Aqueous Synthesis of Covalent Organic Frameworks as Photocatalysts for Hydrogen Peroxide Production

Open AccessCCS ChemistryRESEARCH ARTICLE7 Dec 2022Aqueous Synthesis of Covalent Organic Frameworks as Photocatalysts for Hydrogen Peroxide Production Fanglin Tan, Yuanyuan Zheng, Zhipeng Zhou, Honglei Wang, Xin Dong, Jing Yang, Zhaowei Ou, Haoyuan Qi, Wei Liu, Zhikun Zheng and Xudong Chen Fanglin Tan Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Yuanyuan Zheng Key Laboratory of Low-Carbon Chemistry & Energy Conservation of Guangdong Province, Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Materials Science and Engineering, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Zhipeng Zhou Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Honglei Wang Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Xin Dong Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Jing Yang Key Laboratory of Low-Carbon Chemistry & Energy Conservation of Guangdong Province, Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Materials Science and Engineering, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Zhaowei Ou Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Haoyuan Qi Center for Advancing Electronics Dresden (cfaed) and Faculty of Chemistry and Food Chemistry, Technical University of Dresden, Dresden 01069 Central Facility of Electron Microscopy, Electron Microscopy Group of Materials Science, Universität Ulm, Ulm 89081 Google Scholar More articles by this author , Wei Liu Key Laboratory of Low-Carbon Chemistry & Energy Conservation of Guangdong Province, Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Materials Science and Engineering, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author , Zhikun Zheng *Corresponding author: E-mail Address: [email protected] Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, State Key Laboratory of Optoelectronic Materials and Technologies, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author and Xudong Chen Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, School of Chemistry, Sun Yat-sen University, Guangzhou 510275 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.022.202101578 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Covalent organic frameworks (COFs) are crystalline porous polymers with designable structures and properties. Their crystallization typically relies on trial-and-error experimentation involving harsh conditions, including organic solvents, presenting significant obstacles for rational design and large-scale production. Herein, we present a liquid crystal-directed synthesis methodology and its implementation for up to gram-scale production of highly crystalline COFs in water and air. It is compatible with monomers of different structures, shape, size, length of side chains, and electron-donating, electron-accepting, and heterocyclic substitutions near reactive sites. Seventeen types of donor–acceptor two-dimensional COFs including four types of new ones and a three-dimensional COF with a yield of up to 94% were demonstrated, showing great generality of the method. The as-synthesized donor–acceptor COFs are organic semiconductors and contain macropores besides intrinsic mesopores which make them attractive catalysts. The production of H2O2 under visible light in water was studied and the structure–property relationships were revealed. The production rate reached 4347 μmol h−1 gcat−1, which is about 467% better than that of the benchmark photocatalyst g-C3N4. This study will inspire the mild synthesis and scale-up of a wide spectrum of COFs and organic semiconductors as efficient catalysts, promote their structure–property investigation, and boost their applications. Download figure Download PowerPoint Introduction Covalent organic frameworks (COFs) are crystalline porous polymers constructed by organic monomers termed as knots and linkers from light elements (such as carbon, nitrogen, oxygen, and hydrogen) through strong covalent linkages.1 They have attracted widespread interest in adsorption/desorption,2 separation,3 energy storage,4 catalysis,5 and sensing6 due to designable structures, permanent porosity, high surface areas, light weight, and high thermal and chemical stability.7 Different methods have been developed for the production of COFs, such as the solvothermal method,8 microwave thermal method,9 ionic thermal method,10 mechanical grinding,11 light/electrically promoted-, and plasma-induced synthesis.12–14 Unfortunately, the crystallization of COFs with these methods typically relies on trial-and-error involving harsh conditions, such as high temperature, high pressure, flammable and toxic organic solvents, and inert or vacuum atmosphere.15 Even if a crystallization condition was determined for a specific structure, exhaustive refinements are required upon minor changes of monomer structures in the synthesis of COFs,16 presenting significant obstacles to the rational design of their structures and functionalities as well as large-scale production. A simple, general, and scale-up strategy thus remains to be developed. Herein, we present a liquid crystal (LC)-directed synthesis methodology for the production of highly crystalline COFs with macropores besides intrinsic mesopores in water under air. The method is compatible with monomers of different structures, size, shape, length of side chains, and electron-donating, electron-accepting as well as heterocyclic substitutions near reactive sites even under the same crystallization condition. COFs with different topologies and dimensions were synthesized up to gram-scale amounts with isolated yields of 81–94%. We then explored 17 donor–acceptor COFs to catalyze the reduction of oxygen to produce hydrogen peroxide (H2O2) in water under visible light and revealed the structure–property relationships. The production rates of H2O2 ranged from 580 to 2406 μmol h−1 gcat−1 and from 667 to 4347 μmol h−1 gcat−1 without and with the existence of conventional sacrificial ethanol, respectively. Experimental Methods Synthesis of COFs in liquid crystal The preparation of liquid crystals followed literature reports.17,18 For the synthesis of COFs, 0.1 mmol of monomer with an amine group and 0.15 or 0.2 mmol monomer with an aldehyde group ( Supporting Information Table S1) were added to a 10-mL bottle containing 2 g of Triton X-100/n-decyl alcohol/water liquid crystal. A magnetic stirrer was used to uniformly mix the solution, and then kept at room temperature for 1 h. Afterward, 0.3 mmol of catalyst was added and stirred evenly. The reaction remained undisturbed for 7 days at room temperature. COFs were obtained after subsequent washings with Milli-Q water, ethanol, and anhydrous tetrahydrofuran, and dried in vacuum at 80 °C for 12 h. Preparation and characterization of carbon nitrides (g-C3N4) Melamine (8 g) was heated to 550 °C for 4 h at a rate of 12 °C min−1 in a muffle furnace in an air atmosphere.19 Yellow g-C3N4 powder was obtained after cooling to room temperature. Photocatalytic reduction of oxygen to H2O2 Typically, 15 mg of as-synthesized COFs powder was dispersed in 50 mL Milli-Q water or an EtOH/H2O (EtOH 5 mL and H2O 45 mL) solution in a Schlenk flask (100 mL). The suspension solutions were stirred for 20 min in the dark with continuous O2 bubbling after ultrasound treatment for 10 min to reach the absorption–desorption equilibrium. The reaction temperature was kept at 25 °C (298 K) with a circulating water bath. A 50 W light-emitting diode (LED) lamp (∼420 nm, average intensity: 100 mW·cm−2) was used as the light source. The concentration of H2O2 was determined by UV spectrophotometer after removal of the photocatalyst from the centrifuge at 10,000 rpm followed by filtration with a 0.1 μm filter in the dark. Results and Discussion We explored the synthesis of COF-TTA-DHTA with 2,4,6-tris(4-aminophenyl)-1,3,5-triazine (TTA, knot) and 2,5-dihydroxy-1,4-benzenedicarboxaldehyde (DHTA, linker) monomers to exhibit the synthetic procedure of the LC-directed methodology and determine the crystallization conditions (Figure 1a). The synthesis procedure is illustrated in Figure 1a. Triton X-100, n-decyl alcohol, and water were mixed to form LC according to the phase diagram and achieve phase equilibrium (Figure 1b). Subsequently, monomers and catalysts were added into the LC and mixed uniformly. The mixture was then placed at room temperature to produce COF-TTA-DHTA. To understand the effects of different parameters on the crystallinity of the COF-TTA-DHTA at room temperature under ambient conditions, the ratio of different liquids in the LC, catalyst, and reaction time were investigated with powder X-ray diffraction (PXRD) (Figure 1c and Supporting Information Table S2). Only amorphous or low crystallinity COF-TTA-DHTA was obtained out of the lamellar LC phase region. Within the lamellar LC phase region (blue part, Figure 1b), highly crystalline (full width at half maximum of 100 peak not more than 0.6) COF-TTA-DHTA was achieved, and highest crystallinity was achieved when the ratio of Triton X-100, n-decyl alcohol, and water were 25%, 10%, and 65% in the LC, respectively. When acetic acid (AcOH), trifluoroacetic acid (TFA), trifluoromethanesulfonic acid (TfOH), and p-toluenesulfonic acid (PTSA) were used as catalysts, the crystallinity of COF-TTA-DHTA increased in the sequence of TfOH < AcOH < TFA < PTSA (conditions p, q, r, and d in Figure 1c and Supporting Information Table S2, respectively). Then the molar amount of PTSA was investigated, and it was found that the crystallinity of COF-TTA-DHTA increased with an increase of the molar amount of PTSA and reached a maximum at 3 equiv relative to the amine monomer (condition d), where further increases in concentration led to no significant change in crystallinity. The possible reason is that an appreciable amount of acid can protonate the amino groups of the monomers, increase the reversibility of the Schiff-base reaction, facilitate the error-correction mechanism, and promote the crystallization of the COFs.20 Finally, reaction time was studied. The (100) peak of the PXRD patterns of COF-TTA-DHTA appeared within 30 min and the relative intensity continued to increase, where more peaks emerged with time, and no obvious increase in intensity was observed after 7 days ( Supporting Information Figure S1). We then chose LC with 25% Triton X-100, 10% n-decyl alcohol, and 65% water, 3 equiv. of PTSA relative to amine monomer, room temperature, and a reaction time of 7 days as the general crystallization conditions (condition d) for the production of COFs. COF-TTA-DHTA was obtained after thorough rinsing and subsequent filtering with water, ethanol, and tetrahydrofuran with an isolated yield of 92% ( Supporting Information Table S1). There were no superpositions of characteristic peaks in the solid-state 13C nuclear magnetic resonance (NMR) spectra of COF-TTA-DHTA with 13C NMR spectra from either the monomers or LC, indicating the starting materials were fully removed ( Supporting Information Figure S2). The high purity of the COF-TTA-DHTA was also confirmed with Fourier transform infrared (FT-IR) spectroscopy ( Supporting Information Figure S2). The sharp and multiple PXRD patterns indicated that COF-TTA-DHTA was highly crystalline. It adopted an AA-eclipsed stacking structure from which the simulated PXRD patterns agreed well with the experimental ones (Figure 1d and Supporting Information Figure S3). Figure 1 | Synthesis illustration of COFs by LC-directed method taking COF-TTA-DHTA as an example. (a) The chemical structure of COF-TTA-DHTA and monomers and schematic of the synthesis procedure of the LC-directed strategy. (b) Phase diagrams and proportions of Triton X-100/n-C10H21OH/H2O liquid crystal. Inset image shows the chemical structure of Triton X-100. (c) Comparison of PXRD patterns of COF-TTA-DHTA under different ratios of Triton X-100 [(i) a–j], different ratios of n-decyl alcohol [(ii) d, f, and k–o], different catalysts [(iii) d, p, q, and r: PTSA, AcOH, TFA, TfOH], and amount of PTSA added [(iv) d, s, t, u, and v: 0.3, 0.1, 0.2, 0.4, and 0.6 mmol]. (d) PXRD spectra of synthesized COF-TTA-DHTA and simulated PXRD pattern of COF-TTA-DHTA model (experimental: dark blue dot; Pawley refined: red; simulated: purple; difference: green) (e) SEM images and camera images of COF-TTA-DHTA. Download figure Download PowerPoint Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) showed that COF-TTA-DHTA powder contained macropores in the range of tens to hundreds of nanometers, and was composed of entangled and interwoven nanoribbon crystals with a length and width of ∼500 and 50 nm, respectively (Figure 1e and Supporting Information Figure S4). The synthesis could be easily scaled up to gram-scale level. No notable changes in the morphology, size, or crystallinity were observed in scale-up reactions (22-fold, Supporting Information Figure S5). This LC-directed synthesis methodology shows general applicability. We chose C2 and C3 symmetric knots with different cores [triazine for TTA, triphenyl for 1,3,5-tris(4-aminophenyl)benzene (TPB), and tetraphenylpyrene for 1,3,6,8-tetra(4-aminopheny)pyrene (TPy)] and C2 symmetric linkers with parent core [terephthaldicarbozaldehyde (TP)], electron-donating substitution [2,5-dihydroxy-1,4-benzenedicarboxaldehyde (2,3-DHTA) besides DHTA], electron-accepting substitution [2,3,5,6-tetrafluoroterephthalaldehyde (TFTA)], and different core lengths [biphenyl-4,4′-dicarboxyaldehyde (BDA) and 2,2′-bipyridine-5,5′-dicarboxaldehyde (BPDA)], as well as C3 symmetric linkers 1,3,5-benzenetricarboxaldehyde (TA) and 2,4,6-tris(4-formylphenyl)-1,3,5-triazine (TTTA) to gain COF-TA-TP, COF-TTA-2,3-DHTA, COF-TTA-TFTA, COF-TTA-BDA, COF-TTA-BPDA, COF-TPB-TP, COF-TPB-DHTA, COF-TTA-TA, COF-TTA-TTTA, COF-TPy-DHTA, COF-TPy-BDA, and COF-TPy-BPDA using condition d (Figure 2a and Supporting Information Figure S6). All COFs were highly crystalline despite the different cores, side chains, substitutions, and lengths of the linker notably affecting the reaction activity of aldehyde/amine groups and the crystallization process (Figure 2a). Note that all the COFs were obtained under the same crystallization conditions, which was unprecedented for crystallization of COFs as well as organic polymers in general. Beside intrinsic mesopores, the COFs also had macropores with in the range of tens to hundreds of nanometers due to entangled and interwoven nanoribbons ( Supporting Information Figures S7 and S8). The isolated yields of the COFs were 86–94% ( Supporting Information Table S1), which were higher than those of traditional microwave and solvothermal synthesis methods (66–86%).8,21–27 The formation of imine-linked COFs was further confirmed by solid-state 13C NMR spectrum and FT-IR spectroscopy ( Supporting Information Figures S9 and S10 and Table S3). The experimental PXRD patterns and the Pawley refinement confirmed these two-dimensional (2D) COFs adopted an AA-eclipsed stacking structure ( Supporting Information Figure S11 and Table S4).8,21–27 Thermogravimetric analysis (TGA) showed that the TTA- and TPB-based 2D COFs exhibited thermal stability higher than 400 °C, and TPy-based TPy-BDA and TPy-BPDA higher than 500 °C, pointing out the high thermal stability of the synthesized COFs ( Supporting Information Figure S12). The Brunauer–Emmett–Teller (BET) surface area of the 2D COFs ranged from ∼1076 to 2246 m2 g−1 as determined by N2 adsorption behavior at 77 K ( Supporting Information Figure S13 and Table S5). These values were comparable to COFs synthesized by traditional synthesis methods.8,21–31 The pore sizes of the COFs calculated by the nonlocal density functional theory (NLDFT) model were consistent with the theoretical values ( Supporting Information Figure S14 and Table S5). Finally, the high reproducibility of the LC-directed synthesis method for the crystallization and synthesis efficiency of COFs was further demonstrated with gram-level production of COF-TTA-TFTA and COF-TTA-BPDA under condition d ( Supporting Information Figure S15). Figure 2 | (a) General synthesis of C2 + C3, C2 + C2, and C3 + C3 construction mode TTA-, TPB- and TPy-based 2D COFs in LC. (b) Comparison of the experimental (dark blue), Pawley refined (red), simulated PXRD patterns (purple), Pawley refinement difference (green), and N2 adsorption isotherms of COF-TTA-DMTA, COF-TTA-2-NTA, COF-TTA-PyTA, and COF-TTA-DMEA at 77 K. Inset image shows the model of these COFs. Download figure Download PowerPoint We further explored the LC-directed synthesis methodology to create new COFs with TTA (knot) and TP linker with electro-donating [2,5-dimethyl-1,4-benzenedicarboxaldehyde (DMTA)], -accepting [2-nitroterephthalaldehyde (2-NTA)], heterocyclic substitutions [2,5-pyridine dialdehyde (PyTA)] near aldehyde groups, and long side chains [2,5-di(methyl glycol)-benzene-1,4-dicarbaldehyde (DMEA)] under condition d (Figure 2a and Supporting Information Figure S16). The isolated yields of COF-TTA-DMTA, COF-TTA-2-NTA, COF-TTA-PyTA, and COF-TTA-DMEA were 93%, 90%, 81%, and 82%, respectively ( Supporting Information Table S1). The characteristic peak of the imine carbon (C=N) at ∼159 ppm (parts per million, chemical shift) in solid-state 13C NMR spectra and the obvious peaks of the FT-IR of the COFs at ∼1606 cm−1 revealed the formation of the imine linkage ( Supporting Information Figures S9 and S10 and Table S3). The multiple, sharp, and intense PXRD patterns indicated the synthesized COFs were highly crystalline. The PXRD patterns of the COF-TTA-DMTA, COF-TTA-2-NTA, COF-TTA-PyTA, and COF-TTA-DMEA showed peaks with high intensities at ∼2.7° (±0.1, 2θ) and minor peaks at ∼4.9°, 5.7°, 7.5°, and 9.9° (±0.1, 2θ), which could be attributed to the diffractions of (100), (110), (200), (210), and (310) planes, respectively (Figure 2b). Similarly, broad peaks at higher 2θ values (∼25.5° ±0.3) signified the π–π stacking arising from (001) plane. The experimental PXRD pattern of these COFs also agreed with the simulated AA-eclipsed stacking structure model ( Supporting Information Figure S11), and Pawley refinement using the corresponding structural model provided a very good fit to the experimental data (with Rwp and Rp values converged to ∼3% and 2%, respectively). The produced lattice parameters (a = b = 35.5 Å, c = 3.5 Å, ±1.0 Å, α = β = 90°, γ = 120°) were similar to the expected crystal structure model ( Supporting Information Table S4). The TGA showed that the decomposition temperatures of these newly synthesized COFs were higher than 400 °C, indicating high thermal stability ( Supporting Information Figure S12). We analyzed the pore stability and permanent porosity of these COFs by N2 adsorption at 77 K. The BET surface area of these COFs were 2408 m2 g−1 (COF-TTA-DMTA), 1279 m2 g−1 (COF-TTA-2-NTA), 1427 m2 g−1 (COF-TTA-PyTA), and 1431 m2 g−1 (COF-TTA-DMEA), respectively (Figure 2b and Supporting Information Table S5). SEM also showed that these COFs showed entangled and interwoven nanoribbon morphologies with macropores and intrinsic mesopores ( Supporting Information Figures S7 and S8). The NLDFT pore sizes of the mesopores were 2.9–3.2 nm, which were fully consistent with those calculated from the crystal structure ( Supporting Information Figure S14 and Table S5). We further chose COF-300 as an illustrative example to extend LC-directed synthesis methodology to synthesize three-dimensional (3D) COFs. Under the general crystallization condition d, tetra(4-aminophenyl)methane condensed with TP to afford COF-300 rods with single crystalline domains of up to about 2 μm × 500 nm and an isolated yield of 82% (Figures 3a–3c). The experimental PXRD patterns agreed with the simulated sevenfold interpenetration (dia-c7) model of COF-300 (Figure 3d).32 Pawley refinement provided the lattice parameters (a = b = 19.52 Å, c = 8.89 Å, α = β = γ = 90°) with satisfactory convergence (with Rwp = 8.75% and Rp = 6.30%). The characteristic peak of the solid-state 13C NMR spectra at 159 ppm and FT-IR spectrum at 1620 cm−1 indicated the formation of C=N ( Supporting Information Figure S17).32 By analyzing the N2 adsorption behavior at 77 K, the BET surface area of COF-300 was 245 m2 g−1, and the NLDFT pore size calculation was consistent with literature reports.28 The COF-300 synthesized in LC had less overall N2 uptake, which might be due to its larger crystal size and thus low accessibility of pores ( Supporting Information Figure S17).33 Figure 3 | (a) The chemical structures of COF-300 and monomers. (b) Optical microscope and (c) SEM images of COF-300 synthesized by LC-directed method. (d) Comparison of the experimental (dark blue), Pawley refined (red), simulated PXRD patterns (purple), and Pawley refinement difference (green) for COF-300. Download figure Download PowerPoint The reason behind the formation of highly crystalline 2D COFs in ribbon structures and COF-300 in rod with the LC-directed synthesis methodology should be discussed. The high crystallinity of the COFs might relate to the strong confinement of the monomers as well as their molecular precursors by the LC, which prevents their rapid agglomeration and diffusion among Triton X-100, n-decyl alcohol, and water and slow the polymerization as well as crystallization rate, leaving enough time and space for the ordering of the final products.17,20,34 In the crystallization process, the lamellar phase of the LC facilitates the formation of strips. For 2D COFs, the interaction force in the out-of-plane direction is weaker than that of the in-plane direction, and it decreases with an increase in the thickness of the crystals and of the mismatch generated during their growth,35 thereby reducing the growth trend of the out-of-plane direction and leading to the formation of ribbons. They form entangled and interwoven structures with macropores upon the removal of the LC molecules. For 3D COFs, isotropic growth in three dimensions leads to rods. The synthesized 2D COFs contain photosensitive units such as triphenyls, triazine, and pyrene. They are highly crystalline and can be easily well dispersed in water, and therefore have great potential in photocatalysis. The diffuse reflectance UV–vis spectroscopy (DRS) of the 2D COFs exhibited strong absorption in the visible light region (λ > 400 nm) ( Supporting Information Figure S18). Their optical band (Eg) gaps calculated by the Tauc Plot method were between 2.04 and 2.70 eV ( Supporting Information Figure S18), which made them very suitable as photocatalysts and for photocatalytic H2O2 production.36 The valence band (VB) spectrum of the X-ray photoelectron spectroscopy was explored to determine the VB edges, which amounted to ∼1.47–2.50 eV [vs normal hydrogen electrode (NHE)] ( Supporting Information Figure S19 and Table S6). The conduction band (CB; CB = VB − Eg) value of these 2D COFs were then obtained as ∼0.91–0.12 eV (vs NHE). The VB and CB values of the 2D COFs were more positive and negative than the redox potential of O2/H2O2 (1.23 V vs NHE) and O2/H2O2 (0.68 V vs NHE), indicating that they were promising photocatalysts to reduce oxygen to produce H2O2 and oxidize water to produce oxygen and protons to generate H2O2 under light irradiation, respectively (Figure 4a).37,38 Figure 4 | (a) Band potential alignments of all synthesized 2D COFs in this work vs NHE. (b) The average production rates of H2O2 per hour and per gram for all synthesized 2D COFs nanoribbons and g-C3N4 within 8 h of reaction time (cyan-blue: EtOH used as sacrificial reagent, magenta: no sacrificial reagent). Download figure Download PowerPoint The photocatalytic H2O2 production reaction was carried out in a 100-mL Schlenk flask containing Milli-Q water or a mixed solution of ethanol and water after O2 bubbling with common blue LEDs (420 nm, 50 W, average intensity: 100 mW·cm−2) as the visible light source. Under radiation of the blue LED, 15 mg of 2D COF was used as the photocatalyst for an 8 h reaction at room temperature. The concentration of the produced H2O2 was determined by cerium sulfate [Ce(SO4)2] titration method.39 After 8 h of photocatalytic reaction, the overall production rates of these 2D COFs for photocatalytic production of H2O2 ranged from ∼580 to 2406 μmol h−1 gcat−1 (Figure 4b and Supporting Information Figures S20 and S21). When ethanol was added as the sacrificial agent to provide electrons and protons ( Supporting Information Figure S22), which is common in the production of H2O2, the production rates were greatly improved since ethanol could be more easily oxidized by holes than water [E (O2/H2O) = + 1.23 eV and E (CH3CHO/CH3CH2OH) = + 0.21 eV vs NHE].40 The structure of COFs determines their VB position. The higher the VB position, the stronger the oxidizing ability of the generated holes, and the higher the hydrogen peroxide production rate. The production rates varied from ∼667 to 4347 μmol h−1 gcat−1 (Figure 4b and Supporting Information Figures S20 and S21), which were dozens of times higher than those of organic porous materials (COFs and covalent triazine frameworks) ( Supporting Information Table S7).37,40 In addition, the production rates for photocatalytic hydrogen peroxide production of these 2D COFs were higher than most organic/inorganic materials and their composites reported thus far ( Supporting Information Table S7).41–48 Compared with metal-assisted (Au, TiO2, etc.) catalysts,41–48 these COFs provide a completely metal-free system. The high photocatalytic performance could be due to both the donor–acceptor structures and high crystallinity of the synthesized COFs, which were beneficial to the transport of carriers. The entangled and interwoven macroporous structure could also play a significant role, since it could expose more active sites and facilitate mass transfer.49,50 To verify this assumption, we synthesized bulky COF-TTA-DHTA and COF-TTA-TTTA by a solvothermal method for comparison. In comparison with those produced by the LC-directed synthesis method, these COFs could not form the prerequisite entangled and interwoven macroporous structure. The photocatalytic production rates of H2O2 with the bulky COF-TTA-DHTA and COF-TTA-TTTA decreased by about 35% and 27%, respectively ( Supporting Information Figures S23 and S24). Since reaction conditions (such as light intensity, radiation wavelength, sacrificial agent, and reaction time) might vary among labs, we synthesized g-C3N4, a benchmark photocatalyst, by referring to a method reported in the literature for comparison. We found that g-C3N4 has no photocatalytic activity without the use of sacrificial reagent (ethanol) under the same catalytic conditions as the 2D COFs, which is also consistent with literature reports.51 The H2O2 production rates of g-C3N4 with ethanol as the sacrificial reagent was 767 μmol h−1 gcat−1. The H2O2 production rates of a majority of these 2D COFs were approximately four to six t