Chunhui
Dai
and
Bin
Liu
*
Department of Chemical & Biomolecular Engineering, National University of Singapore, 117585, Singapore. E-mail: cheliub@nus.edu.sg
First published on 1st October 2019
Conjugated polymers have recently been under active investigation as promising alternatives to traditional inorganic semiconductors for photocatalysis. This is due to their unique advantages of low cost, high chemical stability, and molecularly tunable optoelectronic properties. This critical review summarizes the recent advancements in π-conjugated polymers for visible-light-driven photocatalytic applications including water splitting, CO2 reduction, organic transformation and degradation of organic dyes. Special emphasis is placed on how the changes in the polymer structure could influence their physicochemical properties and photocatalytic activities. This structure–activity relationship analysis should guide rational molecular design of conjugated polymers for improved photocatalytic activity.
Broader contextThe exploitation of efficient photocatalysts to directly convert the radiant sunlight into chemical energy is of great importance to address the current energy and environmental challenges. Conjugated polymers (CPs) consisting of photoactive π-systems represent an attractive platform for solar energy utilization. They have been intensively studied for a variety of photocatalytic applications and many exciting performances were reported through facile molecular design. A comprehensive review is thus timely to summarize the progress of the field. This review article systematically presents the recent advances in conjugated polymers for visible-light-driven photocatalysis, including water splitting, CO2 reduction, organic transformation and degradation of organic dyes. The synthesis and design principles of conjugated polymers in these photocatalytic applications are illustrated, with an emphasis on the correlation between polymer structures and their photocatalytic activities. We expect that the systematic discussion in this review will not only provide a general overview of the field, but also promote the development of conjugated polymers with fascinating properties for photocatalysis. |
In a photocatalytic process, the photon energy required for photoexcitation is dependent on the optical gap of the photocatalyst. As compared to the UV region that comprises only 5% of solar light, the visible light (400–800 nm) accounts for about 53% of the incident solar energy (Fig. 1).4,25,26 Therefore, it is crucial to develop photocatalysts with optical gap below 3.0 eV to increase the absorption of solar light with a focus on the visible light region.
Fig. 1 AM1.5 solar light radiation spectrum (red) and photon flux curve (blue) as functions of light wavelength and material band gap. Reprinted with permission from ref. 25. Copyright 2017 Wiley-VCH. |
Over the past few decades, research efforts have been intensively devoted to pursuing suitable photocatalysts for practical applications. Among them, organic π-conjugated polymers have recently garnered significant attention as promising candidates for various photocatalytic applications. Many interesting features allow π-conjugated polymers to stand out in the field of photocatalysis over inorganic semiconductors. Firstly, they are environmentally friendly semiconductor materials, consisting of earth-abundant elements, and they possess tunable energy levels for oxidation and reduction reactions. Secondly, they could be facilely prepared under mild conditions and have excellent chemical stability against photobleaching. Thirdly, their molecular structures can be fine-tuned to better utilize visible light, while many traditional inorganic photocatalysts are only UV-active. Last but not least, the π-conjugation along the polymer backbone enables them with unique photogenerated charge carrier separation and transport properties, which is critical to trigger photoredox reactions.
Upon diverse modification of conjugated polymers (CPs) at the molecular level, many exciting achievements in photocatalysis have recently been reported, which clearly justify a comprehensive review on this research topic. So far, most CP based photocatalysis reviews have been based on the family of graphitic carbon nitride (g-C3N4) materials with a focus on one or two photocatalytic applications.18,27–35 There is an urgent need to provide a review focused on π-conjugated polymers in photocatalysis from a broad perspective to present an overview of the recent developments on this research topic.
This review presents the recent advances of CPs in applications of photocatalytic water splitting, CO2 photoreduction, light-driven organic transformation and photocatalytic degradation of organic dyes (Fig. 2). We focus on an in-depth understanding of how the CP structure modification could influence the optical properties, photogenerated charge separation and transport, and photocatalytic activity. The review starts from a brief discussion of general photocatalytic processes and the synthesis of CP photocatalysts. This is followed by the illustration of CPs for each photocatalytic application. Finally, perspectives toward exploring new CPs for photocatalytic applications are proposed, which is expected to stimulate the future development of more exciting CPs for photocatalysis.
Fig. 2 Overview of π-conjugated polymers for various photocatalytic applications, which are summarized in this review. |
Fig. 3 (a) The main process of polymeric semiconductors for photocatalysis (D: electron donor; A: electron acceptor) and (b) critical factors affecting their photocatalytic performances. |
According to the proposed mechanism of photocatalytic redox reactions, an efficient CP photocatalyst should at least have the following properties: (a) a strong and broad absorption in the visible region to effectively capture solar light; (b) high charge separation and migration efficiency for surface redox reactions; (c) appropriate energy level alignments to enable efficient charge separation as well as the reduction or oxidation half reaction. Therefore, light harvesting, energy levels as well as photogenerated charge separation and migration are three key factors affecting the photocatalytic activities of these CPs (Fig. 3b). This requires rational design and synthesis of CPs for effective photocatalysis.
Interestingly, conjugated polymer nanomaterials could be readily prepared under appropriate reaction conditions. For example, conjugated microporous polymer nanosheets could be obtained by one-step oxidative coupling of aromatic alkynes40 or a condensation reaction.41,42 Conjugated microporous polymer nanoparticles could be directly prepared by polymerization in high internal phase emulsion, which is concentrated emulsion with a dispersed phase volume fraction over 74.05%.43,44 The morphologies of nanoparticles can be varied by introducing monomers with different electron-donating/withdrawing properties.45
Additionally, the condensation reaction between amino and aldehyde groups based on Schiff-chemistry is widely performed to prepare crystalline covalent organic frameworks (COFs) for photocatalysis.51 The reaction is typically performed under harsh experimental conditions with carefully controlled temperature (usually 120 °C), pressure and reaction time (usually 3–5 days). Acetic acid is employed as the catalyst in the reaction. It is believed that an error-correction process occurs during the conversion of amorphous intermediates to crystalline products. The reaction solvent is crucial for the assembly of framework units. In a typical condensation reaction, mesytilene/dioxane and n-butanol/o-dichlorobenzene are commonly used reaction solvents. Besides, a suitable building block is also important to obtain a highly crystalline COF. Rigid and planar π-conjugated units are preferred to prepare 2D COFs.
Covalent triazine frameworks (CTFs) could be prepared by a trimerization reaction of aromatic nitriles in the presence of a Brønsted or Lewis acid catalyst.53 In a typical ionothermal preparation, ZnCl2 is chosen as the catalyst and high temperature (>400 °C) is always required for the reversible reaction. By contrast, a much lower temperature (<100 °C) could be employed when CF3SO3H is used as the catalyst for the condensation reaction. Alternatively, CTFs could be synthesized by condensation of an aldehyde and an amidine dihydrochloride, which involves the generation of a Schiff base followed by Michael addition. The reaction is carried out in DMSO with Cs2CO3 as a base.
Graphitic carbon nitride (g-C3N4) consisting of s-triazine or tri-s-triazine (heptazine) units has been synthesized by thermal polycondensation of nitrogen-rich precursors, such as melamine, thiourea, urea, dicyanadiamide and cyanamide, etc.18,29,31 For more detailed preparation of the above-mentioned CPs, please refer to other reviews46–53 and books.54,55
In a light-driven water splitting process, photogenerated electrons and holes were used for the reduction and oxidation of water, respectively. The photocatalytic reaction is a thermodynamically uphill reaction and a Gibbs free energy (ΔG0) of +237.2 kJ mol−1 is required.3 In principle, to drive the redox half reaction, the energy level alignments of the polymer photocatalyst and the redox potential of water must be well matched. This means that the lowest unoccupied molecular orbital (LUMO) of the polymer photocatalyst should be more negative than the reduction potential of H+/H2 (0 eV vs. normal hydrogen electrode (NHE), pH = 0), while its highest occupied molecular orbital (HOMO) must be more positive than the oxidation potential of O2/H2O (1.23 eV vs. NHE, pH = 0). Therefore, the band gap (Eg) of the polymer should be over 1.23 eV to trigger water splitting. In this part, we summarize the state-of-the-art advancements in photocatalytic H2 evolution using various types of organic CPs, and their applications in photocatalytic O2 evolution and overall water splitting are also briefly described.
As promising photocatalysts, CPs have been employed for hydrogen production in the early 1980s. Unfortunately, they haven’t attracted much attention until three-dimensional (3D) poly(azomethine) networks were demonstrated to be efficient photocatalysts for hydrogen production from water in 2010.65 Since then, various types of organic CPs have been developed to yield greatly improved H2-production efficiency. This was achieved through the optimization of CP photophysical properties, mainly light harvesting in the visible region, band gap alignment, as well as photogenerated charge generation and transport, etc.
Fig. 5 Chemical structures of polymers 1–14, (a) the correlation between the photocatalytic H2 evolution and their optical gaps of polymers 2–7 (>420 nm); (b) transient kinetics probed at 630 nm for polymers 1, 7, and 11 in the Et3N/CH3OH/H2O mixture (excited at 355 nm). Reprinted with permission from ref. 72. Copyright 2018 Nature Publishing Group. |
In 2016, Cooper and co-workers demonstrated the new breakthrough in H2 evolution by incorporating extended planarized units into the main chain of 1, which is expected to facilitate the photogenerated charge dissociation and thus promote the proton reduction reaction.70 Polymers 2–7 show band gaps ranging from 2.70 to 2.86 eV (Fig. 5a). Among them, polymers 2 and 4 exhibited an almost equal HER of ∼128 μmol h−1 g−1, which was 3-fold improvement over 1. More excitingly, polymer 7 with the lowest band gap of 2.70 eV exhibited an impressive HER of 1492 μmol h−1 g−1, which was 46-fold higher than that of 1 (Table 1).
Polymer | Optical gap (eV) | S BET (m2 g−1) | Cocatalyst | Sacrificial agent | Light sourcea | HER (μmol h−1 g−1) | AQYb (%) | Ref. |
---|---|---|---|---|---|---|---|---|
a Xe: xenon lamp. b At 420 nm. NA: no activity; AQY: apparent quantum yield; HER: hydrogen evolution rate; DEA: diethylamine; TEA: trimethylamine; TEOA: triethanolamine; AA: ascorbic acid. | ||||||||
1 | 2.9 | — | — | DEA | >400 nm (Xe) | Trace | — | 66 |
2 | 2.79 | — | — | TEA | >420 nm (Xe) | 136 | — | 70 |
3 | 2.86 | — | — | TEA | >420 nm (Xe) | 16 | — | 70 |
4 | 2.72 | — | — | TEA | >420 nm (Xe) | 128 | — | 70 |
5 | 2.78 | — | — | TEA | >420 nm (Xe) | 36 | — | 70 |
6 | 2.77 | — | — | TEA | >420 nm (Xe) | 432 | 1.1 | 70 |
7 | 2.70 | — | — | TEA | >420 nm (Xe) | 1492 | 2.3 | 70 |
8 | 1.89 | — | — | TEOA | >420 nm (Xe) | 6023 | 4.2 | 71 |
9 | 1.64 | — | — | TEOA | >420 nm (Xe) | 4270 | 0.85 | 71 |
10 | 1.94 | — | — | TEOA | >420 nm (Xe) | 3187 | — | 71 |
11 | 2.62 | — | — | TEA | >420 nm (Xe) | 3260 | 11.6 | 72 |
12 | 2.07 | — | — | TEA | >420 nm (Xe) | 5040 | 2.13 | 74 |
13 | 2.42 | 17 | — | TEA | >420 nm (Xe) | 420 | — | 76 |
14 | 2.45 | 69 | — | TEA | >420 nm (Xe) | 960 | — | 77 |
15 | 2.38 | — | — | AA | >420 nm (LED) | 8300 | ∼0.73 | 82 |
16 | 1.98 | — | — | AA | >420 nm (LED) | 50000 | ∼0.88 | 83 |
17 | 2.46 | — | — | AA | >420 nm (LED) | NA | — | 83 |
18 | 2.03 | — | — | DEA | >420 nm (LED) | 12700 | — | 84 |
19 | 2.11 | — | — | TEA | >420 nm (Xe) | 512 | — | 89 |
20 | — | — | Pt | AA | >420 nm (Xe) | 12800 | 0.13 | 90 |
21 | 2.33 | 995 | — | DEA | >420 nm (Xe) | 174 | — | 96 |
22a | 1.81 | 39 | — | TEA | >400 nm (Xe) | 17 | — | 105 |
22b | 1.95 | 59 | — | TEA | >400 nm (Xe) | 140 | — | 105 |
23a | 2.89 | 228 | — | TEA | >400 nm (Xe) | 397 | — | 105 |
23b | 2.45 | 279 | — | TEA | >400 nm (Xe) | 1460 | — | 105 |
24a | 2.31 | 121 | — | TEA | >400 nm (Xe) | 260 | — | 105 |
24b | 2.19 | 99 | — | TEA | >400 nm (Xe) | 786 | — | 105 |
25a | 2.33 | 80 | — | TEA | >400 nm (Xe) | 414 | — | 105 |
25b | 2.06 | 104 | — | TEA | >400 nm (Xe) | 1900 | — | 105 |
26 | 1.88 | — | — | TEA | >400 nm (Xe) | 2590 | — | 106 |
27 | 2.44 | 280 | Pt | TEOA | >420 nm (Xe) | 400 | — | 107 |
28 | 2.18 | 58 | — | TEOA | ≥420 nm (Xe) | 9600 | 1.8 | 110 |
29 | 2.19 | 499 | Pt | TEOA | >420 nm (Xe) | 30 | — | 111 |
30 | 2.21 | 669 | Pt | TEOA | >420 nm (Xe) | 134 | — | 112 |
31 | 2.28 | 750 | Pt | TEOA | >420 nm (Xe) | 598 | — | 112 |
32 | 2.14 | 564 | Pt | TEOA | >420 nm (Xe) | 908 | 2.0 | 112 |
33 | 2.37 | 834 | Pt | TEOA | >420 nm (Xe) | 620 | — | 112 |
34 | 2.53 | 811 | — | TEOA | >420 nm (Xe) | 2460 | — | 113 |
35 | 2.62 | 915 | — | TEOA | >420 nm (Xe) | 188 | — | 113 |
36 | 2.64 | 447 | — | TEOA | >420 nm (Xe) | 116 | — | 113 |
37 | 2.84 | 895 | — | TEA | >420 nm (Xe) | 120 | — | 114 |
38 | 2.56 | 431 | — | TEA | >420 nm (Xe) | 3106 | 13.2 | 115 |
39 | 2.37 | 723 | — | TEOA | >420 nm (Xe) | 5697 | — | 116 |
40 | 2.11 | 692.5 | Pt | TEOA | >420 nm (Xe) | 2103.2 | 6.4 | 117 |
41 | 2.8 | 1603 | Pt | TEOA | >420 nm (Xe) | 1970 | — | 119 |
42 | 2.67 | 702 | Pt | TEOA | 420–900 nm (Xe) | 22.5 | — | 120 |
43 | 2.68 | 326 | Pt | TEOA | 420–900 nm (Xe) | 90 | 0.08 | 120 |
44 | 2.62 | 1537 | Pt | TEOA | 420–900 nm (Xe) | 438 | 0.19 | 120 |
45 | 2.65 | 1046 | Pt | TEOA | 420–900 nm (Xe) | 1703 | 0.44 | 120 |
46 | 2.10 | 985 | Pt | AA | >420 nm (Xe) | 4440 | — | 121 |
47 | 1.85 | 1288 | Pt | AA | >420 nm (Xe) | 10100 | — | 121 |
48 | 2.28 | 919 | Pt | AA | >420 nm (Xe) | 1600 | — | 121 |
49 | 2.34 | 523 | Pt | TEOA | ≥395 nm (Xe) | 30 | — | 122 |
50 | 2.31 | 758 | Pt | TEOA | ≥395 nm (Xe) | 324 | 1.8 | 122 |
51 | 2.09 | 541 | Pt | TEOA | >420 nm (Xe) | 8300 | — | 123 |
52 | 2.06 | 348 | Pt | TEOA | >420 nm (Xe) | 1560 | — | 123 |
53 | 1.92 | 162 | Pt | TEOA | >420 nm (Xe) | 220 | — | 123 |
54 | 2.03 | 602 | Pt | TEOA | >420 nm (Xe) | 1360 | — | 124 |
55 | 2.94 | 19 | Pt | TEOA | ≥420 nm (Xe) | 200 | — | 128 |
56 | 2.38 | 52 | Pt | TEOA | >420 nm (Xe) | 6600 | 7.3 | 130 |
57 | 2.22 | 807 | Pt | TEOA | >420 nm (Xe) | 1238 | — | 132 |
58 | 2.13 | 764 | Pt | TEOA | >420 nm (Xe) | 1582 | — | 132 |
59 | 2.33 | 663 | Pt | TEOA | >420 nm (Xe) | 1460 | — | 132 |
60 | 2.42 | 757 | Pt | TEOA | >420 nm (Xe) | 2647 | — | 132 |
61 | 2.17 | 569 | Pt | TEOA | >420 nm (Xe) | 10760 | 4.07 | 133 |
62 | 2.47 | 533 | Pt | TEOA | >420 nm (Xe) | 5320 | 4.11 | 133 |
63 | 2.67 | 650 | Pt | TEOA | >420 nm (Xe) | 1440 | 2.10 | 133 |
64 | — | 599 | Pt | TEOA | >420 nm (Xe) | 5100 | — | 138 |
65 | — | 520 | Pt | TEOA | >420 nm (Xe) | 2400 | — | 138 |
66 | — | 598 | Pt | TEOA | >420 nm (Xe) | 650 | — | 138 |
CN–F | 2.63 | — | Pt | TEOA | >420 nm (Xe) | ∼130 | — | 142 |
CN–N | 2.65 | 9.21 | Pt | TEOA | >400 nm (Xe) | 553.5 | — | 143 |
CN–C/P | 2.98 | 141.1 | Pt | TEOA | >420 nm (Xe) | 1493.3 | 2.14 | 144 |
CN–BA | 2.57 | 22 | Pt | TEOA | >420 nm (Xe) | 294 | — | 145 |
CN–P15 | — | — | — | TEOA | >420 nm (Xe) | 722.3 | — | 152 |
CN–P17 | 2.24 | — | — | TEOA | >400 nm (Xe) | 929.3 | 5.7 | 153 |
CN–PyP | 2.46 | 120 | — | TEOA | ≥420 nm (Xe) | 600 | — | 154 |
CN–COF | — | 85.2 | Pt | TEOA | >420 nm (Xe) | 10100 | — | 155 |
CN–CTF | — | — | Pt | TEOA | ≥420 nm (Xe) | 850 | — | 156 |
Encouraged by the good activity of polymer 7, Chen et al. further modified this polymer by introducing an ethynyl group into the main chain, which led to a reduced band gap and accelerated charge separation of the resulting polymers. Polymer 8 has a band gap of 1.89 eV and showed a HER of up to 6023 μmol h−1 g−1 when triethanolamine (TEOA) was used as the sacrificial agent (>420 nm).71 Similarly, compared to their counterparts without ethynyl insertion, polymers 9 and 10 showed large red shifts of 167 and 154 nm in their light absorption edges, respectively. The HER was enhanced to 4270 μmol h−1 g−1 for 9 and 3187 μmol h−1 g−1 for 10.
Notably, dibenzo[b,d]thiophene sulfone has been recognized as one of the most attractive building blocks for the synthesis of H2-production polymer photocatalysts. As shown above, many polymers with incorporated sulfone units show outstanding photocatalytic performances. To study the critical role of the sulfone unit in H2 evolution, Cooper and co-workers investigated the photophysical properties of polymers 1, 7 and 11.72 Compared with 1 and 7, the homopolymer of sulfone 11 showed a much higher HER of 3260 μmol h−1 g−1. The decay kinetics at 630 nm exhibited a significantly larger signal for 11 than 7, while almost no signal was observed for 1 (Fig. 5b). This result indicated a higher polaron concentration of 11 and 7, which is in line with the H2 evolution activities of the three polymers. Recently, they further prepared nanoparticles of about 160 nm using polymer 11 by mini-emulsion polymerization.73 With a polymer concentration of 0.1 mg mL−1, the nanoparticle suspension yielded a HER of 14.52 mmol h−1 g−1, while a significantly lower HER of 6.13 mmol h−1 g−1 was measured for the bulk material of 11, evidently suggesting a size-dependent H2 evolution of the polymer. Moreover, when the polymer concentration was deceased to 13 μg mL−1, the suspension gave a superior average HER of 60.6 mmol h−1 g−1 over 5 h, which is the highest value among the reported organic polymers so far.
To further extend the conjugation degree of polymer 7, we introduced large planar units (fluorene and pyrene) into the main chain.74 Polymer 12 exhibited a smaller band gap of 2.07 eV in comparison with 7 (2.70 eV) and a good HER of 5040 μmol h−1 g−1 was achieved. Very recently, Wang et al. reported an analogue of polymer 12 by replacing the fluorene in the main chain with electron-rich dibenzothiophene to construct a local D–A heterojunction, which led to a reduced exciton binding energy from 91 to 88 meV and 1.3-fold HER as compared to 12 in the presence of TEOA.75
Another modification of PPP (1) is the copolymerization of 2,5-phenylene with heterocycles, which results in optimized band gaps and energy levels of the as-prepared polymers. A maximum HER enhancement of 6.5-fold compared to PPP was achieved by polymer 13 with 33% thiophene fraction and a band gap of 2.42 eV.76 Polymer 14 bearing a 2,5-pyrazine attained a 15-fold higher HER than PPP.77
Compared to those insoluble bulk polymers, the development of soluble CPs provides a good choice for easy processing of polymers into a desirable pattern for photocatalytic H2 production. By virtue of its inherent hydrophobicity, CP could be self-assembled into polymer dots (Pdots) in water.78–80 Pdots have much smaller size relative to the bulk materials; hence, the distance covered by photogenerated charges migrating to the surface of photocatalyst particles becomes short, and the recombination probability of electrons and holes decreases to improve the photocatalytic efficiency.81 In 2016, Tian et al. prepared a Pdot suspension by the nano-precipitation method using polymer 15 and a matrix polymer PS–PEG–COOH.82 In the preparation, the mixture of polymer and matrix in THF was injected into pure water under sonication, and a clear Pdot suspension was obtained after THF evaporation (Fig. 6). The as-prepared Pdots displayed an impressive initial rate of 8300 μmol h−1 g−1. Compared to polymer 15, Pdots of polymer 16 with extended absorption renders an excellent initial HER of up to 50 mmol h−1 g−1.83 The theoretical calculations suggested that N atoms in the 2,1,3-benzothiadiazole (BT) units provided the reactive sites in the formation of hydrogen, polymer 17 without a BT unit did not show any activity although its optical gap and reduction potential are appropriate for proton reduction. Using the nano-precipitation method, Chou et al. prepared cycloplatinated Pdots using polymer 18, which provided an impressive HER of 12.7 mmol h−1 g−1, while only 1.3 and 5.46 mmol h−1 g−1 were obtained for Pdots without a Pt complex and their Pt-complex-blended counterpart under identical conditions, respectively.84 This was likely attributed to the improved charge transfer after grafting the Pt complex into the polymer backbone.
Fig. 6 Chemical structures of polymers 15–20 and preparation of polymer dots (Pdots) by the nano-precipitation method for photocatalytic H2 production. The scheme is adapted with permission from ref. 82, Copyright 2016 Wiley-VCH. |
Conjugated polyelectrolytes (CPEs) are a subclass of interesting water-soluble CPs with wide applications in photovoltaic devices, biomedical imaging and therapy, as well as biosensing, etc.85–88 Their rigid conjugated structures endow them with excellent optoelectronic properties, while the hydrophilic side chains with charged groups allow for unique dispersibility of CPEs in polar solvents. We prepared cationic CPE 19 by one-step quaternization of the precursor polymer attaching a hexyl bromide side chain.89 In the mixture of Et3N/CH3OH/H2O, polymer 19 could self-assemble into homogeneous nanoparticles of ∼180 nm. The suspension gives a HER of 512 μmol h−1 g−1, which is 3-fold that of its counterpart without the side chain. The enhanced HER was attributed to the improved light absorption and charge separation of the nanoparticles.
Different from the cationic CPEs, attaching oligoethylene glycol (OEG) side chains endows polymer 20 with more interesting functions.90 The O atom in the OEG side chain favours the interface contact of the polymer backbone with water, which resulted in a much smaller water contact angle of 20 (27.5°) than that of an alkyl-functionalized polymer (101.2°). Moreover, the side chains could interact with Pt, resulting in more efficient charge transport from the polymers to Pt. Polymer 20 loaded with 3 wt% Pt gave a HER of 12.8 mmol h−1 g−1, which is much higher than its alkyl-attached counterpart.
Besides, soluble CPs could also be processed into films for photocatalytic H2 production. By drop-casting, Cooper's group prepared a film of polymer 4 attaching a 2-ethylhexyl side chain. The film demonstrated a HER of 72 μmol h−1 g−1 when immersed in the Et3N/CH3OH/H2O mixture (>420 nm).91
The statistical copolymerization of aromatic monomers was reported as an important strategy to enhance the catalytic activity of CPPs by improving their physical properties such as band gap, porosity and charge mobility, etc. For instance, Cooper and co-workers synthesized fifteen conjugated microporous polymers (CMPs) by adjusting the feed ratio of the four comonomers shown in Fig. 7.96 The CMPs have high surface areas of 597–1710 m2 g−1 and their band gaps gradually decrease from 2.95 to 1.94 eV with the increase of pyrene content in the polymer. Without adding a metal cocatalyst, the polymers showed gradually enhanced HER when the band gap decreased from 2.95 to 2.33 eV. With the ratio of 1,4-benzene diboronic acid and 1,2,4,5-tetrabromobenzene of 2:1, polymer 21 exhibited the highest HER of 174 μmol h−1 g−1. However, further narrowing the band gap of polymer led to a decrease in the H2 production. This was likely ascribed to the increased nonradiative recombination of electron and hole in the pyrene-rich polymers.
Similarly, Jiang et al. synthesized D–π–A CPPs using pyrene as a donor, benzothiadiazole as an acceptor, and benzene or biphenyl as a π-crosslinker unit.97 Compared to D–A polymers (without π unit), the extended π-conjugation endows the D–π–A polymers with more efficient charge migration to the surface of polymer particles. In addition, the electron push–pull interaction in the D–π–A polymer backbone could facilitate charge separation in the photocatalytic process. Thereby, D–π–A polymers exhibited better hydrogen production than the D–A and D–π–D polymers (without acceptor). When the ratio of pyrene to benzothiadiazole was 9:2, a maximum HER of 296 μmol h−1 g−1 was obtained (>420 nm). Xiang et al. developed a series of CPPs by copolymerizing 1,3,6,8-tetrabromopyrene and 3,8-dibromophenanthroline. A remarkable HER of 4200 μmol h−1 g−1 was achieved by the polymer having a 3:1 ratio for 1,3,6,8-tetrabromopyrene and 3,8-dibromophenanthroline.98 In addition, Wang et al. further modified polymer 21 by gradually replacing the benzene units in the polymer network with dibenzo[b,d]thiophene sulfone units.99 They claimed that the sulfone unit could act as an electron-output “tentacle” to trap photogenerated electrons from polymers, which is beneficial for the proton reduction reaction. The HERs of the resulting polymers were significantly enhanced with increasing content of sulfone unit in the polymer. This molecular modification provided a maximum HER of 400 μmol h−1.
The donor–acceptor combination approach is one of the most prevailing methods used in polymer solar cells to enhance the power conversion efficiency.100–104 The alternating electron-rich donor (D) and electron-deficient acceptor (A) in the main chain could induce intramolecular charge transfer (ICT) interactions between D and A units, resulting in a narrowing of the band gap and molecular energy level modulation. Owing to this, the D–A approach has attracted a great deal of interest in designing CPs for photocatalytic H2 evolution.
In 2016, Yu et al. reported CPPs by the combination of various chromophores (M1–M4) with biphenyl and bipyridyl, respectively.105 With the chromophores from strong acceptors (M1) to strong donors (M4), the band gaps could be fine-tuned from 1.81 to 2.89 eV for 22a–25a. The replacement of biphenyl (weak donor) with bipyridyl (weak acceptor) led to lower band gaps of 2.06–2.45 eV for 23b–25b (Fig. 8a). The overall H2 evolution of bipyridyl containing polymers is better than that of biphenyl containing polymers (Fig. 8b). Polymer 22b gave the highest HER of 1900 μmol h−1 g−1 (>400 nm) as a result of the improved light absorption, wettability and charge separation. Later, a higher HER of 2590 μmol h−1 g−1 was obtained by polymer 26 through further optimizing nitrogen-containing acceptors in M4 based polymers.106
Fig. 8 Chemical structures of D–A type CPPs and (a) the optical gaps as well as (b) photocatalytic H2 production of polymers 22–25 (P-a: 22a–25a; P-b: 22b–25b). Reprinted with permission from ref. 105. Copyright 2016 American Chemical Society. |
2,1,3-Benzothiadiazole (BT) is a strong acceptor unit used in designing D–A copolymers. Wang and co-workers demonstrated that the substitution position of BT on the central phenyl unit has a profound effect on the porosity and morphology of the resulting polymers.107 CPPs with 1,2,4,5 and 1,2,4-substitution exhibited a fused particle-like shape and a fibre shape was observed for their counterpart with 1,3,5-substitution, while the BET surface areas of the three polymers ranged from 40 to 280 m2 g−1 with optical gaps of 2.25–2.44 eV. Consequently, polymer 27 loaded with 3 wt% Pt yielded a HER of 400 μmol h−1 g−1, while 8 μmol h−1 g−1 and 26 μmol h−1 g−1 were observed for the polymers with BT at the 1,2,4 and 1,3,5-positions of the central phenyl ring, respectively. Interestingly, compared to those 3D CPPs, a 1D linear polymer consisting of alternating phenyl and benzothiadiazole units showed a much higher HER of 2320 μmol h−1 g−1, which was ascribed to its narrowest band gap (2.17 eV) and improved charge transport in the polymer.
Fluorine (F) substitution on the CP backbones is an effective way to engineer their optoelectronic properties.108 The van der Waals radii of F atom is only 1.35 Å, which is close to that of the hydrogen atom (1.2 Å).109 Besides, with a Pauling electronegativity value of up to 4.0, the F atom tends to impair the electron density in the D–A type CPs. As a consequence, the introduction of F does not cause a significant steric hindrance, but has a profound influence on the molecular orbits of the polymers.
Zhu and co-workers studied the F substitution effect in the case of BT incorporated linear CPs and conjugated porous polymers on their photocatalytic activities.110 It was found that F substitution had a weak influence on the optical gaps of polymers but the charge transfer efficiency was improved significantly. With the increase of F atom number in BT from 0 to 2, the HOMO and LUMO levels of polymers decrease simultaneously. Polymer 28 incorporating CH3O and F substituted BT units showed a HER of 9.6 mmol h−1 g−1, which was 28.8-fold as large as that of its counterpart without fluorination. The F substitution was also examined by Wang and co-workers.111 They found that the F substitution on BT led to a more negative LUMO level, which could increase the driving force for proton reduction. In this work, polymer 29 showed a 5-fold enhancement of HER compared to its non-fluorinated counterpart.
In general, extending the π-conjugation along the skeleton had multiple impacts on the morphologies, porous properties and photophysical properties of the resulting polymers, which have a synergistic effect on the photocatalytic performances of the CPs. For example, Zhang and co-workers examined the π-extended conjugation effect of polymers 30–33 on their H2 production performances.112 The spacer change led to a profound effect on the porosity of the polymers. Polymer 30 with a small para-phenylene spacer has micropores, while both micro- and mesopores were found for 31 and 33 with a longer spacer. By contrast, polymer 32 showed a much broader pore size distribution, which was due to the longest polyphenylene spacer in the polymer network. As a result, the BET surface areas were 669, 750, 564, and 834 m2 g−1 for 31, 32, 33, and 34, respectively. In addition, the four polymers showed different energy level alignments due to the donor (arene units) and acceptor (tricyanomesitylene) interactions (Fig. 9a). The HERs were found to be 134, 598, 908, and 620 μmol h−1 g−1 for 30, 31, 32, and 33, respectively. Despite the similar band gap, the highest photocatalytic activity of 33 as compared to 30 and 31 was ascribed to its nanoparticle morphology and the largest surface area.
Fig. 9 Chemical structures of polymers 30–40. (a) HOMO and LUMO band positions of polymers 30–33. (b) Nitrogen sorption isotherms, (c) photoluminescence spectra, and (d) hydrogen evolution of polymers 34–36. Reprinted with permission from ref. 112 and 113. Copyright 2017 Wiley-VCH and 2016 American Chemical Society. |
Similarly, Jiang et al. developed polymers 34–36 by changing the central π-linker from benzene to p-terphenyl.113 Because of the higher steric hindrance of the biphenyl than phenyl, 34 showed a smaller BET surface area of 811 m2 g−1 than 35 (915 m2 g−1), while 36 with the longest linker showed the lowest surface area of 447 m2 g−1 (Fig. 9b). Moreover, the introduction of the extended linker seemed to favour the formation of larger pores. The three polymers showed a similar pore distribution but more mesopores or macropores were found for 35 and 36 (Fig. 9c). In addition, polymer 34 has a smaller band gap of 2.53 eV than 35 (2.62 eV) and 36 (2.64 eV). Polymer 34 exhibits better charge separation efficiency as evidenced by a much lower emission than that of 35 and 36 (Fig. 9d). Accordingly, polymer 34 yielded the highest HER of 2460 μmol h−1 g−1, while HERs of 188 and 116 μmol h−1 g−1 were obtained by 35 and 36, respectively.
Cooper et al. found that the optical gaps of the CMPs could be tuned by changing the monomer linker length and type of linkage, which led to different H2 evolution activities.114 With spirobifluorene as the core, polymer 37 exhibited a 15-fold HER as compared to that of the analogue with phenyl as the core. In addition, by extending the central core from phenyl to spirobifluorene, polymer 38 was reported with an improved HER of 3106 μmol h−1 g−1.115
Jiang et al. demonstrated that the substitution position of dibenzo[b,d]thiophene sulfone had a significant influence on the photocatalytic performances.116 They prepared two CMPs by linking pyrene at the 3,7 and 2,8 positions of sulfone, respectively. Polymer 39 with the 3,7-linking pattern showed a narrower optical gap and lower emission intensity than the other polymer, which indicated better photogenerated charge migration and thus led to a much higher HER of 8523 μmol h−1 g−1 than that of the 2,8-linked polymer (2650 μmol h−1 g−1). Similarly, Su and co-workers found that substitution positions of CN groups and carbazole on the central phenyl ring could affect spectral absorption and charge separation of the resulting polymers. With an optimized structure, polymer 40 showed an excellent HER of 2103.2 μmol h−1 g−1 and 6.4% apparent quantum yield at 420 nm.117
In an early contribution, Lotsch et al. synthesized COF 41 (TFPT-COF) by a condensation reaction of 2,5-diethoxy-terephthalohydrazide and 1,3,5-tris(4-formylphenyl)triazine in a 1,4-dioxane/mesitylene mixture.119 COF 41 showed a honeycomb-type layered structure with mesopores of 3.8 nm and its BET surface area was up to 1603 m2 g−1. Pt-loaded COF 41 could produce H2 from water at a constant rate of 1970 μmol h−1 g−1. Subsequently, they reported a series of azine-linked COFs 42–45 (Nx-COFs, x = 0–3) with different N numbers in the central aryl ring (Fig. 10).120 The molecular design is based on the varied dihedral angles between the central aryl ring and peripheral phenyl rings by replacing the H atom (green dots) with a N atom (Fig. 11a), which led to different planarity of the platform and thus varied crystallinity and porosity of the resulting COFs. As displayed in Fig. 11b, compared to 42 (N0-COF) and 43 (N1-COF), 44 (N2-COF) and 45 (N4-COF) have better crystallinity as revealed by their sharper PXRD peaks, which is beneficial for charge migration in the photocatalytic reaction. Besides, the four COFs showed close band gaps of 2.6–2.7 eV, but different BET areas of 326 to 1536 m2 g−1. The HERs of Pt-modified 42, 43, 44 and 45 were 23, 90, 438 and 1703 μmol h−1 g−1, respectively. The progressively enhanced HER was likely attributed to the increased surface area and improved charge migration when the number of N atoms was increased in the frameworks.
Fig. 11 (a) Design principle and (b) PXRD patterns of COFs 42–45. (c) Diffuse reflectance spectra of COFs 46–48 (46a is an amorphous analogue of 46). (d) Transient photocurrent for COFs 49 and 50 measured in 0.5 M KCl solution under visible light irradiation. Reprinted with permission from ref. 120–122. Copyright 2015 Nature Publishing Group, 2018 Nature Publishing Group and 2018 American Chemical Society. |
The facile choice of building blocks did permit a good modulation of the porosity, wettability and optoelectronic properties of β-ketoenamine linked COFs, which led to enhanced photocatalytic performance. When different building blocks were introduced, the corresponding band gaps were 2.10, 1.85 and 2.28 eV for 46 (F-COF), 47 (FS-COF) and 48 (TP-COF), respectively (Fig. 11c).121 Moreover, 46 and 47 containing a sulfone unit exhibited better wettability than 48 as revealed by the water contact angle measurement, which facilitates dispersion of COF particles in water and the interaction between COF and water in the photocatalytic reaction. Loaded with 3 wt% Pt, the HER of 46 was up to 4.44 mmol h−1 g−1, while 10.1 and 1.6 mmol h−1 g−1 were reached for 47 and 48, respectively. The best H2 evolution of 47 was ascribed to its lowest band gap, good wettability, highest BET area (985 m2 g−1 for 46, 1288 m2 g−1 for 47 and 919 m2 g−1 for 48).
In addition, the photocatalytic activity of COFs is susceptible to linking bond length between comonomers. Thomas and co-workers reported that the catalytic activity of the diacetylene-based COF 49 (TP-EDDA) was much better than the acetylene-based COF 50 (TP-BDDA).122 Despite the almost identical band gaps (2.34 eV for 49 and 2.31 eV for 50), COF 50 has an HER of 324 μmol h−1 g−1, whereas a relatively low HER of 30 μmol h−1 g−1 was achieved for 49. The significantly enhanced HER was attributed to the higher charge carrier generation of 50 under visible light, which could be revealed by the larger photocurrent response of 50 as shown in Fig. 11d.
Substituent modification in the COF could also bring good photocatalytic performances. Sun et al. prepared three ketoenamine-based COFs (51–53) and studied the effect of different functional groups on their photocatalytic activities.123 The photocurrent density of COF 53 (TpPa-COF-(CH3)2) is about 2.3-fold and 6.5-fold higher than that of 52 (TpPa-COF-H) and 51 (TpPa-COF-NO2), respectively. Moreover, the electron transfer resistance value decreased gradually when the electron-donating ability of the substituent groups became stronger (from NO2 to H, then CH3). These results indicated that the charge separation of COFs was improved by increasing the electron-donating properties of the substituents. With an optical gap of 2.06 eV and the best charge separation efficiency, COF 52 showed a HER of 8.33 mmol h−1 g−1, which was much higher than that of 51 (1.56 mmol h−1 g−1) and 53 (0.22 mmol h−1 g−1).
In comparison with those reported COFs based on imine-, hydrazone-, or azine linkers, fully π-conjugated COFs are expected to allow for more efficient exciton separation and migration over the framework, which is good for H2 production. Very recently, Jiang and co-workers reported a 2D sp2 carbon-COF 54 (sp2c-COF) as the H2-producing photocatalyst, into which electron-withdrawing cyano groups were incorporated to impart a conjugated electron donor–acceptor structure and the phenylenevinylene further improved π conjugation.124 The BET surface area of COF 54 was 602 m2 g−1 and its optical gap was 2.05 eV. A stable HER of 1360 μmol h−1 g−1 was obtained under visible light. The HER could be enhanced to 2120 μmol h−1 g−1 by attaching 3-ethylrhodanine as an end-capping group to strengthen the push–pull effect.
In 2015, Wu and co-workers prepared CTF 55 (CTF-T1) by a trimerization reaction of terephthalonitrile using CF3SO3H as the catalyst at room temperature (Fig. 12).128 The CTF exhibits a band gap of 2.94 eV and a low HER of 200 μmol h−1 g−1 after loading 3 wt% Pt.
Afterward, Thomas et al. showed that a CTF with a low HER could be converted into a more efficient photocatalyst by further mixing the pre-synthesized CTF in CHCl3/CF3SO3H with ZnCl2 salt for 10 minutes. The HER could increase from 186 to 1072 μmol h−1 g−1 due to the decreased optical gap (2.42 eV) and improved crystallinity.129 Alternatively, Wang et al. further processed the pre-synthesized CTF 55 through thermal annealing at 300 °C, which induced structural reorganization of 55 and yielded new conjugated heterocycles in the polymer. With improved light absorption and charge mobility, 3 times enhancement of HER was achieved under visible light as a result of the thermal treatment of 55.130
The fast recombination of photogenerated electrons and holes of CTFs is another major disadvantage preventing them from achieving higher efficiencies. In this context, Li et al. developed a molecular heterostructure 56 (CTF-Th) via a two-step synthetic strategy.131 CTF 56 exhibited higher photocurrent and charge carrier density than its single-component counterparts, which suggested a largely improved charge separation efficiency. As a result, a HER of 6.6 mmol h−1 g−1 was obtained for 56 under visible light.
Compared to the commonly used ionothermal method (usually at 400 °C), polycondensation under relatively gentle conditions is desirable for the preparation of CTFs with tunable functions and geometries. Tan and co-workers developed CTFs 57–60 by condensation between 1,4-benzenedicarboximidamide and different aldehydes in the presence of Cs2CO3.132 The structural variations have an obvious impact on the porosity and light absorption of the resulting CTFs. The BET surfaces areas are 807, 764, 663, and 757 m2 g−1 for 57 (CTF-HUST-3), 58 (CTF-HUST-3), 59 (CTF-HUST-1) and 60 (CTF-HUST-2), respectively (Fig. 13a). CTFs 57 and 58 showed broader absorption compared to 59 and 60 (Fig. 13b). Under visible light, CTF 60 showed the highest HER of 2647 μmol h−1 g−1, while the HERs were 1238, 1582 and 1460 μmol h−1 g−1 for 57, 58 and 59, respectively. Using a similar synthetic method, Tan et al. prepared donor–acceptor CTFs 61–63.133 Upon incorporating a strong electron-donating carbazole unit, 61 (CTF-N) showed the lowest band gap of 2.17 eV compared to 62 (CTF-S, 2.47 eV) and 63 (CTF-O, 2.67 eV) (Fig. 13c). Meanwhile, the enhanced charge separation/transfer efficiency of 61 was studied by employing fluorescence and photoelectrochemical measurements (Fig. 13d). Therefore, CTF 61 decorated with 3 wt% Pt exhibits the best HER of 10760 μmol h−1 g−1.
Fig. 13 (a) Nitrogen sorption curves and (b) UV-vis absorption spectra of CTFs 57–60. (c) Energy diagram and (d) fluorescence spectra of CTFs 61–63. Reproduced with permission from ref. 132 and 133. Copyright 2017 Wiley-VCH and 2018 Royal Society of Chemistry. |
CTFs could also be fabricated by a Pd-catalyzed coupling reaction. In 2017, Cooper et al. synthesized two series of CTFs with similar structures by either Suzuki coupling of 2,4,6-tris(4-bromophenyl)-1,3,5-triazine and aromatic monomers or trimerization of dicyanobenzene derivatives.134 In this case, Pt loaded CTFs from trimerization exhibited higher activity than their counterparts from Suzuki coupling. Moreover, the length of the para-phenylene spacer has a profound effect on the band gaps and porosity for both series of CTFs and thus affects their photocatalytic activities. Based on the D–A interaction between thiophene derivatives (donor) and triazine units (acceptor), Bojdys et al. designed eight CTFs with fine-tuned band gaps (2.06–2.70 eV) and charge transfer in the polymer.135 As a result, a maximum HER of 3158 μmol h−1 g−1 was achieved by the CTF consisting of benzotrithiophene and triazine.
Instead of introducing different building blocks into CTFs, heteroatom doping is another efficient method to boost the HERs of CTFs. Su et al. prepared S-doped CTF samples (CTFS10) by annealing treatment of CTFs with sulfur.136 The S concentration in the framework was 0.5–0.52 atom% when the CTF was mixed with different amounts of S (5–30 wt%). The characteristic peak of C–S at 286.5 eV revealed by XPS analysis showed that the S atom is covalently bound to the CTF backbone. After depositing 3 wt% Pt, the CTFs mixed with 10 wt% S showed the best HER of 2000 μmol h−1 g−1, which is 5-fold enhanced compared to that of the pristine CTFs. This is due to the sharply decreased band gap (from 2.87 eV to 1.87 eV) and improved charge transport after S doping. Alternatively, the photocatalytic activity of CTFs could be enhanced by phosphorus doping due to the narrower band gap and more efficient charge transfer.137
Besides the above-mentioned amorphous CTFs, Tan and co-workers reported the successful preparation of crystalline CTFs 64–66 by in situ oxidizing alcohols into aldehyde monomers and studied their H2 production activities.138 It was claimed that high crystallinity of CTFs could be achieved by slowing down their nucleation process through controlling aldehyde monomer generation in the polycondensation reaction. All CTFs showed good crystallinity with BET surface areas of 520–599 m2 g−1. The HERs of 64 (CTF-HUST-C1), 65 (CTF-HUST-C5) and 66 (CTF-HUST-C6) were 5100, 2400 and 650 μmol h−1 g−1, respectively. In addition, Zou and co-workers prepared crystalline CTFs by condensation of melamine (MA) and pyromellitic dianhydride (PMDA) under thermal conditions. The obtained CTF displayed a band gap of 2.7 eV and HER of 7 μmol h−1.139
The elemental doping strategy has been intensively developed to boost the photocatalytic performance of g-C3N4 by tuning its band structure, light absorption and charge carrier transport.27–32 Various elements such as F, Br, I, N, P, O, S and B have been doped in the skeleton of tri-s-triazines based g-C3N4, as shown in Fig. 14.141 For instance, Wang et al. prepared F doped g-C3N4 (CN–F) by adding NH4F in the polymerization of dicyanamide. The formation of C–F bonds in the polymer led to a slightly red-shifted optical absorption and 2.7 times higher HER (∼130 μmol h−1 g−1) than that of the unmodified sample (>420 nm).142 Nitrogen-doped g-C3N4 (CN–N) could be synthesized using pre-treated melamine with hydrazine hydrate.143 This modification gave rise to improved charge transfer and a slightly narrowed band gap from 2.72 eV to 2.65 eV, as a result, CN–N showed a 1.8-fold higher HER of 553.5 μmol h−1 g−1 than that of pristine g-C3N4 (>400 nm). On the other hand, the doping of multiple elements in g-C3N4 provided more variable properties.141 Dai et al. reported carbon and phosphorus co-doped g-C3N4 (CN–C/P), which showed a significantly larger surface area of 141.1 m2 g−1 than pristine g-C3N4 (24.9 m2 g−1).144 Moreover, this modulation also led to extended optical absorption and promoted charge transfer in the polymer. Benefiting from this, CN–C/P exhibited a HER of 1493.3 μmol h−1 g−1, which was 9.7-fold as compared to that of pristine g-C3N4.
Fig. 14 Possible elemental doping sites in the single layer of g-C3N4. Reproduced with permission from ref. 141. Copyright 2017 Royal Society of Chemistry. |
Organic small molecules are also capable of doping into g-C3N4 by copolymerization to adjust its π-electronic system and band structure. In 2010, Wang and co-workers prepared modified g-C3N4 (CN–BA) using dicyandiamide and barbituric acid as the precursor and dopant, respectively.145 With an optimization of amount of dopants, CN–BA exhibited a maximum HER of 294 μmol h−1 g−1 mainly due to the extended light absorption. After that, many other small molecules such as pyrimidine,146 quinoline,147 and thiophene148 were incorporated into g-C3N4 by several groups.
Structure defect engineering is another efficient approach for improving the photocatalytic performance of g-C3N4. The introduction of defects into the g-C3N4 backbone can not only modify the band structure by generating mid-gap states to extend the light absorption, but also act as trapping sites for charge carriers to inhibit the recombination of electrons and holes. In the presence of KOH, Zhang et al. introduced nitrogen vacancies and cyano groups into g-C3N4, which resulted in a reduced band gap from 2.68 to 2.36 eV and enhanced charge carrier separation.149 Therefore, the defect-rich g-C3N4 showed a HER of 6.9 mmol h−1 g−1, which was nearly twice that of pristine g-C3N4. In a very recent study, Wang et al. introduced defects into g-C3N4 by NaBH4 reduction of a crystalline pristine polymer under an inert atmosphere, which achieved nearly 8 times higher HER than the pristine polymer.150 Apart from those modification strategies, the development of analogous polymers of g-C3N4 such as C3N3S3151 offered new opportunities for exciting findings in photocatalytic H2 production.
In some cases, the construction of polymer heterojunctions (PHJs) can efficiently modulate the band structure and charge transport of g-C3N4 through inter/intramolecular π–π interactions between g-C3N4 and the aromatic structures. Shen et al. prepared three PHJs by coating different polyfluorene polymers on the surface of g-C3N4, which resulted in extended optical absorption and promoted charge transfer within the PHJs.152 Under visible light, all three pure CPs have no photocatalytic activity and pure g-C3N4 showed a very low HER of 87.4 μmol h−1 g−1. Among the three PHJs, 5% polymer 15 coated g-C3N4 (CN–P15) exhibited a maximum HER of 722.2 μmol h−1 g−1. Using a similar method, Chen et al. deposited Pdots on the surface of g-C3N4 nanosheets, which yielded an enhanced HER via improved charge transfer from Pdots to g-C3N4 nanosheets.153 In this case, 5 wt% polymer 17 deposited PHJ (CN–P17) exhibited a HER of 929.3 μmol h−1 g−1, which is higher than that of its counterpart coated with polymer 15 (439.6 μmol h−1 g−1). Besides, Wang et al. fabricated type II heterojunctions by anchoring pyrene-based polymers onto CN to improve light absorption and interface charge transfer.154 By optimizing the amount and molecular structures of polymers, CN–PyP hybrids attained a maximum HER of 600 μmol h−1 g−1 without Pt deposition. PHJs based on g-C3N4 could also be prepared by linking another component through a chemical bond. Yan et al. prepared heterojunctions (CN–COF) by in situ formation of a COF on the surface of pre-synthesized g-C3N4 through condensation of amino group in g-C3N4 and aldehyde group in the COF.155 CN–COF attained a dramatically enhanced HER of 10.1 mmol h−1 g−1, which is much higher than those of the pure COF (0.16 mmol h−1 g−1), g-C3N4 (1.05 mmol h−1 g−1) and their mixture (0.93 mmol h−1 g−1). Zou and co-workers fabricated heterostructured photocatalysts (CN–CTF) by linking g-C3N4 and CTFs with amide bonds, which demonstrated a good HER of 850 μmol h−1 g−1 and high stability compared to their counterpart (CN and COFs connected via van der Waals forces).156 These investigations offer more opportunities for the design of H2-production polymer photocatalysts. The photocatalytic activity and stability of these materials remain a great concern.
Among the commonly used metal cocatalysts such as Ru, Pd, Rh, Au, and Ag, Pt is a versatile cocatalyst mainly because of its largest work function for trapping photogenerated electrons and lowest overpotential for photocatalytic H2 evolution.158 In general, Pt is deposited on CPs by photoreduction of H2PtCl6 in the reaction system. Under light illumination, Pt(VI) is reduced by the photoinduced electrons of the polymer to produce Pt(0) nanoparticles with sizes of 3–5 nm in situ on the polymer surface. Compared to the pristine polymer, the Pt–polymer composite exhibited greatly enhanced H2 evolution. H2PtCl6 with a Pt mass concentration of 3 wt% was commonly used in the photodeposition experiment, because too low Pt concentration provided only a few active sites for proton reduction, which resulted in a low enhancement of catalytic activity. Excessive amount of Pt deposited on the surface of the polymer photocatalyst could prevent light absorption by shielding the incident light and stimulate the aggregation of Pt nanoparticles, which is detrimental to the activity of Pt.
Pd-catalyzed cross-coupling polymerization reactions are widely used to synthesize CPs for photocatalytic water splitting. Residual Pd inevitably existed in the polymer and may play a vital role in its photocatalytic process.160 Yu et al. have shown that the Pd content in the porous polymer greatly affected the H2 evolution.105 By changing the initial feeding ratio of Pd(PPh3)4 and monomer from 0.005 to 0.40, the Pd content was increased from 0.04% to 1.88% as measured by inductively coupled plasma mass spectrometry (ICP-MS). As shown in Fig. 15a, the HER of polymer 25b was sharply improved with increasing Pd content from 0.04% to 0.46%, which became saturated at a higher residual Pd content. In addition, upon loading with 2 wt% Pt, a ∼5-fold enhancement of the AQY at 350 nm was measured for 25b (Fig. 15b), which verified the importance of the Pt cocatalyst in the enhancement of photoactivity.
Fig. 15 (a) H2 evolution of polymer 25b with different residual Pd contents. (b) Apparent quantum yield (AQY) of polymer 25b with 2% and without Pt loading at different wavelengths. (c) H2 evolution of nanoparticles of polymer 15 with different amounts of Pd or Pt. (d) Transient kinetics of nanoparticles of polymer 15 with different Pd concentrations in a diethylamine/water mixture. Reprinted with permission from ref. 105 and 161. Copyright 2016 American Chemical Society and 2018 Wiley-VCH. |
To study the role of Pd residues in determining the H2 evolution activity, Jan Kosco and co-workers prepared colloidal polymer 15 nanoparticles with different Pd contents.161 In the sample preparation, they first purified the as-prepared polymer 15 by gel permeation chromatography (GPC), and then washed the polymer with diethyldithiocarbamate. After the treatment, the Pd content in polymer 15 was below 1 ppm and the content could be gradually adjusted to 1170 ppm by adding Pd2(dba)3 solution. Fig. 15c displays that the H2 evolution of nanoparticles of polymer 15 is gradually enhanced with increasing Pd content from <1 ppm to ∼100 ppm, at which point the evolved H2 begins to saturate. Transient absorption spectroscopy (TAS) suggested that Pd could mediate electron transfer from the nanoparticles of polymer 15 to protons in the reaction media (Fig. 15d). It was found that photogenerated electrons were increasingly accumulated with decreasing Pd content in nanoparticles of polymer 15. This indicates that the Pd cocatalyst could effectively suppress the accumulation of long-lived electrons and thereby accelerate photocatalytic H2 generation of polymer 15.
To date, most of the O2 evolution photocatalysts have been based on nitrogen-rich CPs. In 2017, Xu et al. reported polymer 67, which has a nanosheet morphology and a very low band gap of ∼1.22 eV.41 With AgNO3 as the electron scavenger, 67 exhibits an oxygen evolution rate (OER) of 1.0 μmol h−1, which could be enhanced to 14.3 μmol h−1 after loading Co(OH)2 as the cocatalyst. More interestingly, polymer 67 can catalyze water oxidation under near-infrared light (>800 nm) and the OER is ∼0.4 μmol h−1 while g-C3N4 is not active.
Inspired by the facile tunability of CTFs, Wang et al. synthesized polymers 68–70 with different numbers of phenyl units in the polymer backbone.164 The band gaps of 68–70 decreased from 2.98 eV to 2.36 eV with increasing number of phenyl units (Fig. 16a). Importantly, their HOMO/LUMO levels straddle the potential of water splitting, suggesting sufficient driving force for both photocatalytic water reduction and oxidation. Furthermore, both the HOMO and LUMO levels decreased with the increase of phenyl number in the polymer. After loading with 3 wt% Co(NO3)2 as a cocatalyst, polymer 69 showed the highest OER of 3 μmol h−1 (>300 nm) compared to 68 (∼1 μmol h−1) and 70 (∼0.4 μmol h−1) in the presence of AgNO3, which was attributed to the synergistic effect of improved light absorption and charge transfer. In addition, the catalytic activity of 69 is also much better than that of C3N4 (Fig. 16b) due to the lower LUMO position of 69. Notably, polymer 69 could produce 7.2 μmol of O2 after 8 h visible-light irradiation. Considering that high crystallinity of CTF materials is desirable for exciton separation and thus enhancing the photocatalytic activity, Tang and co-workers reported highly ordered CTFs made by microwave-assisted polymerization.165 The crystallinity of the resulting CTFs became higher with the increase of microwave power from 20 to 100 W. However, a further increase in microwave power can destroy the hexagonal units. At a power of 100 W, polymer 71 has a well-defined planar structure as evidenced by a G+ peak at 1613 cm−1 from the Raman spectra analysis of CTFs. In addition, polymer 71 has conduction band (CB)/valence band (VB) levels of −0.74/+2.18 eV (vs. NHE, pH = 0), which is suitable for the photocatalytic water oxidation and reduction reaction. Polymer 71 gave an O2 production rate of 140 μmol h−1 g−1 upon loading with 3 wt% RuOx in the presence of AgNO3, which is higher than those of its counterparts prepared under different powers.
Fig. 16 Chemical structures of polymers 67–71; (a) band structure diagram and (b) time course of the O2 production for polymer 69 and C3N4. Reprinted with permission from ref. 164. Copyright 2018 Wiley-VCH. |
In 2016, Wang et al. demonstrated that Pt modified g-C3N4 was an efficient photocatalyst for overall water splitting and the morphology of g-C3N4 significantly affected the catalytic activity.169 Among the three types of g-C3N4, the urea-derived g-C3N4 exhibits a 2D nanosheet structure and compact layer-like morphology, which showed better charge mobility as compared to the other two samples made from dicyandiamide and ammonium thiocyanate as the precursor. All three types of pure g-C3N4 have no photocatalytic activity in the absence of Pt. After depositing Pt on the three samples, g-C3N4 nanosheets exhibited the best activity with H2 and O2 evolution rates of 1.2 and 0.6 μmol h−1, respectively (>420 nm). Very recently, Zhang and co-workers prepared 3D g-C3N4 with high crystallinity, displayed a larger surface area (130 m2 g−1) than those of bulk g-C3N4 (10.83 m2 g−1) and g-C3N4 nanosheets (93.84 m2 g−1).170 Surprisingly, after loading 1 wt% Pt and 3 wt% IrO2, 3D g-C3N4 exhibited H2 and O2 generation rates of 101.4 and 49.1 μmol h−1 g−1, respectively, which are about 11.8 and 5.1-fold higher than those of bulk g-C3N4 and g-C3N4 nanosheets. These examples demonstrated the critical role of the cocatalyst and surface morphology control of g-C3N4 in photocatalytic overall water splitting performance.
In 2017, Xu and co-workers prepared 1,3-diyne-linked CMPs 72–73 by Glaser oxidative coupling of terminal alkynes.40 TEM measurements revealed that both polymers have sheet-like structures (Fig. 17a and b), which favours the separation and transport of photogenerated excitons to the surface of polymer nanosheets for water redox reactions. The CB/VB levels were determined to be −3.55/−6.5 eV for 72 and −2.9/−5.77 eV for 73, indicating that both polymers have appropriate electronic band structures for overall water splitting under the given conditions. As a result, both H2 and O2 are evolved for 73 with expected 2:1 stoichiometry upon light irradiation (Fig. 17c). The HER is 218 μmol h−1 g−1 for 73, which is higher than that of 72 (102 μmol h−1 g−1). This could be attributed to the lower band gap of 73 (2.85 eV) compared to 72 (2.94 eV). This work demonstrates the great potential of polymer nanosheets as unique photocatalysts for overall water splitting.
Fig. 17 Chemical structures of polymers 72 and 73; TEM pictures of (a) 72 and (b) 73 as well as (c) time course of H2 and O2 production of polymer 73. Reprinted with permission from ref. 40. Copyright 2017 Wiley-VCH. |
Entry | Reaction | E 0 vs. NHE (V) |
---|---|---|
1 | CO2 + e− → CO2˙− | −1.85 |
2 | CO2 + 2H+ + 2e− → HCOOH | −0.61 |
3 | CO2 + 2H+ + 2e− → CO + H2O | −0.53 |
4 | CO2 + 2H+ + 4e− → HCHO + H2O | −0.48 |
5 | CO2 + 6H+ + 6e− → CH3OH + 6H2O | −0.38 |
6 | CO2 + 8H+ + 8e− → CH4 + 2H2O | −0.24 |
7 | 2H+ + 2e− → H2 | −0.42 |
Most of the photocatalysts for CO2 reduction are inorganic compounds such as metal oxides or sulphides, e.g. TiO2, CdS, BiVO4, etc.163,171 Organic polymers have been relatively less explored so far. For the development of highly active porous CPs for CO2 photoreduction, some critical factors such as band gaps, surface area, CO2 absorption, energy level alignments and photogenerated charge separation of the polymers are needed to be taken into account for pursuing good photocatalytic activities and selectivity. In addition, the external environment and operating conditions, such as catalyst quantity, solution pH, CO2 pressure, temperature, etc. will also likely influence the activity and selectivity of the reactions. For the liquid–solid system, the very low CO2 solubility in the reaction media is one major hurdle to achieve high CO2 reduction efficiency for the photocatalyst. To address this issue to some extent, Liu et al. used ionic liquids to capture more CO2 from air for the reduction reaction.175 It was estimated that ∼33 mg of CO2 was captured by 1 g of ionic liquid. By introducing different functional monomers, the CBs of polymers 74–78 (Fig. 18) are varied from −0.78 and −1.52 eV, indicating sufficient driving force for the reduction of CO2 to CO. The CO production rate follows the order: 78 > 76 > 77 > 75 > 74. The result was largely determined by the CB values of the CPs. Polymer 78 shows the highest CO production rate of 47.37 μmol h−1 g−1 with an excellent selectivity of up to 98.3%, which is attributed to its most negative CB value and the largest photocurrent.
g-C3N4 also has a favourable CB position to drive the reduction of CO2 into hydrocarbons, which makes them potential candidates in this field.30,176 For example, Huang et al. reported that fluorine doped boron carbon nitride (BCN–F) could efficiently catalyse the reduction of CO2 into CO in the CH3CN/H2O/TEOA mixture, which gave 3-times higher CO generation rate (7.75 μmol h−1) than pristine BCN.177 Wang et al. synthesized barbituric acid modified g-C3N4 nanosheets (CN–BAN) through copolymerization between urea and barbituric acid (BA).178 With an optimal amount of BA, CN–BAN could produce CO and H2 with a rate of ∼14 μmol h−1 and ∼3.1 μmol h−1, respectively. In addition, Guo and co-workers prepared g-C3N4 nanosheets with nitrogen vacancies (CN–NV) by polymerization of dicyandiamide in the presence of tartaric acid.179 It was found that the introduction of defects could not only prolong the lifetimes of charge carriers, but also improve the visible light absorption and charge transfer; thereby, CN–NV exhibited a CO formation rate of 56.9 μmol h−1 g−1, which is about 8 times higher than that of pristine g-C3N4.
Apart from g-C3N4, CTFs with high nitrogen content are expected to have good CO2-philic ability and thus facilitate the reduction of CO2. Wang and co-workers developed triazine-based conjugated polymers 79–81,180 which showed tunable band gaps of 2.24–2.93 eV due to the D–A interaction between the triazine unit and comonomers. The BET surface areas were 409, 52, and 37 m2 g−1 for 79, 80, and 81, respectively. After loading CoCl2 and dipyridyl as cocatalysts, polymer 81 gave the highest CO generation rate of 18.2 μmol h−1 with 81.6% selectivity, which is due to the enhanced charge separation by incorporating electron-withdrawing BT units into the polymer.
Considering the advantages of relatively simple and inexpensive gas–solid systems, Liu et al. reported polymers 82–84 by introducing different arylethynylenes.181 The BET surface areas were 445, 131, and 610 m2 g−1 for 82, 83, and 84, respectively. The highest BET surface area of polymer 84 has thus led to a CO2 adsorption capacity of 62 mg g−1 at 1.0 bar, while lower capacities of 39 and 24 mg g−1 were measured for 82 and 83, respectively. In addition, the band gaps are 2.42–2.53 eV for the three polymers. The combination of electrochemical impedance spectroscopy (EIS) and transient photocurrent response measurements demonstrated that 82 had the best charge transfer efficiency. Therefore, polymer 82 exhibited the highest CO production rate of 33 μmol g−1 h−1 with 92% selectivity (Table 3).
Polymer | Optical gap (eV) | HOMO/LUMO | S BET (m2 g−1) | Cocatalyst | Sacrificial agent | Main products and highest yield | Selectivity (%) | Ref. |
---|---|---|---|---|---|---|---|---|
Reaction conditions:a 50 mg of polymer, 10 g of CO2-saturated ionic liquid, 1 g of TEOA, 220 W xenon lamp (>420 nm).b Polymer (15 mg), CoCl2 (1 mmol), dipyridyl (5 mmol), TEOA (1 mL), solvent (5 mL, acetonitrile/water = 4:1), CO2 (1 atm), white light (>420 nm), 30 °C.c 10 mg of polymer powder, 300 W Xe lamp (>420 nm), CO2 and H2O vapour.d 0.9 mg of polymer, 3 mL of CH3CN, and 0.2 mL of TEOA, 225 W Xe lamp (>420 nm), irradiated for 10 h.e 10 mg of polymer, 500 W Xe lamp (800 nm ≥ λ ≥ 420 nm), CO2 (0.4 MPa) and H2O vapour, the reactor temperature was kept at 80 °C.f 50 mg of catalyst, dipyridine (20 mg), CoCl2·6H2O (1 mmol), solvent (5 mL, acetonitrile:H2O = 3:2), TEOA (1 mL), CO2 (1 atm), 300 W xenon lamp (≥420 nm).g 30 mg of catalyst, 1 μmol CoCl2, 15 mg of bipyridine, 5 mL of solvent containing 1 mL of TEOA, CO2 (1 atm), 300 W xenon lamp (>420 nm).h 5 mg of catalyst, 4 mL of CH3CN, 1 mL of TEOA, bipyridine (10 mmol L−1), as well as 25 μL of 20 mmol L−1 CoCl2 aqueous solution, 300 W xenon lamp (>400 nm).i 50 mg of sample, CO2 and H2O vapour, 300 W xenon lamp (>420 nm).j 30 mg of polymer, CO2 and H2O vapour, 300 W xenon lamp (≥420 nm). | ||||||||
74 | 2.48 | 1.70/−0.78 (eV) | 27.6 | — | TEOA | 14.53 (CO)/5.90 (H2) (μmol h−1 g−1) | 71.1 | 175 |
75 | 2.17 | 1.26/−0.91 (eV) | 14.0 | — | TEOA | 27.45 (CO)/1.16 (H2) (μmol h−1 g−1) | 95.9 | 175 |
76 | 2.86 | 1.40/−1.46 (eV) | 442.8 | — | TEOA | 41.44 (CO)/1.20 (H2) (μmol h−1 g−1) | 97.2 | 175 |
77 | 2.76 | 1.61/−1.15 (eV) | 4.5 | — | TEOA | 31.76 (CO)/1.16 (H2) (μmol h−1 g−1) | 96.5 | 175 |
78 | 2.66 | 1.14/−1.52 (eV) | 23.9 | — | TEOA | 47.37 (CO)/0.81 (H2) (μmol h−1 g−1) | 98.3 | 175 |
79 | 2.24 | 1.55/−0.69 (V vs. NHE) | 52 | CoCl2 and dipyridyl | TEOA | 4 (CO)/2 (H2) (μmol h−1) | 66.7 | 180 |
80 | 2.93 | 1.66/−0.77 (V vs. NHE) | 409 | CoCl2 and dipyridyl | TEOA | 10 (CO)/3 (H2) (μmol h−1) | 76.9 | 180 |
81 | 2.43 | 2.08/−0.85 (V vs. NHE) | 37 | CoCl2 and dipyridyl | TEOA | 18.2 (CO)/4.1 (H2) (μmol h−1) | 81.6 | 180 |
82 | 2.45 | 1.50/−0.95 (eV) | 445 | — | — | 33 (CO)/2.8 (H2) (μmol h−1) | 92 | 181 |
83 | 2.53 | 1.57/−0.96 (eV) | 131 | — | — | ∼26 (CO)/∼3.0 (H2) (μmol h−1) | 90 | 181 |
84 | 2.42 | 1.48/−0.94 (eV) | 610 | — | — | ∼12 (CO)/∼29 (H2) (μmol h−1) | 29 | 181 |
85 | — | — | — | — | TEOA | ∼900 (CO)/∼18 (H2) (μmol h−1 g−1) | 98 | 182 |
44 | 2.57 | — | 1412 | — | — | 37 (CH3OH) (μmol h−1 g−1) | ∼100 | 185 |
86 | 2.69 | — | 1053 | — | — | 57 (CH3OH) (μmol h−1 g−1) | ∼100 | 185 |
BCN–F | 2.81 | 1.93/−0.88 (eV) (V vs. NHE) | — | CoCl2 and dipyridyl | TEOA | 7.75 (CO) (μmol h−1) | ∼100 | 177 |
CN–BAN | — | — | 54 | CoCl2 and dipyridyl | TEOA | ∼14 (CO)/∼3.1 (H2) μmol h−1 | 81.9 | 178 |
CN–NV | 2.49 | 1.66/−0.83 (V vs. NHE) | 29.5 | CoCl2 and dipyridyl | TEOA | 56.9 (CO)/10.3 (H2) μmol h−1 | 84.7 | 179 |
CN–O | 2.61 | 1.73/−0.88 (V vs. Ag/AgCl) | 36 | — | — | 0.88 (CH3OH) μmol h−1 g−1 | 100 | 183 |
CN–CV | 2.79 | 1.60/−1.19 (V vs. NHE) | 135.66 | — | — | 4.18 (CO) μmol h−1 g−1 | 100 | 184 |
COFs have inherent features of crystalline porous structures, which facilitate more efficient charge separation. Thus, this kind of material is used for CO2 photoreduction. Huang and co-workers developed COF 85, in which the photoactive COF served as the photosensitizer and Re complex (Re(bpy)(CO)3Cl) was incorporated as a CO2 reduction catalyst.182 Transient absorption spectroscopy showed a lifetime of 171 ps for the excited state of 85, which was much longer than that of the pristine COF, indicating that the incorporation of the Re moiety inhibits charge recombination in 85. COF 85 could produce ∼15 mmol CO/g with 98% selectivity in the TEOA/water mixture after visible light illumination for 22 h.
g-C3N4 could also be modified for CO2 photoreduction in the gas phase. For instance, Yu et al. prepared O-doped g-C3N4 (CN–O) with a tube-like morphology by a two-step process, which involves thermal exfoliation of the bulk g-C3N4 into a nanosheet and curling-condensation of the nanosheet into a nanotube.183 Compared with bulk g-C3N4, the as-prepared hierarchical nanotube exhibited a narrower band gap, higher CO2 uptake and better charge separation. Therefore, CN–O showed a methanol production rate of 0.88 μmol h−1 g−1, which is 5 times larger than that of bulk g-C3N4. Wang and co-workers prepared modified g-C3N4 with carbon vacancies (CN–CV) by heat treatment of pre-synthesized g-C3N4 under an NH3 atmosphere.184 This modification improves light absorption and CO2 adsorption/activation for CN–CV, as compared to pristine g-C3N4. Moreover, the charge generation and separation of g-C3N4 are also enhanced. As a result, CN–CV displayed a CO production rate of 4.18 μmol h−1 g−1, nearly 3.3 times higher than that of pristine g-C3N4.
COFs are also efficient in a gas–solid system; azine-linked COFs 45 and 86 were reported for the photoconversion of CO2 with gaseous H2O into methanol under visible light.185 COF 45 showed a higher BET surface area of 1412 m2 g−1, much higher than 86 (1053 m2 g−1), which is expected to absorb more CO2 for the reduction reaction. With a smaller optical gap of 2.57 eV than 86 (2.67 eV) and better photogenerated charge transfer, COF 45 produced 13.7 μmol g−1 of CH3OH after 24 h. The amount is much higher than that of 86 (8.6 μmol g−1) and C3N4 (4.8 μmol g−1) under the same conditions.
Besides those critical advances in photoredox organic transformations, several new reactions were recently exploited using CPs as the photocatalysts. For instance, polymer 89 was used for photocatalytic C–C double bond cleavage in aqueous media.209 Various styrene derivatives could be cleaved to afford aldehydes with 27–65% conversion and 92–99% selectivity. Yu and co-workers reported three conjugated microporous polymers for C-3 functionalization of indoles;210 morphology and oxidation of polymers played an important role in determining the catalytic efficiency.
The degradation process of organic dyes using semiconductors is complicated and the detailed mechanism is not well revealed in the literature. A combination of electron paramagnetic resonance (ESR) analyses and radical species trapping experiments revealed that singlet oxygen (1O2), superoxide radical (˙O2−), photogenerated hole (h+), and HO˙ are the main reactive species in the process, and the generation of these species could be proposed as follows:
e− + O2 → O2˙− | (1) |
O2˙− + H+ → HOO˙ | (2) |
HOO˙ → H2O2 + O2 | (3) |
H2O2 + O2˙− → HO˙ + O2 + HO− | (4) |
H2O2 + h+ → 2HO˙ | (5) |
H2O2 + hν → 2HO˙ | (6) |
Linear CPs with unique nanostructures show great promise for organic dye degradation. Remita et al. prepared nanofibers of poly(diphenylbutadiyne) (105) by photopolymerization using a soft templating approach (Fig. 21).214 The nanofibers are a few micrometres long with a diameter of about 19 nm. Without additional sacrificial reagents or cocatalysts, 75% MO were decomposed in water after 240 min irradiation, whereas only 17% MO degradation was achieved by Ag-modified TiO2. The photocatalytic performance of the nanofibers was almost unchanged even after 15 repeated cycles. Using the same method, they further prepared polymer 106 with vesicle and spindle morphologies, respectively.215 The spindles are 40 nm thick and the length is several hundred nanometres, while the vesicles are spherical hollow capsules with a diameter of around 1 μm and the thickness of the wall is around 40 nm. A clear shape-dependent photocatalytic activity was demonstrated as shown in Fig. 22a. The nanospindles displayed 100% degradation of MO after 180 min visible light irradiation; by contrast, the vesicles remained totally inactive under identical conditions. Besides, the photocatalytic activity of the polymer 107 nanospindles is much better than that of the polymer 106 nanofibers.
With good photoactivity and excellent water dispersibility, conjugated polyelectrolytes (CPEs) were reported by Zhang and co-workers for organic dye degradation.216 They prepared polymer 108 by attaching 1-alkyl-3-vinylimidazolium bromide onto the side chains of the poly(benzothiadiazolylfluorene) backbone, which was conducted through a self-initiated radical polymerization process under visible light. With a band gap of 2.11 eV, polymer 108 could almost quantitatively decompose MB and RhB after 70 and 90 min, respectively (Fig. 22b). The good singlet oxygen generation ability of polymer 107 under light excitation is likely responsible for the organic dye photodegradation. In addition, RhB was degraded almost quantitatively in repeated 10 runs, demonstrating excellent stability and reusability of 108.
Fig. 22 (a) Comparative photocatalytic degradation of MO of commercial TiO2, Ag-TiO2, polymer nanofibers of 105, vesicles and nanospindles of polymer 106 under visible light (P25-TiO2 is TiO2 with a diameter of 25 nm). (b) Photocatalytic degradation of RhB and MB using polymer 107. (c) The adsorption isotherms of RhB for polymers 108a, 108b and 109 at 25 °C. (d) Removal of RhB with polymers 108a, 108b and 109 in the supernatant. Reprinted with permission from ref. 215–217. Copyright 2015 Nature Publishing Group, 2015 Wiley-VCH and 2015 Royal Society of Chemistry. |
Fig. 23 (a) The adsorption isotherms for RhB on polymers 110–112. (b) Photocatalytic degradation kinetics of RhB using polymer 110. (c) UV/vis absorption spectra of NPs of polymers 114–116 in a water dispersion. (d) Photocatalytic degradation of RhB in the presence of NPs of polymers 114–116 in water. Reprinted with permission from ref. 218–220. Copyright 2018 American Chemical Society, 2019 American Chemical Society and 2015 Royal Society of Chemistry. |
Recently, Ouyang and co-workers synthesized three polymers with similar backbone structures and studied the effect of structural variation on their photocatalytic activities.217 Among them, polymer 109 had a 3D network structure, while the linear polymers 108a and 108b were obtained by varying the concentrations of the feeding monomers.
The BET areas were 1204, 571, 1265 m2 g−1 for 109a, 109b and 109, respectively. As shown in Fig. 22c, with a high surface area, the adsorption capacities of 110 (661 mg g−1) and 108a (688 mg g−1) toward RhB were much higher than that of 108b (189 mg g−1). Polymers 108a and 108b exhibited a very close band gap (1.95 eV for 108a and 1.97 eV for 108b), which was narrower than that of 109 (2.28 eV). By the synergistic adsorption-photocatalysis, RhB was totally decomposed in 30 min for both 109 and 108a, while 120 min was needed for 108b to achieve the same performance (Fig. 22d). These results implied that the BET surface area of the polymer played an important role in photocatalytic activity.
The adsorption-photocatalysis synergy was also reported by Chen et al.218 They prepared 9,9′-bifluorenylidene-based porous polymers 110–112 by introducing different comonomers such as pyrene, tetraphenylethene and triphenyl. Polymer 110 with incorporated pyrene units exhibited the lowest band gap of 1.55 eV and the largest BET surface area of 1306 m2 g−1, while lower surface areas of 777 and 590 m2 g−1 were determined for 111 and 112, respectively. All three polymers demonstrated fast uptake of RhB in an aqueous solution. After 45 min, the absorbed percentages of RhB were 97%, 96% and 89% for 110, 111 and 112, respectively. As shown in Fig. 23a, polymer 110 exhibited a remarkably high adsorption capacity of 1905 mg g−1 of RhB, which is much higher than those of 111 (1024 mg g−1) and 112 (926 mg g−1). As shown in Fig. 23b, more than 81% of RhB was decomposed in the presence of polymer 110 after 30 min visible light irradiation (>450 nm).
Very recently, Jin and co-workers prepared three CMPs with thiazolo[5,4-d] thiazole as the linker for photodegradation of organic dyes.219 Interestingly, the polymers showed nanosheet morphologies and their porous and photophysical properties could be fine-tuned by introducing diphenyl, triphenylbenzene and tetraphenylbenzene. With the highest surface area of 673 m2 g−1 among the three polymers, polymer 113 could absorb 57% of RhB, while the amount of RhB absorbed by the other two polymers is lower than 15%. Polymer 113 can completely degrade RhB in 30 min and 60 min is needed for the degradation of 80% MO.
Conjugated microporous polymer (CMP) nanoparticles offer unique advantages of high porosity and excellent solution dispersibility compared to bulk catalysts, which is desirable for organic dye photodegradation. Inspired by this, Zhang et al. reported the direct preparation of CMP nanoparticles via palladium-catalysed polycondensation reactions in a miniemulsion.220 Interestingly, the difference in the comonomers and linkages has a clear impact on the morphologies and optical properties of the resulting polymers. CMP nanoparticles obtained via Sonogashira-coupling are larger than the ones obtained via Suzuki-coupling. For BT-containing polymers, polymer 116a showed a rod-like shape, while polymer 116b formed nanorings in the reaction media. As shown in Fig. 23c, the introduction of BT into the backbone has led to a narrower band gap of the resulting polymers. Over 80% of RhB was degraded in the presence of 116b after 25 min, while only about 50% of RhB was decomposed for 116a (band gap: 1.96 eV) after the same period (Fig. 23d). This result is in accordance with the lowest optical band gap of 116b (1.76 eV).
The molecular design has undoubtedly played a vital role in achieving the preferred properties of the resulting polymers, such as strong and wide visible light absorption, appropriate band structure, high efficiency of charge separation and transport, as well as good wettability and surface reaction efficiency. These synergy improvements in principle could result in remarkably enhanced catalytic activity. Among the reported strategies, elemental and molecular doping and the introduction of structural defects into the skeleton of g-C3N4 have been well developed for achieving enhanced photocatalytic water splitting and CO2 reduction. In addition, copolymerization of well-designed donors and acceptors has been widely employed to design CPs for photocatalysis, which have brought extended light absorption and improved charge mobility of polymers. The statistical copolymerization for the synthesis of CPPs offered additional opportunity to adjust the band gaps and porosity of polymers.
Despite the great progress made so far, the investigations in this field are in its infancy. The reported CPs suffer from low photocatalytic efficiency, which are far from the practical requirements. Many issues need to be resolved which may include the following:
(1) The development of photocatalytic H2 production using CPs has evolved rapidly in the past few years. A recorded H2 generation rate of 60.6 mmol h−1 g−1 has been achieved, which is very competitive to inorganic counterparts. It is highly expected that more exciting performance could be achieved in the future by designing new polymers, and the combination of computation and experiment appears to be a promising approach for high-throughput screening of polymer libraries. In a very recent report, Cooper and co-workers computationally studied photocatalytic H2 production using 6354 potential alternating co-polymers, among which more than 170 co-polymers have been synthesized and examined for their H2 evolution activities.222 This work well explored the structure–property-performance relationships of candidate co-polymers and successfully discovered new polymers with HERs greater than 6 mmol h−1 g−1. Apart from photocatalytic water reduction or oxidation, photocatalytic overall water splitting has been reported by conjugated microporous polymer nanosheets.40 However, the fundamentals of the photocatalytic process and structure–activity relationship of the polymer are not well understood. Besides, loading metals as cocatalyst could considerably increase the photocatalytic H2 production activity, but knowledge about the interaction mechanism between the polymer and metal is still not clear.
(2) Until now, relatively few CPs were exploited for photocatalytic CO2 reduction and most of them suffer from low activities. Likewise, knowledge about reaction pathways and fundamental enhancements for CO2 reduction is scarce. In order to increase the catalytic performance of the reaction, more work is in urgent demand to optimize the photoactivity and porosity of polymers.
(3) The use of CPs as heterogeneous photocatalysts is rapidly expanding the toolbox of chemical transformations. They have demonstrated good catalytic activities and recyclability for a variety of photochemical reactions. In particular, some reactions were performed directly under natural sunlight. However, the applied reaction types and substrate scope are still limited as compared to those of traditional transition-metal complexes. It is highly desirable to explore new organic molecule transformations, such as more challenging C-heteroatom or C–C functionalization, for heterogeneous polymer photocatalysts.
(4) Photodegradation of organic dyes using CPs presented an effective technology for environmental remediation. Interestingly, nanostructured CPs with good water dispersion have demonstrated excellent degradation ability. Another important finding is that CPPs could decompose organic dyes very efficiently by the adsorption-photocatalysis synergy. In order to achieve better catalytic activity, advanced research should be devoted to optimize the photophysical properties of conjugated polymers to improve the reactive species generation efficiency.
As CPs have the inherent advantages of diverse molecular structures and readily tunable physicochemical properties, there are unlimited opportunities and challenges in developing high-efficiency CPs for various photocatalytic applications. We expect that this review will be a good reference for the next decade of research in this exciting field. With more advances in polymer chemistry and materials science, it is anticipated that critical innovations in photocatalytic technology using CPs could play an increasingly important role in practical applications.
This journal is © The Royal Society of Chemistry 2020 |