Spontaneous exciton dissociation in organic photocatalyst under ambient conditions for highly efficient synthesis of hydrogen peroxide

Significance

Hydrogen peroxide is a highly competitive ready-to-use product for solar energy transformation. Nevertheless, the contemporary photosynthetic systems are not efficient enough, due to severe charge recombination caused by high activation energy and binding energy of the exciton. Herein, we achieve spontaneous exciton dissociation at room temperature. Moreover, the photosynthesis of H₂O₂ reaches between 9,366 and 12,324 µmol·g⁻¹ from 9 AM to 4 PM in ambient conditions, that is, sunlight irradiation, real water including fresh water and seawater, room temperature, and open air. The ultrahigh photocatalytic efficiency in ambient conditions allows the solar-to-chemical conversion in a real cost-effective and sustainable way, which represents an important step toward real applications.

Abstract

Hydrogen peroxide (H₂O₂) is a highly competitive ready-to-use product for solar energy transformation. However, charge recombination caused by the inefficient dissociation of exciton into free charges severely constrains the photocatalytic efficiencies, especially in ambient conditions. Herein, the photosynthesis of H₂O₂ is achieved in ambient conditions, that is, real water, open air, and sunlight irradiation, by a donor–bridge–acceptor conjugated polymeric photocatalyst with the remarkable productivity reaching between 9,366 and 12,324 µmol·g⁻¹ from 9 AM to 4 PM. The photosynthesis efficiency of H₂O₂ in ambient conditions is even higher than all of the reported systems conducted in pure water and O₂ atmosphere. The remarkably high efficiency is attributed to the spontaneously dissociated exciton at room temperature and the substantially suppressed back electron transfer through storing the photoinduced electron in redox electron acceptors. This efficient photosynthesis in ambient conditions allows the solar-to-chemical conversion in a real cost-effective and sustainable way, which represents an important step toward real applications.

Transformation of solar energy to ready-to-use resources is critical for transitioning toward a sustainable future (1, 2). In the most studied solar-to-chemical conversion (SCC) reactions, that is, CO₂ reduction into C1 and C2 produces, dinitrogen reduction into ammonia, and water reduction into hydrogen, only half-reactions are in demand. By contrast, hydrogen peroxide (H₂O₂) production can fully utilize both the water oxidation and O₂ reduction half-reactions, which avoids the consumption of sacrificial reagents, and thus achieves high SCC efficiency and is more cost effective (3, 4). Moreover, the Gibbs free energy change of producing H₂O₂ from O₂ and H₂O is much lower than the other conversion reactions. Thus, the photons in the full region of visible light (380 nm to 780 nm) are sufficient to actuate this reaction according to the energies of the photons. In addition, the reduction of dissolved O₂ in water to H₂O₂ occurs prior to the reduction of protons, N₂, and CO₂ considering the standard reduction potentials, which indicates that it is thermodynamically possible to exclude the competitive reactions (5). Nowadays, the application of H₂O₂ is still expanding; besides its contemporary wide applications, it is a potential value-added fuel (6).

Organic photocatalysts that are constituted of earth-abundant elements and can be precisely tuned at molecular levels are used in most studies for the synthesis of H₂O₂ (2, 7, 8). However, the photosynthesis of H₂O₂ is not efficient enough, due to severe charge recombination. The generation of an exciton is a distinct characteristic of organic photocatalysts after absorbing photons (Fig. 1A, pathway 1), which is one pair of an electron and a hole bound by Coulombic force, owing to the low dielectric constant for organic materials (9, 10). To acquire completely dissociated charges, the Coulombic force must be overcome within the exciton lifetime (Fig. 1A, pathway 2); otherwise, the charges will undergo geminate recombination (Fig. 1A, pathway 3). The exciton binding energy (E_b) that is used to characterize the strength of the Coulombic force is typically on the order of 0.1 eV to 0.5 eV, which far outweighs the thermal energy (25 meV) at room temperature. Thus, an extra exciton activation energy (E_a) even higher than E_b is required to dissociate the exciton. Generally, the high E_a results in inefficient exciton dissociation, and the positive E_b even indicates that charge recombination is thermodynamically preferred (2, 11).

Fig. 1.

The mechanism behind the excellent performance in photosynthesis of H₂O₂ by TPT-alkynyl-AQ. (A and B) The formation pathways for free charges generation in previous works and this work. (C) Illustration of the pathway for electron transfer and storage.

To overcome this limitation, the strategies of promoting exciton dissociation (Fig. 1A, pathway 2) are reported (12, 13). For example, an external electric field or magnetic field is often required to provide a driving force opposite to the Coulombic force for promoting exciton dissociation (13). Also, to decrease E_a by tailoring the structures of the photocatalysts, the construction of strong charge-delocalized environments by fabricating electron donor–acceptor structures is validated as an effective way (14, 15). Through this method, the E_a can be decreased from larger than 100 meV to dozens of megaelectron volts. Still, the E_a is higher than the thermal energy at room temperature (25 meV). In addition, strong charge delocalization is a double-edged sword: Back electron transfer would be promoted simultaneously (Fig. 1A, pathway 3), which hinders the efficient generation of free charge carriers. Remarkably, the construction of built-in electric fields by fabricating heterojunctions and doping heteroatoms on traditional organic semiconductors is beneficial for decreasing both E_a and E_b, by providing driving forces and stabilizing charge carriers (16–18). Nevertheless, the driving forces originating from the built-in electric fields only exist in limited regions within the photocatalysts, that is, the interfaces of the heterojunctions and the regions around the doped atoms. Moreover, heterojunctions require additional energies for driving electrons or holes across the interfaces (9).

Owing to the lack of efficient strategies for promoting exciton dissociation, most organic photocatalytic systems need a continuous supply of pure O₂ and even organic hole scavengers to accelerate the reactions between charge carriers and reactants (4, 19). Thus, the apparent charge separation efficiency could be elevated. Photocatalytic synthesis of H₂O₂ in ambient conditions, that is, open air, real water, and solar light irradiation, remains insufficiently exploited.

Herein, we rationally design a donor–bridge–acceptor (D-B-A) conjugated polymeric photocatalyst (triphenyltriazine-alkynyl-anthraquinone [TPT-alkynyl-AQ]) exhibiting spontaneous exciton dissociation at room temperature (Fig. 1C). The spontaneous exciton dissociation is achieved by the cooperative action of the rigid alkynyl electron bridges and the redox electron acceptors (AQ) containing electron storage properties. The D-B-A structure forms a strong charge delocalization environment, and the AQ moieties provide a long-lived charge separation state (Fig. 1B, pathway 2 promoted, pathways 3 and 4 suppressed). Moreover, the alkynyl and AQ moieties present abundant reaction sites for efficient oxygen reduction reaction (ORR) and water oxidation reaction (WOR). Notably, the photosynthetic efficiency of H₂O₂ reaches 2,368 µmol·g⁻¹·h⁻¹ in open air and pure water. This efficiency is higher than all of the reported photocatalysts, even those using pure O₂ for the production of H₂O₂. Moreover, the photosynthesis of H₂O₂ is achieved in ambient conditions, that is, sunlight irradiation, real water including fresh water and seawater, room temperature, and open air.

Results

Structure and Activity of the Photocatalysts.

The D-B-A conjugated polymer (CP) was synthesized with TPT moieties as the electron donors, and with built-in redox AQ moieties as the electron acceptors with alkynyl as connectors, which was named TPT-alkynyl-AQ (Fig. 2A). For comparison, D-A conjugated polymer (TPT-AQ) and another D-B-A conjugated polymer (TPT-imine-AQ; imine groups were the bridges between TPT and AQ moieties) were also synthesized (SI Appendix, Figs. S1–S9). AQ moieties exhibit excellent electron accepting ability as the lowest unoccupied molecular orbit of these CPs mainly located in them (SI Appendix, Fig. S1).

Fig. 2.

The structure and photocatalytic performance of CPs for H₂O₂ production. (A) Molecular structures of CPs used in this study. (B) The electronic structure of CPs. (C) The UV-visible diffuse reflectance spectra of CPs and the AQY of TPT-alkynyl-AQ at the specified wavelengths. (D) The photocatalytic performance of CPs for H₂O₂ production in open air and pure water. (E) The photosynthetic amount of H₂O₂ from TPT-alkynyl-AQ in pure water, river water collected from the Pearl River, or seawater from the South China Sea under ambient conditions, that is, open air and sunlight irradiation (from 9 AM to 4 PM).

The powder X-ray diffraction profiles revealed that all CPs exhibited the features of amorphous carbon (SI Appendix, Fig. S2) (20). The photos from scanning electron microscopy and transmission electron microscopy indicated the CPs were exfoliated structures (SI Appendix, Figs. S3 and S4). The formation of triazine ring; the retaining of AQ function groups; the existing of linkages, that is, alkynyl and imine between TPT and AQ in TPT-alkynyl-AQ and TPT-imine-AQ; and the porosity of these three polymers were confirmed carefully and in detail by various characterization methods (SI Appendix, Figs. S5–S9) (7, 21, 22).

The electronic structures of the CPs were also determined (Fig. 2B) (23, 24). As shown in Fig. 2C, CPs exhibited high absorption throughout the visible light range, which was consistent with the ash black colors of the CPs. The absorption band edges indicated that the bandgaps were 2.03, 2.36, and 2.23 eV for TPT-alkynyl-AQ, TPT-imine-AQ, and TPT-AQ, respectively (Fig. 2B). Moreover, the conduction band (CB) minima of TPT-alkynyl-AQ, TPT-imine-AQ, and TPT-AQ were determined to be −0.14, −0.3, and −0.4 eV versus the reversible hydrogen electrode via Mott–Schottky tests (SI Appendix, Fig. S10). Thereafter, the valence band maxima were calculated from the bandgaps and CBs, which were further confirmed by cyclic voltammetry measurement (SI Appendix, Fig. S11). All CPs had the capacities for 2e⁻ ORR, and 2e⁻ or 4e⁻ WOR (25).

The photosynthetic activities of the conjugated polymers were evaluated under xenon-lamp light with the UV light cut off (>400 nm, 100 mW·cm⁻²). No sacrificial agents or continuous O₂ bubbling was adopted. TPT-alkynyl-AQ showed the highest H₂O₂ production rate, that is, 2,368 µmol·g⁻¹·h⁻¹, among the CPs, which was 2 and 5 times that of TPT-AQ and TPT-imine-AQ (Fig. 2D). And the apparent quantum yield (AQY) of TPT-alkynyl-AQ reaches as high as 25%, 18%, and 14% at 365, 425, and 450 nm, respectively, compared to only 10% and 7% at 450 nm for TPT-AQ and TPT-imine-AQ (Fig. 2C). Besides, in pure O₂ atmosphere, the efficiency of TPT-alkynyl-AQ was further elevated to as high as 3,214 µmol g⁻¹·h⁻¹. To the best of our knowledge, the efficiency of TPT-alkynyl-AQ in open air was even higher than the reported photocatalysts used in pure O₂ atmosphere (SI Appendix, Table S1) (4, 5, 17, 26–31). More importantly, the efficiency of SCC is as high as 0.35% in pure water and open air, which was 3 times higher than the average solar-to-biomass conversion efficiency of typical plants (∼0.1%) (SI Appendix, Table S2) (32–36). In addition, TPT-alkynyl-AQ maintained its high efficiency for five cycles without any decrease. Also, the morphology and the component of TPT-alkynyl-AQ were maintained (SI Appendix, Figs. S12 and S13). Meanwhile, the temperature-dependent photocatalytic generation of hydrogen peroxide was investigated. The photosynthetic efficiency of H₂O₂ reaches 564 µmol·g⁻¹·h⁻¹ at 277 K, which was much less than its performance at 298 and 308 K, attaining 2,030 and 2,067 µmol·g⁻¹·h⁻¹. At 333 K, the photosynthetic efficiency of H₂O₂ was only 1,503 µmol·g⁻¹·h⁻¹, because hydrogen peroxide was decomposed with the increase in temperature (SI Appendix, Fig. S14).

Furthermore, in ambient conditions, that is, open air and sunlight irradiation (from 9 AM to 4 PM), 9,295 µmol·g⁻¹ of H₂O₂ was produced in pure water (Fig. 3 and SI Appendix, Figs. S15 and S1). For comparison, 12,042 and 9,366 µmol·g⁻¹ of H₂O₂ were produced in river water; also, 10,634 and 12,324 µmol·g⁻¹ of H₂O₂ were produced in seawater (Figs. 2E and 3 and SI Appendix, Figs. S15 and S16). TPT-alkynyl-AQ maintained its high efficiency for five cycles in real water without any decrease, which indicated that the complex matrices did not influence the photocatalytic efficiencies (SI Appendix, Fig. S17). Photosynthesis of H₂O₂ has never been realized under sunlight irradiation and in real water before.

Fig. 3.

Photosynthesis of H₂O₂ in ambient conditions. (A) The conditions for photosynthesis of H₂O₂ in ambient conditions, that is, sunlight, collected location of water samples, weather conditions, and temperature at the corresponding time. (B) Generated rates of H₂O₂ in various water samples under ambient conditions.

Charge Separation and Transfer Performance.

The reason behind the remarkable photosynthetic efficiency in ambient conditions was investigated. The overall recombination efficiencies of excitons (Fig. 1B, pathways 3 and 4) were monitored by steady-state photoluminescence (PL) emission spectroscopy (SI Appendix, Fig. S18) (17). The radiative recombination of excitons was negligible in TPT-alkynyl-AQ, while it was still significant in TPT-AQ and TPT-imine-AQ, indicating that the geminate charge recombination was greatly inhibited in TPT-alkynyl-AQ. The attribution of fluorescence peaks was also determined by fluorescence lifetime (SI Appendix, Fig. S19). The lifetimes of TPT-alkynyl-AQ, TPT-AQ, and TPT-imine-AQ are 0.74, 0.69, and 0.8 ns, respectively, by fitting PL decay curves, which were more consistent with the lifetime of exciton recombination than free carrier recombination. To further evaluate the influences of electron bridges and electron acceptors in CPs for inhibiting the exciton geminate recombination, exciton activation energy (E_a) was characterized through temperature-dependent PL spectra (Fig. 4 A and B and SI Appendix, Fig. S20) (8, 9). Surprisingly, the PL intensity increased with temperature for TPT-alkynyl-AQ, which was in sharp contrast to the other two organic semiconductors. This indicated that the energy barrier for exciton dissociation into free charge was lower than the thermal energy at room temperature, and the energy level of the charge-separated state was even lower than that of the exciton state (Fig. 4 C and D) (37, 38). In other words, E_a was smaller than 25 meV, and E_b was negative. This was rarely observed in organic photoactive materials, and demonstrated that the separation of excitons was spontaneous in TPT-alkynyl-AQ at room temperature. Thus, the geminate recombination (Fig. 1B, pathway 3) was highly depressed through promoting pathway 2. Based on the temperature-dependent PL, the E_a was estimated to be 42 meV for TPT-imine-AQ. In comparison, the unusual increase in PL with temperature indicated that E_a was smaller than the reverse potential barrier (E_ar), as shown in Fig. 4D. Only the activation energy of charge recombination (E_ar) could be estimated for TPT-alkynyl-AQ, due to the reversed temperature-dependent PL intensity, which was 22 meV. Thus, E_a for TPT-alkynyl-AQ should be even smaller than 22 meV, which verified the assumption above. Neither E_a nor E_ar could be estimated for TPT-AQ, as the PL intensity varied irregularly with the temperature. Furthermore, E_a for TPT-alkynyl-TPT (a new photocatalyst without AQ moieties; SI Appendix, Figs. S20 and S21) was estimated to be 70 meV. These results demonstrated that both the alkynyl bridges and the AQ moieties were crucial for the decreasing of E_a.

Fig. 4.

Characterization of intrinsic properties of CPs. (A) Temperature-dependent photoluminescence (PL) spectra for TPT-alkynyl-AQ. (B) Evolution of PL intensity as a function of temperature from 100 K to 300 K for TPT-alkynyl-AQ. (C) Illustration of mutual transitions between the charge separated state (CS) and the lowest singlet excited state (S₁) in the common materials in which the value of exciton binding energy (E_b) is positive and E_a is the activation energy from S₁ to CS. (D) Illustration of mutual transitions between CS and S₁ in TPT-alkynyl-AQ, in which the value of E_b is negative and E_ar is the activation energy from CS to S₁. (E) The EPR spectra of CPs under ambient conditions without additional illumination. (F) The odd-electron density in a single TPT-alkynyl-AQ unit, through DFT calculation.

To further investigate the embedded mechanism behind the low E_a, density functional dispersion (DFT) calculation was adopted. The introduction of the alkynyl electron bridges largely promoted the coplanarity of the TPT and AQ moieties. The dihedral angle between TPT and AQ was only 0.8° in TPT-alkynyl-AQ, while it was as large as 36° and 46° in TPT-AQ and TPT-imine-AQ, respectively (SI Appendix, Fig. S22). Thus, it was easier for AQ to withdraw the electron cloud from TPT after introducing the alkynyl bridge, to achieve a higher charge delocalization state. The calculated dihedral angle between two TPTs beside the alkynyl bridge was 0.036° in TPT-alkynyl-TPT.

It was highly remarkable that a stable charge separation state was observed in TPT-alkynyl-AQ at room temperature, as evidenced by the radical signal probed by electron paramagnetic resonance (EPR) analysis (Fig. 4E). The EPR signal was at g = 2.006, demonstrating that the radicals were mainly formed on the O atoms in the quinone groups. DFT calculations also confirmed the photoinduced electrons were mainly stored in the AQ moieties (Fig. 4F). These results indicated that AQ moieties could store photoinduced electrons under ambient conditions to avoid back electron transfer (Fig. 1B, pathway 3). Moreover, the strongest radical signal of TPT-alkynyl-AQ verified that the electron storage capacity of the AQ moieties was improved by the introduction of the alkynyl electron bridges, which should also be attributed to the better coplanarity for TPT-alkynyl-AQ.

In addition, the intramolecular charge transfer was characterized by transient absorption spectroscopy (TAS). Very broad TAS signals, between 900 and 1,200 nm, were observed, due to a range of energies for trap states (SI Appendix, Fig. S23), which were further assigned to electrons based on the amplitude decreased in the presence of AgNO₃ as electron scavengers (39). The lifetimes of photoinduced electrons were at picosecond levels in TPT-alkynyl-AQ and TPT-AQ, and at nanosecond level in TPT-alkynyl-TPT (SI Appendix, Fig. S24). The ultrashort carrier lifetime was attributing to the intramolecular electron transfer from electron donor to acceptor. The photoinduced electron did not directly react with the free oxygen, but was stored in AQ moieties or used to reduce the oxygen absorbed on alkynyl moieties, which is discussed in more detail in the following section. As a result, although the carrier lifetime was very short, the ORR reaction was still very efficient. The extremely short lifetimes of electrons in TPT-alkynyl-AQ and TPT-AQ were consistent with the reported electron lifetimes for intramolecular transfer (9). The intramolecular transfer in TPT-alkynyl-AQ was also supported by DFT calculations, as no interlayer interaction between electron donors was observed (SI Appendix, Fig. S25). Thus, the recombination of free charges that occurred during intermolecular charge transfer in TPT-alkynyl-AQ was diminished (Fig. 1B, pathway 4).

Active Sites and Pathways for H₂O₂ Generation.

As the 4e⁻ ORR that produces H₂O instead of H₂O₂ is a competitive reaction of the 2e⁻ ORR, the selectivity of ORR when using these three photocatalysts, that is, TPT-alkynyl-AQ, TPT-AQ, and TPT-imine-AQ, was investigated by rotating disk electrode voltammetry (SI Appendix, Fig. S26) (40). All three of these photocatalysts showed high ORR selectivity. The average electron transfer number involved in ORR was calculated to be 2.124, 2.097, and 2.004 for TPT-alkynyl-AQ, TPT- AQ, and TPT- imine-AQ, respectively, which were close to the theoretical value for directly and selectively reducing O₂ into H₂O₂.

To elucidate the reaction routes, the active sites for ORR were studied first. DFT calculations results indicated that the spontaneous chemisorption of O₂ dominantly occurred on the alkynyl electron bridges to form the endoperoxide species rather than on triazine ring or AQ moieties (SI Appendix, Fig. S27) (7). Furthermore, to investigate the subsequent ORR for the O₂ absorbed on alkynyl, photosynthesis of H₂O₂ by TPT-alkynyl-AQ and TPT-AQ was conducted under illumination in Ar atmosphere with a hole scavenger to eliminate the O₂ from the surrounding environment and water (SI Appendix, Fig. S28). The generation efficiencies of H₂O₂ by TPT-alkynyl-AQ (124 µmol⋅g⁻¹⋅h⁻¹) was 9 times higher than that in the TPT-AQ system (14 µmol⋅g⁻¹⋅h⁻¹) (SI Appendix, Fig. S28C). These strongly evidenced that the preadsorbed O₂ on the alkynyl electron bridges could be subsequently reduced to H₂O₂.

Moreover, it was observed that the electrons stored in AQ moieties could reduce O₂ into H₂O₂. The photosynthesis of H₂O₂ was conducted in Ar atmosphere, and then pure O₂ gas was injected into the photocatalytic system in the dark. The generation of H₂O₂ lasted for 5 min in the TPT-alkynyl-AQ photocatalytic system after stopping the illumination, while no H₂O₂ was generated in TPT-alkynyl-TPT photocatalytic system (SI Appendix, Fig. S29). These phenomena indicated that the AQ moieties were another active site for ORR in TPT-alkynyl-AQ. And the reaction rate between reduced AQ and O₂ was relatively slow, as it took about 5 min to consume the photoinduced electrons stored in AQ under high concentrations of O₂.

Moreover, the rate-determining step of ORR in TPT-alkynyl-AQ was verified. Increasing the concentration of O₂, both the photosynthesis efficiencies of H₂O₂ and photocurrent densities were enhanced in TPT-alkynyl-AQ but not in TPT-AQ (SI Appendix, Fig. S30). The photosynthesis efficiencies of H₂O₂ in open air and in pure O₂ atmosphere achieved by TPT-alkynyl-AQ were 2,368 and 3,241 µmol⋅g⁻¹⋅h⁻¹, respectively (SI Appendix, Fig. S30A). And the photocurrent density of TPT-alkynyl-AQ in O₂ was nearly twice that in Ar atmosphere (SI Appendix, Fig. S30C). All of these indicated that the supply rate of O₂ determined the ORR rate in TPT-alkynyl-AQ, while the reaction rate between reduced AQ and O₂ determined the ORR rate in TPT-AQ (41).

On the other hand, the products of WOR were determined by the rotating ring-disk electrode (RRDE) tests (SI Appendix, Fig. S31) (7). TPT-alkynyl-AQ TPT-AQ and TPT-imine-AQ exhibited significant reduction currents when the applied potentials at the Pt ring electrode were −0.23 V and +0.6V, which were attributed to the reduction of O₂ and H₂O₂, respectively. Both of the reduction currents for O₂ and H₂O₂ were improved in TPT-alkynyl-AQ. These results indicated that the efficiency of water oxidation was increased through introducing alkynyl moieties. Meanwhile, H₂O₂ was photosynthesized in Ar atmosphere by TPT-alkynyl-AQ (296 µmol⋅g⁻¹⋅h⁻¹). By contrast, only a small amount of H₂O₂ was generated in TPT-AQ (14 µmol⋅g⁻¹⋅h⁻¹), which was consistent with the results in RRDE tests that the alkynyl moieties were crucial for WOR in TPT-alkynyl-AQ.

Furthermore, time-dependent DFT calculation revealed that the photoinduced hole of excited TPT-alkynyl-AQ was mainly located on the alkynyl and AQ moieties, which demonstrated that the alkynyl and AQ moieties were the WOR active sites (Fig. 5A). Moreover, other calculation results revealed that the alkynyl moieties were the dominant sites for the generation of H₂O₂ by adsorbing OH (OH*), which was a crucial intermediate step in 2e⁻ WOR (SI Appendix, Fig. S32). On the other hand, DFT calculation demonstrated that free hydroxyl radical (·OH) was another potential product at the AQ or triazine active sites for WOR, which was detected by 5,5-dimethyl-1-pyrroline-N-oxide capturing experiments (SI Appendix, Fig. S33).

Fig. 5.

The proposed mechanism of photocatalytic H₂O₂ production. (A) The analysis for the distribution of the hole (blue) and electron (green) for TPT-alkynyl-AQ through TD-DFT. (B) The proposed pathways of photosynthetic H₂O₂ production.

According to the above characterizations and analyses, the photosynthetic pathways of H₂O₂ by TPT-alkynyl-AQ were proposed in Fig. 5B. When the O₂ supply was adequate, the O₂ was first adsorbed on the alkynyl electron bridges. Then the photoinduced electrons quickly reduced the O₂ to generate H₂O₂, and the photoinduced holes were located in the alkynyl moieties to further oxidize the water to generate H₂O₂ (Fig. 5B, pathway 1). On the other hand, when the O₂ supply was inadequate, the photoinduced electrons were first stored in the AQ moieties, and the holes were located in alkynyl moieties. Then, the holes oxidized the water to generate H₂O₂, and the electrons stored in AQ moieties reacted with O₂ to produce H₂O₂ (Fig. 5B, pathway 2).

Wastewater Treatment.

In order to explore the potential of TPT-alkynyl-AQ for in situ river restoration, TPT-alkynyl-AQ was used for organic pollutant degradation under ambient conditions. The potential biological risk was investigated (SI Appendix, Fig. S34) (42). The model wastewater contained triclosan, 17α-ethinyl estradiol, and tetracycline. All of them were 10 mg⋅L⁻¹ in the model wastewater. The wastewater was treated by dispersing 10 mg of TPT-alkynyl-AQ in 500 mL of the wastewater for 10 h under ambient conditions (xenon lamp, λ > 400 nm, 100 mW⋅cm⁻¹, open air). Then, the treated water was mixed with the E3 nutrient solution. The normal hatching rate of zebrafish in the treated wastewater was 97%, while the normal hatching rate in nonpolluted E3 nutrient solution was 88%, and no zebrafish embryos were hatched in the pretreated wastewater after 3 d. The much higher hatching rate was attributed to the pollutant degradation and antibacterial properties of the in situ generated H₂O_2.

Discussion

In summary, we have reported a highly active and strong antiinterference D-B-A organic photocatalyst for H₂O₂ production in ambient conditions. The photosynthetic rates of H₂O₂ reach 9,366 and 12,324 µmol⋅g⁻¹ during 9 AM to 4 PM in river water and seawater, respectively. To the best of our knowledge, the photosynthetic efficiency of H₂O₂ is the highest under ambient conditions to date, even higher than all the systems consuming pure O₂. The outstanding efficiency of H₂O₂ production is owing to the spontaneous exciton dissociation at room temperature. Moreover, the rigid electron bridges and redox electron acceptors provide abundant active sites for both ORR and WOR to accelerate the overall rate of H₂O₂ production. The ultrahigh photocatalytic efficiency in ambient conditions allows SCC in a real cost-effective and sustainable way.

Materials and Methods

Synthesis of TPT-alkynyl-AQ.

Three milliliters of trifluoromethanesulfonic acid was charged into a predried flask under Ar atmosphere at 0 °C; 200 mg of 4-(2-{6-[2-(4-cyanophenyl)ethynyl]-9,10-dioxo-9,10dihydroanthracen-2-yl}ethynyl)benzonitrile in 20 mL of CHCl₃ was added into the flask dropwise over 30 min and stirred for 2 h. Then the suspension was warmed up to 25 °C and stirred for another 2 h before being left overnight at 100 °C. The obtained solid was quenched in cold water and washed with excess diluted ammonia and water. After being vacuum dried at 100 °C, the finally obtained solid was ground into powder and subjected to ultrasonication for 36 h in water.

TPT-AQ was synthesized via the same procedure and the same conditions as TPT-alkynyl-AQ.

Synthesis of TPT-alkynyl-TPT

Two milliliters of trifluoromethanesulfonic acid was charged into a predried flask under Ar atmosphere at 0 °C; 200 mg of 4,4′-(ethyne-1,2-diyl)dibenzonitrile in 10 mL of CHCl₃ was added into the flask dropwise over 30 min and stirred for 2 h. Then the suspension was warmed up to 25 °C and stirred for another 2 h before being left for 2 h at 100 °C. The obtained solid was quenched in cold water and washed with excess diluted ammonia, CHCl₃, and water. After being vacuum dried at 100 °C, the finally obtained solid was ground into powder and subjected to ultrasonication for 36 h in water.

Synthesis of TPT-imine-AQ.

One hundred seventy-five milligrams (0.75 mmol) of 2,6-diamino-9,10-dihydroanthracene-9,10-dione, 197 mg (0.5 mmol) of 4-[bis(4-formylphenyl)-1,3,5-triazin-2-yl]benzaldehyde, N,N-dimethylformamide (DMF)/1,4-dioxane (6 mL, 2/1 by volume), and acetic acid solution (1 mL, 6 M) were charged into a predried flask. After being quickly frozen, the flask was sealed and degassed by three freeze–pump–thaw cycles. Subsequently, the reaction flask was slightly warmed to room temperature and then kept at 160 °C for 72 h. After being cooled to room temperature naturally, the brownish red precipitate was collected by filtration and washed by DMF, dichloromethane, and ethanol, sequentially. After being vacuum dried at 100 °C, the finally obtained solid was ground into powder and subjected to ultrasonication for 36 h in water.

Photocatalytic Experiments.

One milligram of photocatalyst was added to 50 mL of deionized water in a 100-mL beaker. The catalyst was dispersed by ultrasonication for 30 min and irradiated by using a Xe lamp (100 mW·cm², λ > 400 nm) under magnetic stirring.

SCC Efficiency Measurements.

The SCC efficiency was determined by the photocatalytic experiments using an AM 1.5 G solar simulator as the light source (100 mW·cm⁻²). The photoreaction was performed in pure deionized water (100 mL) with photocatalyst (100 mg) in a glass bottle. The SCC efficiency (η) was calculated by following equation:

$$\eta \left(\text{\%}\right)=\frac{\mathrm{\Delta }G{H}_2{O}_2\times n{H}_2{O}_2}{t_{ir}\times S_{ir}\times I_{AM}}\times 100\text{\%},$$

where ΔGH₂O₂ is the free energy for H₂O₂ generation (117 kJ·mol⁻¹), nH₂O₂ is the amount of H₂O₂ generated, and t_ir is the irradiation time, 3,600 s. The overall irradiation intensity (I_AM) of the AM1.5 global spectrum (300 nm to 2,500 nm) is 100 mW·cm⁻², and the irradiation area (S_ir) is 3.14 × 10⁻⁴ m².

Apparent Quantum Efficiency Analysis.

The photocatalytic reaction was carried out in pure deionized water (50 mL) with photocatalyst (50 mg) in a quartz tube. The tube was irradiated by an Xe lamp for 1 h with magnetic stirring. The number of incident photons (M) is calculated by the following equation:

$$\text{M}=\frac{\text{Eλ}}{\text{hc}}.$$

In the equation, E, λ, h, and c are the average intensity of irradiation, the wavelength of the irradiation, the Planck constant, and the speed of light, respectively. The quantum efficiency was calculated from the following equation:

$$\text{AQY}=\frac{2\times {\text{number} \text{of} \text{evolved} \text{H}}_2{\text{O}}_2 \text{molecules}}{\text{number} \text{of} \text{incident} \text{photons}} \times 100\%.$$

Determination of H₂O₂ Concentration.

The concentration of H₂O₂ was quantified by a TMB-H₂O₂-HRP enzymatic assay, and the horseradish peroxidase (HRP) could instantaneously catalyze the reaction between H₂O₂ and TMB,

$${\text{H}}_2{\text{O}}_2+\text{TMB}\stackrel{\text{HRP}}{\to }{\text{H}}_2\text{O}+\text{oxTMB}.$$

Preparation of 3,3′,5,5′-tetramethylbenzidine (TMB) solution was as follows: 0.015 g TMB was dissolved in 0.3 mL of DMSO, followed by adding 5 mL of glycerol and 45 mL of deionized water containing 0.02 g of ethylenediaminetetraacetic acid and 0.095 g of citric acid. Then the solution was filled to 500 mL with deionized water.

Preparation of HRP solution was as follows: 0.002 g of peroxidase (from horseradish) was dissolved in 10 mL of deionized water.

Determination of the calibration curve was as follows: TMB and HRP were added into the H₂O₂ solution with a known concentration. Note that HRP concentration should be kept at 12.5 µg·mL⁻¹. After 3 min, 10 µL of HCl was added and measured by UV-visible spectroscopy at 450 nm. Based on the liner relationship between signal intensity and H₂O₂ concentration, the H₂O₂ concentration of the samples could be obtained.

Verification of the ORR Sites at the Alkynyl Moieties.

Photocatalyst (1 mg) and disodium ethylenediaminetetraacetate (2 mM) were added to 50 mL of deionized water in a 100-mL beaker. The catalyst was dispersed by ultrasonication for 30 min. The suspension was bubbled with O₂ for 30 min to make alkynyl moieties absorb oxygen and was irradiated under Ar atmosphere for 1 h.

Verification of the ORR Sites at the AQ Moieties.

Ten milligrams of photocatalyst was added to 20 mL of deionized water in a beaker. The catalyst was dispersed by ultrasonication for 30 min. The suspension was first irradiated under Ar atmosphere for 1 h and then injected with O₂ in the dark for 30 min.

TAS Measurements.

TA spectra of TPT-alkynyl-AQ, TPT-AQ, and TPT-alkynyl-TPT were measured on a Helios femtosecond transient absorption spectrometer (Ultrafast Systems, LLC). A 375-nm pump pulse was generated via an optical parametric amplifier (OPerA Solo, Coherent). The 375-nm laser intensity was 70 µW. The sample preparation was as follows: 20 mg of photocatalyst was dispersed in 3 mL of deionized water. Then the suspension was subjected to ultrasonication for 36 h. The initial suspension was centrifuged at 3,000 rpm for 10 min to remove large aggregates. The suspension was bubbled with oxygen for 20 min before atmosphere measurements.

Computational Methods.

The DFT calculation was carried out as implemented in the Gaussian 09 D.01 program package (43), using Grimme-D3 dispersion correction (44). GaussView6 was applied to visualization (45). Geometry optimization and frequency analysis were calculated at the PBE0/6-31G(g,d) level of theory (46). The optimized structures were used for single-point energy calculations with PBE0/ma-TZVP (46–48). Time-dependent density functional theory (TD-DFT) was calculated at CAM-B3LYP/6-31G(d,p) level of theory and was used to investigate the transfer direction of electrons (49–51). Multiwfn was used for hole–electron analysis (52, 53). The Gibbs free energy change (ΔG) during water oxidation on the surface conjugated polymers was calculated based on a computational hydrogen model (54).

Zebrafish Embryos Experiments.

Two fish culture media were prepared by adding the synthetic polluted water and the treated effluent to the standard E3 medium at a volume ratio of 1:2. The pure standard E3 medium was used as a blank control, and every group had 96 zebrafish zygotes for experiments. E3 medium includes 5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, and 0.33 mM MgSO₄. We used ethinylestradiol, triclosan, and tetracycline as the model water pollutants. The synthetic polluted water was obtained by adding 5 mg of ethinylestradiol, 5 mg of triclosan, and 5 mg of tetracycline to 100 mL of deionized water, which was sonicated for 6 h, filtered by water filtration membrane, and filled to 500 mL with deionized water. Ten milligrams of TPT-alkynyl-AQ catalyst was added into 100 mL of solution containing 5 mg of ethinylestradiol, 5 mg of triclosan, and 5 mg of tetracycline for sonication for 6 h; then the above solution was irradiated for 10 h, filtered by water filtration membrane, and filled to 500 mL with deionized water, and named the treated effluent. Zebrafish zygotes were obtained from wild-type adults and cultured in standard E3 medium for 7 d. Dead and unfertilized eggs were discarded at 4 h postfertilization. The fertilized embryos were transferred into 48-well plates and filled with corresponding medium. Embryo development was monitored, and representative images were captured.

Data Availability

All study data are included in the article and/or SI Appendix.