Han
Wang†
a,
Hui
Wang†
b,
Ziwei
Wang†
a,
Lin
Tang†
a,
Guangming
Zeng
*a,
Piao
Xu
*a,
Ming
Chen
a,
Ting
Xiong
a,
Chengyun
Zhou
a,
Xiyi
Li
b,
Danlian
Huang
a,
Yuan
Zhu
a,
Zixuan
Wang
a and
Junwang
Tang
*b
aCollege of Environmental Science and Engineering, Hunan University and Key Laboratory of Environmental Biology and Pollution Control (Hunan University), Ministry of Education, Changsha 410082, P. R. China. E-mail: zgming@hnu.edu.cn; piaoxu@hnu.edu.cn
bDepartment of Chemical Engineering, University College London, Torrington Place, London, WC1E7JE, UK. E-mail: junwang.tang@ucl.ac.uk
First published on 18th May 2020
In the light of increasing energy demand and environmental pollution, it is urgently required to find a clean and renewable energy source. In these years, photocatalysis that uses solar energy for either fuel production, such as hydrogen evolution and hydrocarbon production, or environmental pollutant degradation, has shown great potential to achieve this goal. Among the various photocatalysts, covalent organic frameworks (COFs) are very attractive due to their excellent structural regularity, robust framework, inherent porosity and good activity. Thus, many studies have been carried out to investigate the photocatalytic performance of COFs and COF-based photocatalysts. In this critical review, the recent progress and advances of COF photocatalysts are thoroughly presented. Furthermore, diverse linkers between COF building blocks such as boron-containing connections and nitrogen-containing connections are summarised and compared. The morphologies of COFs and several commonly used strategies pertaining to photocatalytic activity are also discussed. Following this, the applications of COF-based photocatalysts are detailed including photocatalytic hydrogen evolution, CO2 conversion and degradation of environmental contaminants. Finally, a summary and perspective on the opportunities and challenges for the future development of COF and COF-based photocatalysts are given.
In the 1970s, Fujishima and Honda realized water splitting under ultraviolet (UV) radiation by using a titanium dioxide (TiO2) electrode for the first time.4 And Carey et al. carried out the photocatalytic degradation of organic pollutants with TiO2 in aqueous suspensions four years later.5 These have sparked intense interest among researchers in artificial photosynthesis. Traditional inorganic semiconductor photocatalysts such as TiO2,6–8 cadmium sulphide (CdS),9–11 zinc oxide (ZnO)12,13 and silver phosphate (Ag3PO4)14 have occupied a leading position over the past several decades. Among them, TiO2 is the most important and well-known photocatalyst due to its low cost, relatively high availability and durability. However, its wide band gap of 3.2 eV that only allows for ultraviolet light absorption limits its utilization of the solar spectrum, leading to low photocatalytic efficiency and photocurrent quantum yield.15 Besides, Ag3PO4, CdS and other transition metal sulfides and oxides with a suitable band gap to absorb visible light and with good carrier transportation capacity have stimulated the attention on photocatalytic studies, whereas the heavy metal toxicity and photo-corrosion effect block their practical applications.9 As the research progressed, organic semiconductors like graphitic carbon nitride (g-C3N4),16,17 metal–organic frameworks (MOFs),18,19 and covalent organic frameworks (COFs)20–24 have been used as the photocatalyst and show promising performance towards solar energy conversion. g-C3N4 as a metal-free polymer possesses many fascinating features including “earth-abundant” nature, high physicochemical stability and favorable band gap structure. However, there are drawbacks: its synthesis is often conducted at high temperature (>500 °C) and its molecular backbone consists of either triazine or heptazine units, leading to limited structural diversity.25 As a type of porous crystalline materials, MOFs constructed from organic linkers and transition-metal nodes are attractive due to their large surface area, structural tailorability and easy pore functionalization. Unfortunately, most MOFs are unstable and can easily deteriorate under humid conditions which limits their repeated use.19
Covalent organic frameworks (COFs), as newly developed organic polymers, have caused ripples of excitement among researchers striving to exploit their promising photocatalytic potential. COFs with low density are crystalline porous materials composed of organic molecules linked by covalent bonds through reticular chemistry, and they have been widely used in areas such as heterogeneous catalysis,26–29 gas storage and separation,30,31 energy storage and optoelectronic devices.32–34 Compared with traditional semiconductors, COFs possess not only some common features but also many special advantages pertaining to photocatalysis: (i) the structural designability of COFs enables them to realize the design of targeted structures and special properties related to photocatalytic reactions such as excellent visible-light absorption, and fast electron–hole separation and transfer; (ii) the large surface area of COFs enriches accessible catalytic sites, and the highly crystalline and porous structure endows COFs with accelerated charge transport to the surface and decreases the possibility of charge trapping caused by defects, thus contributing to suppressed electron–hole recombination; (iii) COFs with strong covalent bonds show high chemical and thermal stability, and photoactive units fixed in the robust framework can avoid photo-corrosion and enhance the lifetime of the excited states; and (iv) the extended π-conjugated structure both in-plane and in the stacking direction enables high charge carrier mobility. These fascinating inherent features endow COFs with great potential in photocatalytic energy conversion and environmental remediation, and they are deemed to match or even exceed MOFs and conventional photocatalytic semiconductors in performance. Lotsch and co-workers first reported the discovery of a COF-based photocatalyst.35 A high visible-light-induced hydrogen production efficiency has been achieved based on hydrazine-based TFPT-COF (evolution rate: 1970 μmol h−1 g−1, triethanolamine (TEOA) as a sacrificial donor), which was competitive with other representative photocatalysts including Pt-modified amorphous melon (720 μmol h−1 g−1), g-C3N4 synthesized at 600 °C (840 μmol h−1 g−1),36 and crystalline poly(triazine imide) (864 μmol h−1 g−1).37 This success has initiated the exploration of COF-based photocatalysts in the whole community (Fig. 1).
The number of publications in the area of COF-based photocatalysts has increased sharply, and a comprehensive review of COF photocatalysts is needed. In this review, we begin by summarizing different connections of COF building blocks including boron-containing connections, nitrogen-containing connections and double-stage connections combining imine linkages and boronate ester linkages. Subsequently, we compare the performance of COFs with different morphologies, such as 0-dimensional (0D) structures, 1-dimensional (1D) structures, 2-dimensional (2D) structures, and 3-dimensional (3D) structures. Strategies related to the enhanced photocatalytic performance of COF materials are then presented. Afterwards, the solar-driven application of COFs is discussed, including water splitting, CO2 conversion as well as photocatalytic degradation of pollutants in wastewater. Finally, a perspective on the challenges and opportunities in this area, including synthesis, functions and application, is discussed. Complementary to this review, the readers are also suggested to read another review about the design of COF if they are interested in the materials design.38–44
As a representative example, COF-1 was designed and fabricated through the self-condensation of 1,4-benzenediboronic acid (BDBA), which was based on molecular dehydration to form six-membered boroxine connections.45 The as-prepared COF-1 possessed a layered graphitic structure with a hexagonal pore diameter of 15 Å and a Brunauer–Emmett–Teller (BET) surface area of 711 m2 g−1. In this method, it is essential to keep the reaction under a closed condition for water equilibrium to guarantee reversibility of COF formation. Similarly, the same group further successfully constructed the first 3D COFs (COF-102 and COF-103) with the self-condensation of the tetrahedral molecular building block tetra(4-dihydroxyborylphenyl)methane (TBPM) or its silane analog (TBPS).46 The crystalline COF-102 and COF-103 exhibited a higher BET surface area of 3472 m2 g−1 and 4210 m2 g−1, respectively. Since then, this self-condensation strategy has been widely used to fabricate boron-containing COFs based on various monomers, such as biphenyldiboronic acid,47 pyrene-2,7-diboronic acid (PDA),48 and 4,4′-phenylazobenzoyl diboronic acid.49
Besides self-condensation, co-condensation of two or more building blocks such as boronic acids with catechols has also been reported. The dehydration condensation of 2,3,6,7,10,11-hexahydroxytriphenylene (HHTP) and BDBA resulted in the formation of layered COF-5 with five-membered BO2C2 rings, which exhibited an eclipsed boron nitride arrangement.45 It is worth mentioning that COF-5 has been widely regarded as a representative to examine various new synthesis strategies.50–52 Likewise, the first crystalline boronate-linked 3D COFs (COF-105 and COF-108) were obtained by replacing BDBA with tetrahedral molecules TBPM and TBPS, respectively.46 COFs with different properties and functions could be designed and synthesized by a diverse combination of building units. For instance, a novel photoactive donor–acceptor TP-Por COF was prepared based on triphenylene and porphyrin units.53 The resulting TP-Por COF film with enhanced charge separation showed broad optical absorption covering the entire visible range up to 680 nm. In a conventional condensation, donor–acceptor DTP–ANDI–COF with a large pore size of 5.3 nm was obtained from N,N′-di-(4-boronophenyl)naphthalene-1,4,5,8-tetracarboxylic acid diimide and HTTP.54 The charge-separation state lifetime of 2.5 μs was determined by time-resolved electron spin resonance spectroscopy, indicating the presence of long-lived radicals produced through effective charge transfer from the donor triphenylene to the acceptor naphthalene diimide. Notably, polyfunctional catechols are easily oxidized and are difficult to dissolve in most organic solvents, leading to difficulty in the fabrication of functional building blocks and related COFs. Thus, a new Lewis acid-catalyzed strategy protecting catechols from oxidation was put forward.55 A boronate ester-linked Pc–PBBA COF with a pore size of 2.3 nm was constructed from 1,4-phenylenebis(boronic acid) (PBBA) and phthalocyanine tetra(acetonide) (Pc) in the presence of Lewis-acid catalyst BF3·OEt2. The as-prepared eclipsed COF with broad absorbance showed great potential for effective charge transfer through stacked phthalocyanines. In contrast to the conventional condensation of two components, a multiple-component (MC) strategy was also studied.56 For example, a three-component [1+2] co-condensation was attempted by using the shortest unit BDBA and a longer molecule PDA as the linkers to react with HHTP as the knots. Two MC-COFs (termed MC-COF-TP-E11E27 and MC-COF-TP-E21E17) with slipped AA stacking were generated to possess a BET surface area of 1892 and 1534 m2 g−1 and a pore size of 3.2 and 2.9 nm, respectively. This co-condensation strategy could also be used to tailor the functionality of COFs. A highly emissive 2D COF TPE-Ph COF was designed by introducing an aggregation-induced emission active tetraphenylethene (TPE) unit to condense with TPE-cored boronic acids and 1,2,4,5-tetrahydroxybenzene.57 Considering that the boronate linkages in the TPE-Ph COF formed a Lewis acid–base pair when interacted with ammonia, the TPE-Ph COF could be used as a fluorescence sensor for ammonia.
Generally, COFs with boron-containing linkages possess low density and high surface area, leading to various applications.58,59 However, boroxines and boronate esters are easily hydrolysed and oxidized, which limits their application as catalysts or their long-term usage. Still, it is undeniable that boron-containing COFs are of particular importance for mechanistic studies.50,60,61
Despite the crystalline problems, high BET surface area, remarkable thermal and chemical stability, and controllable C/N/H composition endow CTFs with potential for catalysis. 2D CTFs with triazine subunits can be regarded as analogs of g-C3N4, which has been studied extensively as a photocatalyst.71,72 On the one hand, the incorporated nitrogen in the backbone benefits metal nanoparticle loading, which provides a platform for the introduction of active sites for the catalytic reaction. On the other hand, the tunable structures with unlimited organic subunits allow for the controllable band alignment and optimal light absorption.73 Studies demonstrated that the photocatalytic hydrogen production of CTF-1 can be varied with different reaction conditions. For example, a well-ordered CTF-1 was synthesized via a mild microwave-assisted condensation.74 An apparent quantum efficiency (AQE) of 3.8% and 6% at 420 nm for oxygen and hydrogen evolution under visible light irradiation was determined, respectively. In particular, the oxygen evolution rate and hydrogen evolution rate of CTF-1-100 W were 140 μmol g−1 h−1 and 5500 μmol g−1 h−1, respectively, both of which are higher than those of g-C3N4.71,75 The examples verified the promising properties and applications of triazine-linked COFs. The successful synthesis of crystalline CTFs on a large scale will be the focus of future research.
Various building blocks have been involved in imine-based COF formation.80–85 For instance, a highly conjugated π-electron porphyrin unit and its metal derivatives have been largely employed in the construction of functional imine-linked COFs. One study introduced two porphyrin-based COFs, termed COF-66 and COF-366, with the feature of extended planar π-conjugation.86 COF-66 and COF-366 were obtained from the solvothermal reaction of porphyrin and TA and tetrahydroxy anthracene, respectively, and the formed imine bond was characterized by FT-IR and 13C cross-polarization magic-angle spinning (CP-MAS) NMR spectroscopic techniques. Both COFs exhibited high charge carrier mobility owing to the close intermolecular π–π distances. A series of porphyrin COFs MP-DHPh COFs with varied H-bonding sites were synthesized via a three-component condensation strategy. Specifically, porphyrin derivatives (MP; M = H2, Cu, and Ni) were used to react with a mixture of TA and dihydroxyterephthalaldehyde (DHTA, H-bonding edges) at different molar ratios. As determined by ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS), H2P-DHPh COF, CuP-DHPh COF, and NiP-DHPh COF possessed narrower band gaps of 1.31, 1.36, and 1.54 eV compared to that of 1.36, 1.40, and 1.58 eV for the corresponding amorphous MP-Ph polymers, respectively. The H2P-DHPh COF displayed higher photocatalytic singlet oxygen evolution than the CuP-DHPh COF and NiP-DHPh COF, and the photocatalytic performance of COFs increased with the increasing content of the H-bonding site. More recently, a conjugated imine-linked metalloporphyrin COF was prepared through the Schiff-base reaction of Zn-5,10,15,20-tetrakis(4-aminophenyl)-21H,23H-porphyrin (Zn-TAPP) and Cu-5,10,15,20-tetrakis(4-formylphenyl)-21H,23H-porphyrin (Cu-TFPP) in the presence of n-butanol, o-dichlorobenzene and aqueous acetic acid. The resulting ZnCu-Por-COF possessed effective π-conjugation and high charge-transfer transition.
Interestingly, COFs with two types of covalent linkage were realized by the orthogonal (interference-free) reaction strategy. Binary NTU-COF-1 with both boroxine ring and imine group was constructed from the copolymerization of 1,3,5-tris(4-aminophenyl)-benzene (TAPB) and 4-formylphenylboronic acid (FPBA), which possessed ditopic units of aldehyde and boronate. As indicated by FT-IR spectra, the appearance of B–O stretching bands (1336 cm−1 and 1305 cm−1), B–C band (1221 cm−1), B3O3 band (711 cm−1) and a strong CN band (1627 cm−1) verified the existence of B3O3 rings and imine linkage. Likewise, ternary NTU-COF-2 was successfully synthesized based on TAPB, FPBA, and HHTP with the formation of the C2O2B boronate ring and imine group. Accordingly, there are two paths for the design of bifunctional linkages. First, one of the building units possesses at least two functional moieties, which enables the simultaneous reactions of co-condensation and self-condensation with other functional building blocks, such as TATTA-FPBA COF (TATTA: 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)trianiline) and NTU-COF-1.83 Second, three functional building blocks were employed, and at least one of them has two different functional moieties to enable two non-interfering co-condensation reactions, like NTU-COF-2 and HHTP-FPBA-TATTA COF.83
Fig. 4 Schematic illustration of the formation of TpPa which involved the steps of reversible Schiff-base reaction and irreversible enol-to-keto tautomerism. |
Understandably then, the remarkable chemical stability endows β-ketoenamine linked COFs with exceptional potential for diverse applications such as photocatalytic reactions. Moreover, keto functionalities present in the β-ketoenamine core could help to enhance the lifetime of the excited triplet state.90 For example, two chemically stable β-ketoenamine COFs were prepared for photocatalytic hydrogen production.91 The designed TP-EDDA COF bearing acetylene functional groups was constructed from Tp and 4,4′-(ethyne-1,2-diyl)dianiline (EDDA), while the Tp-BDDA COF with diacetylene moieties was based on the reaction of Tp and 4,4′-(buta-1,3-diyne-1,4-diyl)dianiline (BDDA). The appearance of characteristic signals corresponding to CC and C–N bonds at ∼1451 and ∼1251 cm−1 confirmed the formation of β-ketoenamine functionalities. A much higher photocatalytic hydrogen evolution rate of TP-BDDA (324 ± 10 μmol h−1 g−1) was observed compared to that of TP-EDDA (30 ± 5 μmol h−1 g−1). Similarly, thioether-functionalized Thio-COF was fabricated via the acid-catalyzed reaction of Tp with thioether substituted diamine, which was highly stable toward water and common organic solvents (acetone, dichloromethane, ethanol, tetrahydrofuran, etc.).92 The introduction of the thioether group was beneficial for metal deposition and nanoparticle growth, paving the way for various applications, including optical and electronic devices.
In addition, Michael's addition–elimination strategy can also be used to construct β-ketoenamine linked COFs.93 A series of COFs were fabricated in a one-step process via the reaction of aromatic amines with di- and tritopic ketoenols. The disappearance of the N–H and C–N stretching at 3470, 3420, and 1206 cm−1 together with the appearance of a new C–N band at 1200 cm−1 in FT-IR spectra confirmed the formation of β-ketoenamine linkage. The obtained β-ketoenamine linked COFs exhibited improved hydrolytic stability owing to the intramolecular hydrogen bonding. The electron delocalization in these COFs generated a narrower band gap and reversible electrochemical doping. Moreover, a wide range of nucleophilic and electrophilic building units can be employed to form this kind of COFs.
The high robustness and easy processible nature of hydrazone COFs make them popular in various applications.96,97 The first visible-light-active COF was designed and prepared based on hydrazone linkage with the copolymerization of 2,5-diethoxy-terephthalohydrazide and 1,3,5-tris-(4-formyl-phenyl)triazine (TFPT).35 In the presence of Pt, the system produced 230–1970 μmol h−1 g−1 of hydrogen. Later, a hydrazone-linked TFB-COF was constructed from TFB and 2,5-dimethoxyterephthalohydrazide with a BET surface area of 1501 m2 g−1, which can be used as a photocatalyst for cross-dehydrogenative coupling reactions.98 Another two hydrazone COFs with rich hydroxy units were synthesized using water and then incorporated with CoII to investigate their Lewis acid catalytic activity.94 As a result, the metallated COFs were effective in catalyzing the cyanosilylation reactions of various aldehydes.
In addition, two crystalline benzobisoxazole-linked (BBO) COFs were prepared by the condensation of 2,5-diamino-1,4-benzenediol dihydrochloride with TFB or 1,3,5-tris(4-formylphenyl)benzene (TFPB) under the catalysis of cyanide.114 A three-step mechanism was proposed to explain the BBO linkage formation: (1) a phenolic imine linked intermediate was first formed, (2) then ring closure took place with the addition of cyanide to the imine and a benzoxazoline intermediate appeared, and finally (3) the benzoxazoline intermediate was oxidized under air, thereby promoting the BBO linkage formation.115 The resulting BBO-COF 1 and BBQ-COF 2 displayed excellent water stability and a high surface area of 891 m2 g−1 and 1106 m2 g−1, respectively. In another study, a room-temperature solution-phase reaction was employed to synthesize an azodioxy-linked COF (POR-COF) with I2-doping-enhanced photo-current generation.116 A series of spiroborate-linked ionic COFs (ICOFs) were synthesized with a high BET surface area up to 1259 m2 g−1, constructed from the transesterification of diol and trimethyl borate.117 Recently, an unsubstituted olefin-linked COF (COF-107) was first synthesized by Aldol condensation of 4,4′-biphenyldicarbaldehyde and 2,4,6-trimethyl-1,3,5-triazine (TMT).26 FT-IR and 13C CP-MAS spectroscopic techniques were utilized to verify the formation of –CHCH– linkage. The as-synthesized COF-701 possessed a BET surface area of 1715 m2 g−1 and high chemical robustness owing to the existence of unsubstituted olefin linkage.
As discussed above, various linkage motifs have been designed relating to the COF formation. Different linkages lead to different structures and properties, which usually correlated with the stability. It is easy to understand that the stability of COFs, especially in water and under light irradiation, is of crucial importance for their photocatalytic application. COFs based on boroxine and boronate ester linkages are susceptible to hydrolysis under humid conditions.118 Though enhanced stability has been achieved by protecting electron-deficient boron centers from degradation, such as in the case of the ionic spiroborate-linked COF117 and the alkylated COF-14Å,119 their applications to photocatalysis have still been hindered and limited studies have been done. Different from boron-based COFs, imine-linked and other nitrogen-containing COFs are more stable. Interlayer complementary π-interactions and intralayer hydrogen-bonding interactions have been developed to improve the stability of imine-linked COFs.120,121 Similarly, β-ketoenamine-linked COFs originating from the enol–keto tautomerization of their imine counterparts show much higher stability, and have been used in photocatalysis.87,122 The TzDTz COF (TpDTz: Tp and 4,4′-(thiazolo[5,4-d]thiazole-2,5-diyl)dianiline) was stable in boiling water and strong acids for up to 7 days, and the morphology, structure, and crystallinity were retained after a 72 h long photocatalysis experiment.123 COFs with hydrazone and azine linkages are also active in the photocatalytic process.124–126 The studies revealed that COFs obtained after photocatalysis retained connectivity and photoactivity, but lost a part of long-range order which could be ascribed to exfoliation in water and can be recovered in the original reaction conditions. Compared to the imine, hydrazine and azine COFs, triazine and phenazine-linked COFs show exceptional chemical stability, and the triazine unit as a photoactive group has been widely explored in photocatalysis.127–129 As for the newly developed CC-linked sp2 COF, extremely high stability has been found in the photocatalytic experiment. Under the light irradiation of 16 h, while imine-linked COF-LZU1 nearly lost its crystallinity in 4 h, g-C18N3-COF with CC linkages exhibited retained structure and activity despite a slight decay of crystallinity.130 Remarkably, the excellent photostability of g-C40N3 was proved by the nearly constant photocurrent density within the measurement period of 2600 s.131 Unlike MOFs, most COFs show enhanced stability because of the covalent bond, but it is still the key point to improve the water- and photostability of COF photocatalysts for practical application.
Fig. 6 (a and b) AFM of COF colloids prepared at 75% solvent concentration of CH3CN. (c) Representative VT-LCTEM image of COF-5 nanoparticles (55% growth solution). Reproduced with permission from ref. 139. Copyright 2017, American Chemical Society. (d) SEM images of g-C18N3-COF. (e) Top view SEM micrograph of g-C18N3-COF film. Reproduced with permission from ref. 139. Copyright 2019, American Chemical Society. |
0D structures with favorable surface speciation are deemed to display excellent photocatalytic activity compared with their bulk-phase counterparts. In addition, by reducing the particle size, the discrete energy levels arise at the band-edges of both the CB and VB considering the quantum confinement effect, thus improving the redox potential of photogenerated electrons and holes.12,144 However, the study of COF photocatalysts with 0D structure remains challenging.
Likewise, g-C18N3-COF with fibrillar morphology was prepared by Knoevenagel condensation of 1,4-diformylbenzene (DFB) with 2,4,6-trimethyl-1,3,5-triazine (TMTA) (Fig. 6d and e).130 Ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS) displayed that the absorption band edge of g-C18N3-COF was at 450 nm, indicating a strong visible-light harvesting. And π-conjugated g-C18N3-COF with an average lifetime of 7.25 ns revealed the suppressed photogenerated electron–hole recombination. With ascorbic acid as a sacrificial agent and Pt as a co-catalyst, an average H2 production rate of 292 μmol g−1 h−1 was achieved over g-C18N3-COF. In addition, COFs bearing Tp and melamine (MA) building units with visible-light-response features were synthesized as exfoliated thin ribbon-like and interwoven thread-shaped structures under different conditions (catalyst-assisted, solvent-assisted, and liquid-free) by ball milling.134 Compared to the thread-shaped COF, the optical absorption edge of the ribbon-like COF displayed a red-shift, enhancing solar utilization efficiency, and therefore leading to a higher photocatalytic degradation rate of phenol. These findings suggested that the morphology affected the photocatalytic activity of COF-based materials, which may be ascribed to the aggregation behavior, dispersity, and incident light-harvesting capability in water.
In another study, the BDT-ETTA COF based on amine-functionalized 1,1′,2,2′-tetra-p-aminophenylethylene (ETTA) and donor-type benzo[1,2-b:4,5-b′]-dithiophene-2,6-dicarboxaldehyde (BDT) was grown on an indium tin oxide substrate to yield BDT-ETTA COF thin films.161 The obtained COF thin films displayed strong visible light absorption with a threshold of ca. 550 nm and a band gap of 2.47 eV, indicating the photoactive potential. The results suggested that the BDT component could be the reason for the photoactivity, and the oriented COF thin films were beneficial to the photoresponse and stability. A new synthetic method was employed by directly condensing 3,4,9,10-perylenetetracarboxylic diimide (PDI) and cyanuric chloride (CC) to yield a CTF film photocatalyst.162 The CTF film with excellent photocatalytic activity showed an enhanced NADH regeneration of 75.88% and HCOOH production of 204.14 μM. Also, ultrathin 2D porphyrin nanodisks with enhanced photocatalytic activity were prepared by COF exfoliation via axial ligand incorporation. Porphyrin-containing COF DhaTph (Dha: 2,5-dihydroxyterephthalaldehyde, Tph: 5,10,15,20-tetrakis(4-aminophenyl)-21H,23H-porphyrin) was exfoliated by simultaneously incorporating 4-ethylpyridine and copper (Cu) ion ligands into the porphyrin center to yield e-CON(Cu, epy) (Fig. 7a and b).163 The resulting e-CON was further incorporated with Pt nanoparticles and reduced-graphene oxide (RGO) to obtain the composite material e-CON(Cu, epy)/Pt/RGO for photocatalytic reaction (Fig. 7c). Compared with DhaTph/Pt/RGO, an enhanced visible/NIR-light-induced hydrogen evolution of the e-CON(Cu, epy)/Pt/RGO system was observed owing to the higher surface area of e-CON and Pt/RGO. The abovementioned results demonstrated that 2-dimensional COF thin films and nanosheets with broad light absorption, optical band gap, and efficient charge separation and transfer have great potential for photocatalytic activity improvement.
Fig. 7 (a) Scheme for the preparation of e-CON. (b) SEM image of e-CON (Cu, epy) deposited on the silicon wafer. (c) H2 evolution upon irradiation with visible (>420 nm) and NIR (>780 nm) light using e-CON(Cu, epy)/Pt/RGO and DhaTph/Pt/RGO. Reproduced with permission from ref. 163. Copyright 2019 Springer Nature Limited. |
Fig. 8 Scheme of the formation of hollow COFs via a templating strategy. Reproduced with permission from ref. 137. Copyright 2015, Royal Society of Chemistry. |
Apart from hollow morphology, flower-shaped morphology was also obtained by controlling the synthetic conditions of TpPa-1 and TpPa-2 COFs.87 Each flower was assigned to the aggregation of sheet-like petals with 1–3 μm in length as a result of the π–π stacking of COF layers. Specifically, the petals of TpPa-1 with spike-like tips were grown out from a core, while TpPa-2 with longer and broader petals showed a plate-like structure. Similar morphology has also been found in the Tp-Azo COF, in which petals with an average length of 40–50 nm aggregated to form a flower-like structure.174 Later, a bouquet-shaped magnetic TpPa-1 COF was successfully fabricated through a room-temperature solution-phase approach.138 Clustered Fe3O4 NPs were used as the template for growing TpPa-1 with a thread-like structure. With the increase of reaction time to 30 min, the branched TpPa-1 interconnected with each other to form bouquet-like morphology. Magnetic TpPa-1 with enhanced reactant and active site accessibility exhibited potential for photocatalysis as a result of its high porosity, large surface area and supermagnetism.
Fig. 9 (a) A tunable triphenylarene structure. (b) Formation of Nx-COFs based on hydrazine and Nx-aldehydes. (c) PXRD patterns of Nx-COFs compared with the simulated pattern calculated for the representative N3-COF. (d) H2 production monitored over 8 h using Nx-COFs as a photocatalyst in the presence of triethanolamine as a sacrificial electron donor. Reproduced with permission from ref. 102. Copyright 2015 Macmillan Publishers Limited. |
Various functional building blocks such as triazine,181 sulfone,124 pyrene,182 benzothiadiazole,183 and thiophene184 have been utilized for constructing COF photocatalysts with high performance. For example, diacetylene-bridged COFs were of great interest owing to the highly conjugated structures, accessible active sites, and accelerated charge transfer.185 Porous and stable acetylene (–CC–) and diacetylene (–CC–CC–) functionalized β-ketoenamine COFs, TP-EDDA and TP-BDDA, were prepared and their photocatalytic properties were studied (Fig. 7).91 Ketoenamine linkage was introduced to ensure the chemical stability of the COFs. To well-determine the influence of acetylene and diacetylene functional groups, an isoreticular COF, namely, TP-DTP COF (DTP: 4,4′′-diamino-p-terphenyl), with similar pore apertures based on terphenylene edges was designed and prepared. As determined by UV-vis spectra, TP-BDDA showed an absorbance edge of 525 nm and the tail extended up to 675 nm, while the absorbance edge of TP-EDDA and TP-DTP was 520 nm and 500 nm, respectively. Similarly, the optical band gaps followed the order of TP-BDDA (2.31 eV) < TP-EDDA (2.34 eV) < TP-DTP (2.42 eV). Photocatalytic experiments indicated that the conjugated diacetylene group played a vital role in enhancing the photoactivity. Apart from narrowing the band gap, diacetylene-moieties were also considered to possess higher charge carrier mobility and enable the accelerated migration of photogenerated excitons to the surface of the photocatalyst. In addition, electron acceptors such as benzothiadiazole (BT) and electron donors such as tris-(4-aminophenyl)triazine (TAPT) and tris(4-aminophenyl)benzene (TPB) were employed to construct COFs with tailored band gaps and improved charge separation and transfer.183 The resultant BT-COFs showed extended absorption bands ranging from 400 nm to 800 nm. Compared to TAPT-BT-COF, TPB-BT-COF with a narrower band gap and a more negative conduction band was found to exhibit promoted visible-light harvesting efficiency and produce more charge carriers. And the photocurrent intensity and electrochemical impedance spectra further confirmed that the structure of TPB-BT-COF was beneficial for enhanced charge carrier separation and reduced charge transfer impedance.
Fig. 10 (a) Electrochemical impedance spectroscopy plots of CTF-1 and CTFX samples. (b) Photocurrent responses under visible-light irradiation of CTF-1 and CTFX samples. (c) H2 evolution rates of CTF-1, CNCl and CTFX samples. (d) The band gap structures of CTF-1 and CTFX samples. Reproduced with permission from ref. 188. Copyright 2016, Royal Society of Chemistry. |
In addition to the abovementioned non-metal doping, metals such as Fe, Zn, and Re have also been doped into COFs for the modulation of their optical and electrical properties by narrowing the band gap, extending visible-light absorption, facilitating electron charge transfer, and increasing the lifetime of charge carriers.189–191 The nitrogen pots in the COFs provide rich binding sites for the incorporation of metal ions via ion coordination. The inclusion of Re in CTF-py (based on 2,6-dicyanopyridine) was developed by Cao et al. for the first time.189 Compared with CTF-py, Re-modified Re-CTF-py showed a lower charge transfer resistance and a higher charge carrier separation efficiency. By incorporating Re into CTF-py, photogenerated electrons could transfer from CTF-py to Re and the recombination of electron–hole pairs was retarded, thus leading to enhanced photocatalytic activity. Furthermore, BpZn-COP was synthesized by the coordination of Zn2+ with N atoms of pyridine units in Bp-COF.191 It was found that BpZn-COP showed broader light absorption (from 550 nm to more than 600 nm) and a narrower band gap (from 2.35 eV to 2.18 eV) compared to that of Bp-COP. The presence of Zn2+ played an important role in promoting the electron transfer inside the bulk and across the interface of semiconductor and electrolyte, suppressing electron–hole recombination and improving the utilization efficiency of charge carriers. By this way, BpZn-COP displayed a much higher photocatalytic activity.
Similarly, a novel MOF@COF core–shell hybrid material was constructed to possess high photocatalytic performance.201 By virtue of its available amino functional groups and high stability under harsh experimental conditions, NH2-MIL-68 with 2-aminoterephthalic acid ligand and infinite chains of InO4(OH)2 was selected. As depicted in the scheme, NH2-MIL-68 was first synthesized through solvothermal reaction, and then functionalized with the tris(4-formylphenyl)amine (TFPA) molecule to obtain aldehyde-functionalized NH2-MIL-68, denoted as NH2-MIL-68(CHO). And TPA-COF was grown on the NH2-MIL-68(CHO) surface by covalently linking tris(4-aminophenyl)amine with TFPA via conventional solvothermal condensation, generating core–shell structured hybrid NH2-MIL-68@TPA-COF (Fig. 11). NH2-MIL-68@TPA-COF displayed higher photocatalytic activity, which was about 1.4 times higher than that of NH2-MIL-68, due to its large BET surface area as well as smaller band gap.
Fig. 11 (a) General schematic of the synthesis of NH2-MIL-68@TPA-COF hybrid material. (b) N2 sorption isotherms for NH2-MIL-68, TPA-COF, and NH2-MIL-68@TPA-COF measured at 77 K. (c) UV-vis DRS spectra, and (d) the plots of the Kubelka–Munk function of NH2-MIL-68, TPA-COF, NH2-MIL-68@TPA-COF, and the mixture of NH2-MIL-68 and TPA-COF (NH2-MIL-68 + TPA-COF). Reproduced with permission from ref. 201. Copyright 2018, Wiley-VCH. |
As discussed above, strategies including functional building block incorporation, elemental doping, incorporation of sensitizers, and hybrid construction have been utilized in enhancing the photocatalytic performance of COFs. Among them, while functional building block incorporation as a distinctive feature of COFs has been widely used, the exploration is far from enough in view of the unlimited building molecules, and much work still needs to be done regarding the synthesis of new functional COFs. And post-synthetic modification will also be a promising strategy to utilize photoactive groups which are difficult for ab initio construction. Besides, heterojunction construction attracts considerable interest. By building suitable band positions, it is able to transfer photogenerated electron–hole pairs from the interface to the surface of two components, which leads to redox and reduction reactions. Certainly, new strategies with high performance are highly desired.
The first use of a COF for photocatalytic hydrogen production was reported in 2014.35 The triazine-based building block was selected because of its high electron mobility and electron-withdrawing characteristic.208 Specifically, the crystalline hydrazone-linked COF (TFPT-COF) was prepared by condensation of 2,5-diethoxy-terephthalohydrazide with TFPT. Then, Pt as a proton reduction catalyst and TFPT-COF as the photosensitizer were integrated to form the TFPT-COF/Pt photocatalyst for visible-light-induced hydrogen evolution with sodium ascorbate or TEOA as an electron donor. A hydrogen evolution rate of 1970 μmol h−1 g−1 was achieved with 10 vol% TEOA, which was nearly 3 times higher than that achieved with other outstanding photocatalytic systems including crystalline poly(triazine imide) and Pt-modified amorphous melon.36 Moreover, the quantum efficiency was determined to be 2.2% at 500 nm. Interestingly, on the one hand, TFPT-COF with retained photoactivity lost its crystallinity after 92 h photocatalytic reaction, probably due to its exfoliation in the process; on the other hand, this filtered amorphous product could be easily reconverted to the crystalline TFPT-COF just by putting it under the original experimental conditions without additional new building units, which suggested that the connectivity and photoactivity of TFPT-COF were retained.
As discussed before, one of the most intriguing characters of COFs is structural tunability, which allows for structure-to-function design at an atomic level. Indeed, many kinds of research studies on COF-based photocatalysts for water splitting have been done by tailoring the building blocks and linkages. For instance, a series of planar pyrene-based A-TEXPY-COFs were designed and synthesized by extending alkynes with the variation of peripheral heteroaromatic building units.182 The visible-light-driven hydrogen production by COF photocatalysts was studied by using Pt as the co-catalyst and 10 vol% TEOA as a sacrificial electron donor. A-TEBPY-COF constructed from 1,3,6,8-tetrakis(4-ethynylbenzaldehyde)-pyrene (TEBPY) and hydrazine with the lowest nitrogen content and thereby the most advanced donor features exhibited the highest hydrogen production rate of 98 μmol h−1 g−1 in this system. The results were in accordance with an increasing thermodynamic driving force for hydrogen reduction with decreasing nitrogen content.
Previous studies revealed that the rigid, planar dibenzo[b,d]thiophene sulfone (DBTS) unit was conducive to visible-light-induced photocatalytic evolution.209 The DBTS unit was incorporated into ordered COFs to investigate their photocatalytic activity.124 The as-prepared FS-COF exhibited a high hydrogen generation rate, up to 16300 μmol h−1 g−1, which is almost ten times higher than that of N3-COF. Later, three ketoenamine-based COFs were prepared to investigate the effect of different groups on photocatalytic performance.210 Specifically, TpPa-COF-X (X = –H, –(CH3)2, and –NO2) were constructed from the same host backbone with different functional groups anchored on the framework. In the photocatalytic experiment, H2 evolution efficiency decreased in the order of TpPa-COF-(CH3)2 > TpPa-COF > TpPa-COF-NO2. The order was attributed to the electron-donating ability of the three groups, –CH3 > –H > –NO2, which resulted in more efficient charge transfer within the COF framework. Besides, benzothiadiazole as the electron-withdrawing unit and thiophene as the electron-donating moiety were selectively introduced into CTFs.184 The as-prepared CTF-BT/Th was dispersed in water containing 3 wt% Pt as a co-catalyst and 10 vol% TEOA as a sacrificial agent under visible-light irradiation, and it exhibited a maximum hydrogen evolution rate of 6600 μmol h−1 g−1 and an AQE of 7.3% at 420 nm. Notably, the AQE was the highest value compared to the triazine-based polymer photocatalysts that existed at that time. To further enhance the activity, an attractive COF-based hybrid material was prepared based on benzoic acid-modified CTF-1 (B-CTF-1) and NH2-MIL-125(Ti) or NH2-UiO-66(Zr).211 The results showed that the hydrogen evolution rate over 15 wt% NH2-MIL-125(Ti)/B-CTF-1 (15TBC) was 360 μmol h−1 g−1 under visible light irradiation, which was twice higher than that of B-CTF-1. This enhanced photocatalytic activity of 15TBC could be ascribed to the appearance of amide bonds between MOFs and B-CTF-1, which facilitated charge separation and improved the photocatalyst stability.
Notably, considering the charge recombination and the kinetic overpotential for hydrogen production, there is no evidence for current COFs to produce H2 without a co-catalyst. Metallic Pt with a large work function has been widely used for electron trapping among photocatalysts, which also provides efficient proton reduction sites, making the facile H2 formation.212 Thus, the COF backbone with Pt coordination sites enables the specific interaction of COF and Pt, leading to the enhanced charge transfer. However, the stability of Pt in this environment limits its development.213,214 Developing earth-abundant, scalable, low-cost co-catalysts, which are water-soluble and can also interact with a heterogeneous photoabsorber, is urgent. Apart from Pt, MoS2 quantum dots (QDs) with high quantum confinement and small-size effect also represent a prominent candidate as the hydrogen generation co-catalyst.215,216 MoS2 QD modified CTF (MoS2/CTF) composites were reported to yield higher photocatalytic hydrogen production from water under visible-light illumination. MoS2 QDs were easily distributed on the surface of CTFs uniformly via an in situ photo-deposition method.217 The obtained MoS2/CTF composites showed obviously enhanced photocatalytic hydrogen evolution compared to the original CTFs and MoS2/g-C3N4 composite. This high activity was ascribed to the interactions between CTFs and MoS2, which enabled the efficient electron–hole transfer and separation. Cobaloximes, as the most efficient transition metal-based co-catalysts, feature easy synthesis and low overpotentials for hydrogen evolution, and can be easily introduced into the photocatalytic system.218 Lotsch and co-workers firstly selected noble-metal-free cobaloximes as a co-catalyst in the N2-COF-based photocatalytic proton reduction.219 Several factors influenced the H2 evolution rate including the solvent, sacrificial donor, reaction pH, and the fundamental properties of COFs such as crystallinity and porosity. By selecting azine-linked N2-COF as the photosensitizer, chloro(pyridine)cobaloxime as the co-catalyst, and TEOA as a sacrificial donor, a H2 evolution rate of 782 μmol h−1 g−1 and a TON of 54.4 were obtained in a mixture of water and acetonitrile. Herein electrons were transferred from the LUMO of the COF to the co-catalyst, following a monometallic pathway of H2 evolution from the CoIII-hydride and/or CoII-hydride species. As the cobaloxime tends to be inactive within few hours owing to decomposition or hydrogenation, an earth-abundant, noble-metal-free nickelthiolate hexameric cluster was further employed.123 A visible-light-induced hydrogen evolution system was constructed with TzDTz COF (TpDTz: Tp and 4,4′-(thiazolo[5,4-d]thiazole-2,5-diyl)dianiline) as a photosensitizer, Ni-thiolate cluster (NiME) as a co-catalyst, and TEOA as a sacrificial agent (Fig. 12). As a result, a sustained high H2 evolution rate of 941 μmol h−1 g−1 and a TONNi > 103 were observed over 70 h visible-light illumination.
Fig. 12 (a) Schematic illustration of photocatalytic H2 evolution. (b) The proposed key steps of the photocatalytic H2 evolution reaction with TpDTz COF and NiME cluster cocatalyst. [Ni-L] denotes a ligand-coordinated co-catalyst state which is attained fast compared to the [R] state, [R] denotes the catalyst resting state, which is catalytically active nickel cluster species, [D] denotes the deactivated species, and [I] denotes an intermediate reduced catalyst species able to run the HER step. Reproduced with permission from ref. 123. Copyright 2019, American Chemical Society. |
Fig. 13 Photocatalytic O2 evolution of BpCo-COF under visible light irradiation (λ ≥ 420 nm). Reproduced with permission from ref. 224. Copyright 2020, Elsevier B. V. |
Azine-based COFs with the existence of π-stacking aromatic units have been regarded as one of the most attractive candidates for photocatalysis. A large conjugated structure could facilitate the separation and transfer of photo-induced electrons/holes. Recently, two azine-linked crystalline COFs ACOF-1 (hydrazine, TFB) and N3-COF were utilized as photocatalysts for visible-light-induced reduction of CO2 with H2O as a hole scavenger.126 Understandably, in the reaction of CO2 photoreduction, the CO2 absorption capability of the catalyst is the key point. In this study, the high surface area of ACOF-1 (1053 m2 g−1) and N3-COF (1412 m2 g−1) with abundant accessible nitrogen sites rendered them with high CO2 absorption, leading to the facilitated photocatalytic reduction of CO2 to CH3OH. Upon 24 h visible light irradiation, the total amount of CH3OH generated over N3-COF was 13.7 μmol g−1, which was much higher than that of ACOF-1 (8.6 μmol g−1). Compared with ACOF-1, N3-COF with electron-poor triazine moieties was able to stabilize the negative charge generated on the COF which was important for the enhanced photocatalytic activity. It should be noted that the activity of these COFs outperformed that of other materials such as g-C3N4 (4.8 μmol g−1) under similar reaction conditions.231,232 Furthermore, the electronic properties and configuration of N3-COF and ACOF-1 were calculated with density functional theory (DFT). The results suggested that the potential of their LUMO was enough to drive CO2 reduction although the band gap was not suitable for the visible light response. Under visible light irradiation, the excited electrons at the LUMO energy level could reduce the adsorbed CO2 on the catalyst surface to produce methanol.
Apart from using COF itself as a photocatalyst for the reduction of CO2, crystalline COFs have also been considered as a photosensitive supporter to stabilize metallic active moieties for CO2 conversion.233 Rhenium(I) bipyridine (bpy) complexes are widely used in constructing photocatalysts to selectively reduce CO2 into CO under visible light irradiation.234,235 A pyridine-based CTF, namely CTF-py, constructed from 2,6-dicyanopyridine (DCP) with abundant N,N-chelating sites allowed for coordination of rhenium complexes targeting CO2 photoreduction. CTF-py was firstly synthesized via traditional trimerization reaction, and then the rhenium complex Re(CO)5Cl was introduced into the nitrogen sites of CTF-py to obtain Re-CTF-py through post-synthetic modification.189 The photocatalytic CO2 conversion was investigated in a solid–gas system under the irradiation of UV-Vis light, which could avoid dimerization and leaching of reactive species. The production of CO linearly increased with the irradiation time. The highest CO production rate of 353.05 μmol g−1 h−1 was observed on Re-CTF-py after 10 h continuous irradiation, while in the case of pristine CTF-py and the physical mixture it was only 13.4 μmol g−1 h−1 and 156.2 μmol g−1 h−1, respectively. The photogenerated electrons could easily transfer from CTF-py to Re via the coordination bond, indicating the efficient separation of photo-induced carriers. Using a similar strategy, a triazine COF derived from the condensation of 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)trianiline (TTA) and 2,2-bipyridyl-5,5-dialdehyde was selected as a photosensitizer to incorporate with a Re complex (Re(bpy)(CO)3Cl) for photocatalytic conversion of CO2 to CO.127 When using a Xe lamp as a light source (λ ≥ 420 nm) and TEOA as an electron donor, the resulting Re-COF showed a steady CO generation of 15 mmol g−1 for more than 20 h after the 15 min induction period with a TON of 48, which was 22 times better than that of its homogeneous Re(bpy)(CO)3Cl.
Very recently, COFs were also developed as functional supporters, like TpBpy COF, to anchor active sites for photocatalytic CO2 conversion.236–238 Compared with TpBpy, the introduction of Ni resulted in a red-shifted absorption edge and a narrower band gap due to the increased delocalization. Moreover, Ni-TpBpy helped to enhance the CO2 absorption capacity and isosteric heats, which could be ascribed to the Lewis acid–base interaction between adsorbed CO2 molecules and loaded Ni ions.217 In the experiment of Ni-TpBpy photocatalytic CO2 reduction, [Ru(bpy)]3Cl acted as a photosensitizer and TEOA served as an electron donor. Upon illumination, Ru(bpy)32+ was excited and transferred electrons to reduce the coordinated CO2 molecules on Ni-TpBpy (Fig. 14). The affinity of CO2 on Ni sites over H+ was crucial for the inhibition of H2 formation. As a result, the generated amount of H2 and CO from the Ni-TpBpy catalytic system was 170 and 4057 μmol g−1 within 5 hours, respectively, indicating a higher selectivity to CO. This CO production was comparable to other previous reported MOFs and COFs. Control experiments revealed that single Ni sites in the TpBpy framework acted as catalytic sites while TpBpy facilitated the activity as well as selectivity as a functional support.
Fig. 14 Schematic illustration of CO2 photoreduction over Ni-TpBpy. Reproduced with permission from ref. 217. Copyright 2019, American Chemical Society. |
Considering the similar features of nitrogen-rich rings and π-conjugated structure to g-C3N4, COFs with visible-light catalytic active moiety C3N4 exhibited great potential to become a qualified photocatalyst. Over the years, triazine-based COFs have been explored due to their superior photodegradation efficiency compared with g-C3N4.247–249 Likewise, ultrastable TpMA with the C3N4 active center was synthesized by the co-condensation of Tp and MA under solvothermal conditions, which involved a two-step path of reversible Schiff-base reaction and irreversible enol–keto tautomerization.249 This subtly designed structure endowed TpMA with enhanced light-harvesting capability and photooxidation property as a result of the reduced band gap and positive-shifted VB position. MO was selected as a model pollutant to assess the photocatalytic performance of TpMA under visible-light illumination. MO molecules could be degraded by the TpMA photocatalyst within 40 min, whereas the bulk g-C3N4 photocatalytic system showed almost no degradation under the same conditions. In order to exclude the photosensitive effect, colorless organic contaminant phenol was also chosen to evaluate the photocatalytic performance of TpMA. Notably, 90% of phenol was decomposed by TpMA in comparison with 8% decomposition by g-C3N4 after 40 min irradiation. Upon visible light irradiation, TpMA could be excited when the energy was greater than or equal to its band gap (2.30 eV). Then the dissolved O2 quickly captured the electrons from the CB to form O2− (E0, O2/O2− = −0.33 eV vs. NHE),250–252 and the O2− radicals thus formed reacted with H2O to further produce active OH˙. Consequently, MO molecules could be effectively oxidized and mineralized by reactive oxygen species O2− and OH−. According to the total organic carbon (TOC) measurements, TpMA achieved 36.7% of MO mineralization after 40 min-irradiation.
Recently, three imine-linked COFs with a visible-light catalytic active triazine ring were prepared by condensation of three different nitrogen-containing building blocks with the same aldehyde 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)tribenzaldehyde (A), which yielded COFA+B, COFA+C, and COFA+D, respectively. Specifically, the three different monomers were 1,3,5-tris(4-aminophenyl)benzene (B), 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)trianiline (C), and 2,4,6-tris(4-hydrazinylphenyl)-1,3,5-triazine (D) (Fig. 15a).253 The MO dye and colorless phenol as model pollutants were selected to assess the photocatalytic performance of the as-prepared COFs. The BET surface area and the corresponding pore volume followed a similar trend, as COFA+D (458 m2 g−1, 0.434 cm3 g−1) < COFA+B (907 m2 g−1, 0.436 cm3 g−1) < COFA+C (1903 m2 g−1, 0.455 cm3 g−1). Although the large surface area and accessible porous nature are beneficial for mass transfer,112,254 the interaction between adsorbent and adsorbate should be given high priority, especially in the liquid phase. In the case of COFA+D, it exhibited higher absorption activity than the other two COFs due to the existence of H-bonding between azo groups of MO and hydrazine groups of COFA+D. A similar phenomenon was also observed with phenol since the N-containing group could interact with the hydroxyl groups of phenol. Therefore, the absorption of MO and phenol followed a reverse trend compared to the BET surface area, as COFA+B > COFA+D > COFA+C. MO molecules could be completely degraded by COFA+C under 30 min visible light irradiation, while only 29.6% MO could be removed by COFA+B, and COFA+D showed almost no degradation. Similarly, the photocatalytic degradation of phenol followed the order of COFA+C > COFA+B > COFA+D (Fig. 15b). The hydrazine groups on COFA+D broke the π-delocalized electron system, leading to the reduction of electron-transfer conductivity and decreasing the interfacial charge transfer; and the CB edge potential of COFA+D was too positive to reduce the molecular oxygen to O2− species, resulting in poor photocatalytic performance. On the other hand, different from COFA+B, the interdigitated triazine-benzene heterojunctions in COFA+C enabled decreased electron–hole recombination. As a result, COFA+C with a higher density of active centres and conjugation degrees showed the highest photocatalytic performance. COFA+C was excited to generate electrons and holes under visible-light illumination. The dissolved O2 captured the accumulated electrons to yield abundant O2−, and then the obtained O2− further reacted with H2O to produce OH˙. On the other hand, the holes could easily transfer to water or oxidized pollutants, which enabled effective charge separation. Thus, the finally generated reactive radicals including O2− and OH˙ could degrade pollutants effectively (Fig. 15c). In addition, COFA+C did not show any major loss of activity after four photocatalytic cycles, indicating its high stability and renewability.
Fig. 15 (a) Structure and band alignment of COFA+B, COFA+C and COFA+D. (b) Photocatalytic performance of COFA+B, COFA+C and COFA+D under visible light irradiation. (c) Schematic illustration of pollutant photodegradation over COFA+C under visible light irradiation. Reproduced with permission from ref. 253. Copyright 2017, Elsevier B. V. |
Besides, other functional building units have also been utilized to construct COFs with high photocatalytic performance. For example, a heptazine unit was embedded into the framework of CTF (forming PCN-1 and PCN-2), which was demonstrated to possess high photocatalytic performance toward degradation of RhB.255 In detail, PCN-1 was prepared by the polymerization of melem and 2,4,6-triformylphloroglucinol using a solvent of dimethyl sulfoxide, whereas PCN-2 with crystalline structure was obtained by incorporating melem moieties into CTF. Compared with the traditional polymer semiconductor g-C3N4, the PCN polymers showed broader absorption wavelength, even extended to the entire visible region. Moreover, the enhanced surface area of PCN-2 ensured more active surface sites, thereby giving more chance for reactants to access photoredox reactions. As for the photocatalytic performance, PCN-1 and CTF could degrade RhB within 120 min and 60 min, respectively, while PCN-2 could degrade RhB within 25 min under visible light irradiation. Triptycene with 3D spatial orientation containing three benzene rings is another attractive conjugated building unit for microporous materials synthesis. A triptycene-based imine-linked covalent organic polymer (TP-COP) was prepared for organic dye degradation.256 Graphene-like layered TP-COP was achieved by manual grinding of terepthaldehyde and triaminotriptycene at room temperature. The DRS analyst indicated that TP-COF responded to visible light, and a narrow band gap of ∼2.49 eV was determined by the Tauc plot. 95% of RhB degradation efficiency could be achieved within 160 min under sunlight irradiation. Meanwhile, TP-COP possessed remarkable reusability in RhB degradation without any visible performance decay.
In addition to the building block design, morphology control has also been regarded as an essential method to optimize the catalytic efficiency of photocatalysts.53,257,258 Hollow architectures have been investigated to not only promote the interaction between catalysts and substrates by decreasing the thickness of the structure but also enhance light absorption by multiple light reflections.171 As for other morphologies, TpMA with thread-like morphology could be synthesized by ball milling by varying the amount of liquid added during the process.134 With the addition of p-toluenesulfonic acid and 1 mL solvents, crystalline TpMAC(1mL) was achieved with well-defined morphology of an interwoven thread shape. When the solvent volume was increased to 3 mL, crystalline TpMAC(3mL) with thin ribbon-like morphology was presented. Both TpMAC(3mL) and TpMAC(1mL) were able to respond to visible light and the optical band gap of TpMAC(3mL) and TpMAC(1mL) was 2.29 eV and 2.56 eV, respectively. 10 mg L−1 phenol as a model environmental contaminant was selected to evaluate the photocatalytic performance of TpMAC(3mL) and TpMAC(1mL). Consequently, phenol was completely decomposed after 60 min over TpMAC(3mL) under visible light irradiation, while only 83.5% phenol was degraded by TpMAC(1mL).190,200,201,259
In addition to the morphology control, heterojunction construction has also been used to improve the photocatalytic degradation performance of COFs. For example, a Z-scheme MOF/COF heterojunction was firstly reported by the incorporation of NH2-MIL-125(Ti) with TTB-TTA (TTB: 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)tribenzaldehyde).259 The NH2-MIL-125(Ti)/TTB-TTA composite exhibited enhanced photocatalytic performance for MO degradation because of efficient charge separation through the covalent heterojunction interface. In addition, a BiOBr/CTF-3D composite was designed and prepared, showing enhanced photocatalytic activity toward antibiotic removal.200
Photocatalysts | Building blocks of COFs | Conditions | Applications | Ref. |
---|---|---|---|---|
CN-COF | Tp, 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)trianiline (TTA), g-C3N4 | 100 mg of photocatalyst dispersed in 200 mL of aqueous solution containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp with a cut-off filter (λ ≥ 420 nm) | Photocatalytic H2 evolution, 10100 μmol h−1 g−1 | 181 |
FS-COF | 3,9-Diamino-benzo[1,2-b:4,5-b′]bis[1]benzothiophene-5,5,11,11-tetraoxide,2,4,6-triformylphloroglucinol | 5 mg of photocatalyst dispersed in 25 mL of 0.1 M ascorbic acid water solution, Pt as a co-catalyst, lactic acid (LA) as the sacrificial agent, 300 W Xe lamp with a cut-off filter | Photocatalytic H2 evolution, 16300 μmol h−1 g−1 | 124 |
g-C40N3-COF | 4,4′′-Diformyl-p-terphenyl (DFPTP), 3,5-dicyano-2,4,6-trimethylpyridine (DCTMP) | 50 mg of photocatalyst dispersed in 100 mL of deionized water containing 10 vol% TEoA as a sacrificial agent, 3% Pt as a co-catalyst, 300 W Xe lamp with a 420 nm cut-off filter | Photocatalytic H2 evolution, 4120 μmol h−1 g−1 | 180 |
CdS-COF (90:10) | 1,3,5-Triformylphloroglucinol (Tp), 2,5-dimethyl-p-phenylenediamine (Pa-2) | 30 mg of photocatalyst dispersed in 10 mL of deionized water, 0.5 wt% Pt as a co-catalyst, lactic acid (LA) as the sacrificial agent, 400 W xenon arc lamp with a UV-cut-off filter (λ ≥ 420 nm) | Photocatalytic H2 evolution, 3678 μmol h−1 g−1 | 196 |
sp2c-COFERDN | 3-Ethylrhodanine, 1,3,6,8-tetrakis(4-formylphenyl)pyrene (TFPPy), 1,4-phenylenediacetonitrile (PDAN) | 50 mg of photocatalyst powder dispersed in 100 mL of aqueous solution containing 10 vol% TEoA as a sacrificial electron donor, 3 wt% Pt as a co-catalyst, 300 W Xe lamp with a water-cooling filter | Photocatalytic H2 evolution, 2120 μmol h−1 g−1 (λ ≥ 420 nm) | 223 |
TFPT-COF | 1,3,5-Tris-(4-formyl-phenyl)triazine (TFPT), 2,5-diethoxy-terephthalohydrazide (DETH) | 4 mg of photocatalyst dispersed in 9 mL of water containing 1 mL of TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 1970 μmol h−1 g−1 | 35 |
N3-COF | Hydrazine hydrate, 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)tribenzaldehyde | 5 mg of photocatalyst dispersed in 10 mL of PBS containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp with a cut-off filter (900 nm > λ > 420 nm) | Photocatalytic H2 evolution, 1703 μmol h−1 g−1 | 102 |
N2-COF | 4,4′,4′′-(Pyrimidine-2,4,6-triyl)tribenzaldehyde, hydrazine | 5 mg of photocatalyst dispersed in 10 mL of water containing 100 μL of TEoA as a sacrificial agent, chloro(pyridine)cobaloxime(III) (Co-1) as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 782 μmol h−1 g−1 | 219 |
g-C18N3-COF | 1,4-Diformylbenzene (DFB), 2,4,6-trimethyl-1,3,5-triazine (TMTA) | 50 mg of photocatalyst dispersed in 100 mL of 1 M aqueous ascorbic acid solution, 3% Pt as a co-catalyst, 300 W Xe lamp with a 420 nm long pass cut-off filter | Photocatalytic H2 evolution, 292 μmol h−1 g−1 | 130 |
A-TEBPY-COF | 1,3,6,8-Tetrakis(4-ethynylbenzaldehyde)-pyrene (TEBPY), hydrazine | 10 mg of photocatalyst dispersed in 9 mL of water containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 98 μmol h−1 g−1 | 182 |
ter-CTF-0.7 | 4,7-Bis(4-formylphenyl)-2,1,3-benzothiadiazole (M-BT), 3,6-dicarbaldehyde-N-ethylcarbazole (M-CBZ), terephthalimidamide dihydrochloride | 50 mg of photocatalyst dispersed in 100 mL of deionized water containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 19120 μmol h−1 g−1 | 207 |
CdS-NP/5%CTF-1 | 2,6-Dicyanopyridine | 20 mg of photocatalyst dispersed in 80 mL of ultrapure water, 1 wt% Pt as a co-catalyst, lactic acid (LA) as the sacrificial agent, 300 W Xe lamp with a UV-cut-off filter (λ ≥ 420 nm) | Photocatalytic H2 evolution, 7500 μmol h−1 g−1 | 197 |
CTF-BT/Th | 4,4′-(Benzothiadiazole-4,7-diyl) dibenzonitrile, 4,4′-(thiophene-2,5-diyl) dibenzonitrile | 50 mg of photocatalyst dispersed in water containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp with a cut-off filter (λ > 420 nm) | Photocatalytic H2 evolution, 6600 μmol h−1 g−1 | 184 |
CTF-1-100 W | 1,4-Terephthalonitrile (microwave-assisted synthesis) | 50 mg of photocatalyst dispersed in 200 mL of deionized water containing 23 mL of TEoA and 7 mL of methanol, 2.01 wt% Pt as a co-catalyst, 300 W Xe lamp with a 420 nm long pass filter | Photocatalytic H2 evolution, 5500 μmol h−1 g−1 | 74 |
CTF-0-M2 | 1,3,5-Tricyanobenzene (microwave-assisted synthesis) | 100 mg of photocatalyst dispersed in 230 mL of water containing 10 vol% TEoA as a sacrificial agent, 3 wt% Pt as a co-catalyst, 300 W Xe lamp with a 420 nm long pass filter | Photocatalytic H2 evolution, 4900 μmol in total during seven day-long runs | 222 |
CTF-0-I | 3-Ethylrhodanine, 1,3,6,8-tetrakis(4-formylphenyl)pyrene (TFPPy), 1,4-phenylenediacetonitrile (PDAN) | 50 mg of photocatalyst powder dispersed in 100 mL of aqueous solution containing 10 vol% TEoA as a sacrificial electron donor, 3 wt% Pt as a co-catalyst, 300 W Xe lamp with a water-cooling filter | Photocatalytic H2 evolution, 2120 μmol h−1 g−1 (λ ≥ 420 nm) | 222 |
CTFS10 | 1,4-Dicyanobenzene (trifluoromethanesulfonic acid catalysed synthesis) | 20 mg of photocatalyst dispersed in 50 mL of aqueous solution containing 10 vol% TEoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 2000 μmol h−1 g−1 | 129 |
15 wt% NH2-MIL-125(Ti)/B-CTF-1 (15TBC) | 1,4-Dicyanobenzene | 20 mg of photocatalyst dispersed in 80 mL of aqueous solution containing TEoA as a sacrificial agent, 3% Pt as a co-catalyst, 300 W Xe lamp with a cut-off filter (780 nm ≥ λ ≥ 420 nm) | Photocatalytic H2 evolution, 360 μmol h−1 g−1 | 211 |
CTF-T1/CTF-T2 | 1,4-Terephthalonitrile (trifluoromethanesulfonic acid catalysed synthesis) | 80 mg of photocatalyst dispersed in 70 mL of ultrapure water containing 10 mL of TeoA as a sacrificial agent, Pt as a co-catalyst, 300 W Xe lamp with a 420 nm cut-off filter | Photocatalytic H2 evolution | 271 |
CTFCl | 1,4-dicyanobenzene | 20 mg of photocatalyst dispersed in 50 mL of ultrapure water containing 10 vol% TEoA as a sacrificial agent, 5 wt% Pt as a co-catalyst, 300 W Xe lamp with a 420 nm band-pass filter | Photocatalytic H2 evolution | 188 |
Pd0/TpPa-1 | Tp, p-phenylenediamine (Pa-1) | 10 mg of photocatalyst dispersed in 100 mL of water containing 10 vol% TEoA as a sacrificial agent, Eosin Y as the sensitizer, 300 W Xe lamp with a cut-off filter (λ ≥ 420 nm) | Photocatalytic H2 evolution, 10400 μmol h−1 g−1 | 194 |
TpDTz | Tp, 4,4′-(thiazolo[5,4-d]thiazole-2,5-diyl)dianiline (DTz) | 5 mg of photocatalyst dispersed in 10 mL of water containing 10 vol% TEoA as a sacrificial agent, Ni-thiolate hexameric cluster (NiME) as a co-catalyst, 300 W Xe lamp | Photocatalytic H2 evolution, 941 μmol h−1 g−1 | 123 |
TP-BDDA | Tp, 4,4′-(buta-1,3-diyne-1,4-diyl)dianiline (BDDA) | 10 mg of photocatalyst dispersed in 34 mL of water containing 4 mL of TEoA as a sacrificial agent, 3 wt% Pt as a co-catalyst, 300 W Xe lamp with a cut-off filter of 395 nm | Photocatalytic H2 evolution, 324 ± 10 μmol h−1 g−1 | 91 |
e-CON(Cu, epy) | 2,5-Dihydroxyterephthalaldehyde (Dha), 5,10,15,20-tetrakis(4-aminophenyl)-21H,23H-porphine (Tph) | 0.25 mg of e-CON(Cu, epy) and 0.25 mL of Pt/RGO dispersed in 4.05 mL of water and 0.2 mL of EDTA (0.5 M), 1.0 sunlight, long-pass 420 nm filter for visible light or 780 nm filter for NIR light | Photocatalytic H2 evolution | 163 |
g-C40N3-COF | 4,4′′-Diformyl-p-terphenyl (DFPTP), 3,5-dicyano-2,4,6-trimethylpyridine (DCTMP) | 50 mg of photocatalyst dispersed in 100 mL of deionized water containing 0.01 mol L−1 AgNO3 as an electron acceptor, 3 wt% Co2+ as a co-catalyst, 300 W Xe lamp | Photocatalytic O2 evolution, 50 μmol h−1 g−1 | 180 |
sp2c-COF | TFPPy, PDAN | 50 mg of photocatalyst dispersed in 100 mL of aqueous solution containing AgNO3 as an electron acceptor and Co(NO3)2 as a co-catalyst, 300 W Xe lamp with a 420 nm long-pass cut-off filter | Photocatalytic O2 evolution, 22 μmol h−1 g−1 | 223 |
CTF-1-100 W | 1,4-Terephthalonitrile (microwave-assisted synthesis) | 50 mg of photocatalyst dispersed in 200 mL of 0.02 M AgNO3 aqueous solution, 3 wt% RuOx as a co-catalyst, 300 W Xe lamp with a 420 nm long pass filter | Photocatalytic O2 evolution, 140 μmol h−1 g−1 | 74 |
CTF-T1 | 1,4-Terephthalonitrile (trifluoromethanesulfonic acid catalysed synthesis) | 50 mg of photocatalyst dispersed in 50 mL of 0.01 M AgNO3 aqueous solution, RuO2 as a co-catalyst, 300 W Xe lamp | Photocatalytic O2 evolution | 271 |
N3-COF | Hydrazine hydrate, 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)tribenzaldehyde | 10 mg of photocatalyst dispersed in 5 mL of deionized water, pre-injected CO2, 500 W Xe lamp with a UV and IR cut-off filter (800 nm ≥ λ ≥ 420 nm) | Photoreduction of CO2, CH3OH (13.7 μmol g−1) was generated in 24 h | 126 |
Ni-TpBpy | 1,3,5-Triformylphloroglucinol, 5,5′-diamino-2,2′-bipyridine | 10 mg of photocatalyst dispersed in 5 mL of mixed solution of acetonitrile, H2O, and TEOA, pre-injected CO2, 300 W Xe lamp with a UV cut-off filter (λ ≥ 420 nm) | Photoreduction of CO2, CO was generated at a rate of 4057 μmol h−1 g−1 for 5 h | 218 |
Re-CTF-py | 2,6-Dicyanopyridine | 2 mg of photocatalyst dispersed on a porous quartz film in the reaction cell, TEoA as a sacrificial agent, 300 W Xe lamp | Photoreduction of CO2, 353.05 μmol h−1 g−1 for 10 h | 189 |
Re-TpBpy | Tp, 2,2′-bipyridine-5,5′-diamine | 15 mg of photocatalyst dispersed in the AcN/H2O mixture (10/1.8 mL) containing 0.1 M TEoA as an electron donor, 200 W Xe lamp with a high-pass filter at 390 nm | Photoreduction of CO2 to CO | 237 |
CTF-BT | 4,4′-(Benzothiadiazole-4,7-diyl)dibenzonitrile | 5 g of photocatalyst dispersed in 4-nitrophenol solution, white LED light | Complete reduction of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP) after 50 min | 167 |
NH2-MIL-68@TPA-COF | Tris(4-formylphenyl)amine (TFPA), tris(4-aminophenyl)amine (TAPA) | 5 mg of photocatalyst dispersed in 5 mL of Rh B aqueous solution, 300 W Xe lamp with a UV cut-off filter (λ ≥ 420 nm) | Degradation of Rh B | 201 |
PCN-2 | Melem, 1,3,5-triformyl phloroglucinol, 2,4,6-tris(4-aminophenyl)-1,3,5-triazine | 4 mg of photocatalyst dispersed in 80 mL of Rh B, 300 W Xe lamp with a cut-off light filter (λ > 420 nm) | Complete degradation of Rh B within 25 min | 255 |
TP-COP | Triaminotriptycene, terephthaldehyde | 100 mg of photocatalyst dispersed in 100 mL of Rh B solution, sunlight | 95% degradation of Rh B within 160 min irradiation | 256 |
Fe–TiO2@COF | Tp, 1,1′:4′,1′′-[terphenyl]-4,4''-diamine (Ta) | 0.4 mg of photocatalyst dispersed in 4 mL of MB solution (40 mg L−1), UV light, visible light, ambient light | 96% degradation of MB after 240 min irradiation | 190 |
COP-NT | 20 mg of photocatalyst dispersed in 20 mL of aqueous solutions of MO, RhB or MB respectively in the presence of 30% H2O2, 10 W LED | Degradation of MO, RhB and MB, 67% MO, 78% RhB and 57% MB degraded within 10 h, 4 h, and 100 min, respectively | 243 | |
CTF-A | 1,4-Dicyanobenzene | 5 mg of photocatalyst dispersed in 50 mL of MB solution (10 mg L−1), 300 W Xe lamp with a UV cut-off filter (λ ≥ 420 nm) | Degradation of MB, totally degraded within 60 min | 248 |
COFA+C | 4,4′,4′′-(1,3,5-Triazine-2,4,6-triyl)tribenzaldehyde (A), 4,4′,4′′-(1,3,5-triazine-2,4,6-triyl)trianiline (C) | 15 mg of photocatalyst dispersed in 50 mL of organic pollutant solution (10 mg L−1), 300 W Xe lamp with an optical cut-off filter (λ ≥ 420 nm) | Complete degradation of MO and 79% degradation of phenol after 30 min irradiation | 253 |
TpMA | Tp, melamine (MA) | 30 mg of photocatalyst dispersed in 50 mL of MB solution (10 mg L−1), 300 W Xe lamp with a UV cut-off filter (λ ≥ 420 nm) | Complete degradation of MO within 40 min | 249 |
BiOBr/CTF-3D | 1,4-Dicyanobenzene | 40 mg of photocatalyst dispersed in 200 mL of TC (10 mg L−1) or CIP (10 mg L−1) solution, 500 W Xe lamp | 90.9% degradation of TC within 50 min irradiation, higher degradation of CIP compared to BiOBr | 200 |
(1) The structures, morphologies and properties of COFs are most likely to be changed with different synthesis methods and reaction conditions, thereby leading to the different photocatalytic performance of COFs. Synthetic strategies such as solvothermal synthesis,260,261 ionothermal synthesis,262,263 microwave synthesis74,222,264,265 and room temperature synthesis266,267 have been developed for COF synthesis. While solvothermal synthesis is the most widely used method, the harsh synthesis conditions such as long reaction time and high temperature and pressure make it difficult for large-scale production. Ionothermal synthesis here is utilized for the synthesis of the photoactive triazine core, but the high reaction temperature and low crystalline products hamper the development. The microwave-assisted method and room temperature reaction seem to be better choices. However, only a few examples were reported,74,222,264–266 and thus further improvement is needed. Operative, low cost but effective synthetic methods with mild reaction conditions are eager to be introduced for the development of COFs with enhanced photocatalytic activity.
(2) New stable COFs with high efficiency are necessary. How to facilely control the band gap structure of COFs should be taken seriously. Efficient utilization of the solar spectrum is a significant prerequisite for photocatalysis, and efforts must be made to broaden the light absorption. Besides, the molar absorption coefficient, as a representative factor of light absorption at a specific wavelength, is highly connected to the photocatalytic activity that photocatalysts with a high molar absorption coefficient are able to utilize sunlight more effectively and generate more electron–hole pairs. Thus, constructing COF photocatalysts with enlarged light absorption as well as high molar absorption coefficient is encouraged. For example, as learnt from other traditional photocatalysts, long-wavelength-light-responsive building blocks such as lanthanide-based molecules and phthalocyanine units could be incorporated into COFs to extend the light absorption from visible light to NIR light.268–270 On the other hand, problem still exists in the high recombination rate of photogenerated charge carriers, which retards the effective transfer of electrons and holes. A two-photocatalyst system is found to replace single photocatalysts in nature to avoid inevitable back reaction. Similarly, Z-scheme systems are preferred considering that the photogenerated electrons and holes tend to be separated on divided subsystems, which minimizes the possibility of electron–hole recombination and enables longer-lived charge carriers.
(3) The fundamental mechanism of the COF-based photocatalytic system still remains unclear. Theoretical calculation as a very useful tool is capable of predicting the structures and properties as well as simulating the photocatalytic process. Physicochemical properties of COFs pertaining to the high photocatalytic activity, including surface area, crystallinity, conjugated structure, band gap configuration, visible-light absorption, charge separation and transfer, should be fully investigated. For example, by the utilization of first-principles calculations, three 2D-CTF models CTF-0,63 CTF-1,62 and CTF-264 were investigated including electronic band structures, conduction band minimum (CBM)/valence band maximum (VBM) position, work functions, and optical absorption spectra.271 As a result, 2D-CTFs with controllable construction are better candidates for visible-light-induced water splitting, which stimulated the experimental research on their photocatalytic properties. Besides, advance characterization, especially in situ and even operando technologies, should be taken into consideration to reveal the mechanism behind all the photocatalytic processes, which would provide the insight for further development of efficient COF-based photocatalysts. Technologies such as in situ FT-IR, in situ X-ray absorption spectroscopy (XPS) and in situ extended X-ray absorption fine structure (EXAFS) are highly recommended to monitor the reaction process, distinguishing reactive intermediates and investigating the active sites. More specifically, spectroscopy technologies, such as photoluminescence (PL) spectroscopy, transient absorption (TA) spectroscopy and Kelvin probe force microscopy-based spatially resolved surface photovoltage technique, are also needed for optical and electronic property analysis, corresponding to charge carrier transfer and recombination.
(4) Studies of O2 evolution and CO2 photoreduction using COF-based photocatalysts should also be strengthened in the near future, which are far less than the research on H2 evolution and pollutant degradation. It is a long-term goal to find high performance photocatalysts for visible-light-induced overall water splitting. On the other hand, as for CO2 photoreduction, increasing the product selectivity demands prompt solutions. The design of COFs with highly selective photocatalysis by elaborately choosing functional building blocks and components is highly desired. Additionally, in the current photocatalytic systems, uneconomical sacrificial electron donors and cocatalysts, such as TEOA and noble metal Pt, respectively, have often been used. Strategies like reducing the usage or using highly active but economical alternatives are to be achieved for the development of this area. Besides, the photocatalytic activity is known to be affected by varied conditions such as the amount of photocatalyst, the solvent volume, the kind of cocatalysts, the light source and intensity, and the temperature. It is difficult to compare the activity of photocatalysts reported by different groups. The standardization of photocatalytic activity evaluation has become an urgent necessity.
Footnote |
† These authors contribute equally to this article. |
This journal is © The Royal Society of Chemistry 2020 |