Substrate scope. (Reaction conditions are listed in the Section 4.)

The reaction solvent was also optimized (Table S4). The non-polar solvent cyclohexane prevented this reaction. Common weakly polar solvents, such as chloroform and dichloromethane, caused low conversion rates, whereas ethyl acetate, acetone and acetonitrile had relatively higher conversion rates. Among polar solvents, alkanol solvents such as ethanol and isopropanol, inhibited the oxidation reaction. This was attributed to the preferential oxidation³³ of the solvents, not the substrates. Stronger polar solvents, such as DMF and DMSO, caused the increasing conversion. In particular, the conversion rate was low when water was used as the solvent, due to the poor solubility of the reactants. However, if 6% acetonitrile was added to water, then the conversion rate was the same as that of pure acetonitrile. Therefore, the optimal solvent for this reaction was H₂O with 6% (v/v) acetonitrile. As shown by the optimized conditions, this reaction only required a small amount of organic solvent to assist the oxidation of various substrates in an aqueous solution, which is in line with the concept of green chemistry. Additionally, for gram-scale reactions, DHIQ was obtained with an isolated yield of 71% (Figure S8), proving this method may be used not only for small-amount preparations but also for batch synthesis.

Catalytic activity enhancement and proposed reaction mechanism

Prior to the exploration of the catalytic activity enhancement, the main active species in this reaction were investigated, and the trapping of active species (Figure 4A and Table S6) showed that upon triethanolamine addition, the conversion rate was greatly inhibited to only 6%, indicating that the main active species was photoinduced h⁺. As for the superoxide radical (·O₂⁻), it was generated by O₂ through a single electron transfer and did not directly oxidize the substrate. Namely, ·O₂⁻ played a role as a hydrogen acceptor (Figures S9 and S10). When TEMPO was added, the reaction was nearly suspended, indicating that the reaction was conducted via the free radicals pathway. In addition, the control experiment (Table S6) showed that for this reaction, O₂, light and the photocatalyst were all indispensable. A fluorescence quenching experiment (Figure 4B) was performed. Under O₂-free conditions, tetrahydroisoquinoline was added to the triphenylamine–PDI acetonitrile dispersion, and the fluorescence intensity was greatly reduced, indicating that electron transfer occurred.³² Thus, electrons were transferred from THIQ to the catalyst. In the absence of O₂, the electrons from the catalyst were unable to transfer to the dissolved O₂. This confirmed the role of photogenerated h⁺ as the main oxidative active species. Obvious signals can be observed when DMPO was used as the spin trapping reagent, and they corresponded to superoxide radicals (Figure 4C), g = 2.003, confirming the existence of ·O₂⁻.

FIGURE 4. Investigation of the reaction mechanism and catalytic activity enhancement. (A) Confirmation of active species using scavengers; (B) Photoluminescence quenching spectra; (C) EPR spectra of the superoxide radical; (D) Photocurrent spectra; (E) Electrochemical impedance spectra; and (F) Photoluminescence spectra

Importantly, the high photocatalytic activity of triphenylamine–PDI is caused by triphenylamine and PDI forming a D–A structure, which enhances the intrinsic driving force of the charge separation. The improvement in photocatalytic activity was then verified using multiple characterizations. Photocurrent spectra³³ (Figure 4D) showed that the photocurrent of triphenylamine–PDI increased, indicating that the photocatalytic performance was elevated. A Tafel polarization curve³⁴ (Figure S11) showed that under the same bias voltage, the current of triphenylamine–PDI was larger, implying the enhanced charge separation ability. Electrochemical impedance spectroscopy³⁵ (Figure 4E) showed a charge transfer impedance of 1190 Ω/cm² after fitting, and the value of triphenylamine–PDI was 14.7% that of hydrazine–PDI (Table S7), corresponding to a stronger charge conduction capability. The excitonic process was investigated using fluorescence spectroscopy (Figure 4F). The emission intensity of triphenylamine–PDI was significantly reduced compared with that of hydrazine–PDI, demonstrating a suppressed photogenerated charge recombination. These results proved that the D–A structure of triphenylamine–PDI provides a stronger charge separation ability for triphenylamine–PDI than hydrazine–PDI, thereby improving performance and increasing the conversion rate. In contrast, the as-prepared mpg-C₃N₄, SA–PDI, and SA–TCPP do not possess the intermolecular D–A structure and lack the driving force for the migration of photogenerated charges, resulting in the decreased performance.

To completely propose the reaction mechanism, some other crucial intermediates of this reaction were examined. The oxidation reaction of tetrahydroisoquinoline to dihydroisoquinoline followed a two-electron pathway. The O₂ reduction product was H₂O₂. Using the KI-starch method,³⁶ a dark blue reaction solution was observed, indicating H₂O₂ was generated (Figure S12). However, the concentration of H₂O₂ was too low, making it prone to decompose under irradiation. Consequently, it hardly affected the oxidation reaction, and the final product was H₂O. In addition, the two specially designed substrates were both unreactive (Figure S13), indicating that the positions of single electron transfer and hydrogen atom transfer were on the residual amino group and α-methylene group, respectively. When oxidized by h⁺, tetrahydroisoquinoline was simultaneously transformed into an imine cation and H radicals,³⁷ which were trapped by PBN as shown in the ESR spectrum (Figure S14). Using TEMPO as the free radical quencher significantly inhibited the reaction, indicating that the reaction mostly followed a free-radical pathway.³⁸ Therefore, the mechanism of the reaction, illustrated in Figure 5, was proposed as follows: Under light illumination, the efficient charge separation of triphenylamine–PDI was generated. The photogenerated electrons on CB were accepted by the O₂ to form ·O₂⁻. While the photogenerated holes on VB oxidized tetrahydroisoquinoline. Electrons were transferred from THIQ to the photocatalyst to generate imine cations.³⁹ The hydrogen radicals were simultaneously expelled, and then the H radicals and H⁺ were trapped by ·O₂⁻, converted into H₂O₂ and finally decomposed to H₂O.⁴⁰

FIGURE 5. Proposed reaction mechanism for photocatalytic oxidation of THIQ to DHIQ

CONCLUSION

Here, we prepared a PDI polymer photocatalyst with an oxidative ability. By introducing an electron-donating motif, the D–A structure was constructed. The introduction of the triphenylamine motif promoted charge separation, improved photocatalyst performance and enhanced the conversion rate. The covalent linkage overcame the structural disintegration and performance degradation of the SA–PDI supramolecular photocatalyst under neutral conditions caused by the dissociation of hydrogen bonds and greatly enhanced the catalytic stability. The reaction used a green oxidant air in a water-based solution. The main active species in the oxidation reaction was h⁺, and ·O₂⁻ was the hydrogen acceptor. The catalyst had potential applications, and the PDI polymer photocatalyst can be used in green chemistry and selective oxidation.

EXPERIMENTAL SECTION Preparation of triphenylamine–PDI

Perylene-3,4,9,10-tetracarboxylic dianhydride (141.2 mg, 0.362 mmol, 1 eq) and imidazole (40 g) were added to an oven-dried round-bottom flask. The reaction mixture was heated to 150°C under an argon atmosphere until molten. Tris(4-aminophenylene)amine (52.5 mg, 0.181 mmol, 0.5 eq) was added to this mixture and stirred at 190°C for 24 h. Afterward, the reaction mixture was cooled to 100°C, and methanol (20 mL) was added. The precipitate was collected by filtration, followed by washing with DMSO (40 mL) and methanol (20 mL). The residual solid was dried at 100°C overnight in vacuo to produce triphenylamine–PDI, with a yield of 90%.

Typical procedure for photocatalytic oxidation reaction

A 10-mL quartz tube was charged with the substrate (0.1 mmol), solvent (1.0 mL) and photocatalyst (3 mg). The reaction mixture was irradiated by a loop of 20 W 420 nm LED (obtained from Suncat Instruments, 32 mW/cm²), cooled using an electric fan to keep the reaction mixture at ~27°C. After the reaction for 15 h, the mixture was centrifuged to remove undissolved impurities, and the supernatant was then analyzed by gas chromatography using n-dodecane as the internal standard.

For cyclic catalysis, the photocatalyst was collected by centrifugation and filtration, washed with acetonitrile, water and methanol, and then dried at 80°C for 3 h.

ACKNOWLEDGMENTS

This work was supported by the National Science Foundation of China (21872077), the National Key Research and Development Project of China (2020YFA0710304), and the Collaborative Innovation Center for Regional Environmental Quality. H. Zhang and K. Yu contributed equally to this work.

CONFLICT OF INTEREST

The authors declare no conflict of interest.