Hongjie
Yang
a,
Zhen
Lu
b,
Xiangyu
Yin
a,
Shengjin
Wu
a and
Linxi
Hou
*abc
aDepartment of Materials-Oriented Chemical Engineering, School of Chemical Engineering, Fuzhou University, Fuzhou 350116, P. R. China. E-mail: lxhou@fzu.edu.cn
bQingyuan Innovation Laboratory, Quanzhou 362801, P. R. China
cFujian Key Laboratory of Advanced Manufacturing Technology of Specialty Chemicals, Fuzhou University, Fuzhou 350116, P. R. China
First published on 5th July 2023
Boosting the dissociation of excitons is essential to enhance the photocatalytic efficiency. However, the relationship between the structure of the catalyst and the exciton effect on the photocatalytic activity is still unclear as the main problem. Here, it is proposed that as a descriptive factor, an experimentally measurable dielectric constant (εr) is available to quantitatively describe its relationship with exciton binding energy (Eb) and photocatalytic activity. With tuning the linker of covalent organic frameworks (COFs), the “air gap” structure is oriented to shrink, leading to an increased εr of COFs and a lower Eb to facilitate exciton dissociation. Meanwhile, taking “water-/oxygen-fueled” photo-induced electron transfer reversible addition–fragmentation chain transfer (PET-RAFT) polymerization as a demonstration platform, it can be seen that COFs with a small “air gap” structure have relatively superior photocatalytic activity. This provides important implications for the evolution of efficient photocatalysts.
Covalent organic frameworks (COFs) comprise an attractive category of crystalline organic polymer photocatalysts with pre-programmed structures capable of fulfilling functionally targeted design requirements.11,12 COF-based photocatalysts have made substantial progress in promoting exciton dissociation, such as donor–acceptor strategies,13 heterojunctions,14 and introduction of electron sacrificial agents.15 However, relatively few reports have proposed theories and methods for the inverse on-demand design of COF-based photocatalytic materials based on the correlation between the catalyst structure and photocatalytic activity. Compared with the high dielectric constant (εr) of inorganic semiconductor photocatalysts, the εr of organic polymer photocatalysts is small, and their strong binding may even trigger the decay of excitons.10 In particular, the interlocking iso-network structure and low-polarity periodic skeleton of COFs make the dissociation of excitons difficult.
The εr is directly related to the polarizability of the material and depends heavily on the chemical structure of the material.16 According to the Wannier–Mott model:17Eb = EHm*/εr2 (where Eb is the exciton binding energy, EH is the energy of the 1s orbital of the H atom and m* is the effective mass), the εr has the potential to quantitatively evaluate photocatalytic activity.18 In this regard, Wang's group inserted ionic sites into the skeleton of COFs by polarization engineering, in which highly dipolar ionic groups enhance the directional polarization of COFs, resulting in ionic COFs with high εr relative to neutral COFs and displaying more active exciton behavior.10 It is evident that the εr is a descriptive factor to quantitatively characterize its relationship with Eb along with the photocatalytic activity which is important for the development of efficient photocatalysts.9,10,19–22
Pursuant to the International Roadmap for Semiconductors (ITRS), porous materials and “air gap” structures are an effective way to develop low dielectric materials.23–25 Porous materials are primarily designed to increase the porosity in the intrinsic structure of the material in order to effectively reduce the density of the material and thus decrease the number of polarizable dipoles in the interlayer medium.1,10,19,26–29 The “air gap” structure refers to where air (with a minimum εr of 1 for air) is mixed into a dense material , with the air eventually acting as the dielectric.30 Conversely, to lower the Eb, the design principle of the photocatalyst entails a target along high εr. As is known, oxygen is a radical sudden extinguisher, which greatly impedes the occurrence of controlled radical polymerizations (CRPs) and confines the industrial scalability and wide application of CRPs, and it is extremely significant to tackle this difficulty for advancing the research progress in polymer science.31
In the present study, we take “water-/oxygen-fueled” photo-induced electron transfer reversible addition–fragmentation chain transfer (PET-RAFT) polymerization as a demonstration platform to optimize the photocatalytic activity of COFs by narrowing the “air gap” structure of COFs through ligand tailoring, enhancing the εr and lowering the Eb. The study confirms that the εr can be applied to describe the photocatalytic activity of COFs, correlates the “air gap” structure of COFs with the dielectric properties and exciton effect, and reveals the influence law of the “air gap” magnitude on the photopolymerization reaction activity.
The descriptor εr of Py-TPA-COF and Py-BPDA-COF was measured by using the parallel plate capacitor principle on an impedance analyzer. The εr of Py-TPA-COF decreases with increasing frequency and outperforms the εr of Py-BPDA-COF in the frequency range of 40 Hz to 10 MHz (Fig. 2a). As is known, the relative εr of materials is frequency dependent and depends on the density of the material and the total polarizability of the molecules.25,32 The total polarizability can be attributed to the total contribution of electron polarization, distortion polarization or orientation polarization.19,25,33 Since no ionic sites are inserted in Py-TPA-COF and Py-BPDA-COF, the distortion polarization essentially does not contribute. Considering the strongly ordered skeleton of the crystalline organic polymeric material COFs and its building blocks interlock with each other through the synthesis of net chemistry, the spatial positional degrees of freedom are reduced and thus the orientation polarization is greatly suppressed.19,20,32 As the linker shrinks, it is possible to increase the density of the material, narrow the “air gap” structure, and raise the εr (Fig. 2b). To further verify that the increase in the descriptor εr contributes positively to the decrease in Eb, temperature-dependent photoluminescence (PL) spectra were examined. As seen from the inset of Fig. 2c and d, the PL intensity of Py-TPA-COF and Py-BPDA-COF decreases upon raising the temperature. The temperature-dependent PL spectra were fitted by integration according to the Arrhenius equation (eqn (S1)†) to obtain the Eb.34 The fitting results showed that Py-TPA-COF (Eb = 45.0 ± 2.6 meV) has a lower Eb than Py-BPDA-COF (Eb = 53.8 ± 5.9 meV), implying a relatively better charge–hole separation of Py-TPA-COF after absorbing photon energy. The above experimental results suggest that the directional reduction of the “air gap” structure by ligand modulation can lead to an increase in the descriptor εr, which effectively reduces Eb, which is consistent with the reported correlation pattern.10
Density functional theory (DFT) is applied to interpret the dielectric properties and exciton dissociation behavior of the different “air gap” structures. From the electrostatic potential analysis of Py-TPA-COF and Py-BPDA-COF (Fig. 3a and b), it is seen that the positive potential is uniformly distributed over the PyTTA and linker components, while the negative potential is concentrated in the N atoms of the imine linker mode. The distribution conditions of positive and negative charges are similar and uniform for both, showing no local polarization. The material density of Py-TPA-COF with a short linker is larger than that of Py-BPDA-COF, which means that more dipoles can be polarized by the alternating electric field, coupled with a smaller “air gap” structure, resulting in a larger εr (Fig. 2b). In the difference of the ground state charge density (Fig. 3c and d), the electrons accumulate and dissipate uniformly in the skeleton, indicating again that Py-TPA-COF and Py-BPDA-COF have a uniform charge distribution.
Contraction of the “air gap” structure to raise the descriptor εr and thereby suppress the exciton effect enables efficient charge–hole separation and photocatalytic activity. Firstly, the light harvesting ability of Py-TPA-COF and Py-BPDA-COF was evaluated by UV-vis DRS. As seen in Fig. 3e, the absorption band edges of Py-TPA-COF and Py-BPDA-COF are 592 nm and 602 nm, respectively. The optical band gaps (Eg) of Py-TPA-COF and Py-BPDA-COF were calculated to be 2.23 eV and 2.27 eV, respectively, according to the Tauc plots deduced from the Kubelka–Munk function (Fig. S6a and b†). The above results are consistent with the reported work noting the trend that the εr of the COF is inversely proportional to the square of the optical band gap.9 Subsequently, carrier transport efficiency and redox reaction potential during the photoelectrochemical conversion were studied by photoelectrochemical measurements. Transient photocurrent spectroscopy (Fig. 3f) demonstrates a stronger photocurrent signal for Py-TPA-COF than Py-BPDA-COF, signifying the more efficient charge–hole pair dissociation.35 The flat-band potentials (Efb) were determined by extrapolating the Mott Schottky curves to 1/c2 = 0.36 The Efb of Py-TPA-COF and Py-BPDA-COF was estimated to be −0.82 and −1.12 V vs. NHE, respectively (Fig. S6c and d†). The bottom of the conduction band (CB) of n-type semiconductors is usually considered to be equal to the Efb.37 Hence, the CB potentials of Py-TPA-COF and Py-BPDA-COF are −0.82 and −1.12 V vs. NHE, respectively. With the Eg, the valence band (VB) potentials of Py-TPA-COF and Py-BPDA-COF are +1.41 and +1.15 V vs. NHE, respectively (Fig. 3g). Meanwhile, its redox reaction potential was further verified by cyclic voltammetry with compatible results (Fig. S7†). Negative potentials compared to O2/O2˙− (−0.33 V vs. NHE)15 suggest the feasibility of generating superoxide anions (O2˙−), and insufficiently positive potentials relative to H2O/OH˙ (2.73 V vs. NHE)38 make the direct oxidation of water by holes to form OH˙ an infeasible path. Electrochemical impedance analysis (Fig. 3h) shows that Py-TPA-COF possesses a smaller semicircle radius, implying a relatively small interfacial charge transfer resistance.39 Instantaneous photoluminescence emission spectra (Fig. 3i) at ambient temperature demonstrate the weaker charge–hole recombination efficiency of Py-TPA-COF.38 Meanwhile, the time-resolved photoluminescence decay (TRPL) curves (Fig. 3j) record the shorter fluorescence lifetime of Py-TPA-COF, suggesting a fast exciton dissociation behavior.40
Previous studies revealed that with air/oxygen as the oxidant and assisted by water, the superoxide anion was capable of rapid disproportionation to generate OH˙,41,42 which enabled a deep reaction. Then, the reduction capacity of photogenerated electrons of Py-TPA-COF and Py-BPDA-COF was assessed by means of EPR (Fig. 3k and l). As DMSO can suddenly inactivate OH˙ and O2˙− are observed in DMSO.43 Signal peaks of DMPO-OOH˙ adducts (Fig. 3k) can be noticed in DMSO solutions, which reflect the generation of O2˙−.44 Among the higher DMPO-OOH˙ adduct signal peaks reveal the higher photogenerated electron reduction capacity of Py-TPA-COF leading to a relatively superior ability to generate OH˙. EPR spectra show the detection of DMPO-OH˙ adducts in H2O solution after illumination (Fig. 3l). The signal response intensity of DMPO-OH˙ adducts was stronger for Py-TPA-COF than for Py-BPDA-COF, indicating that more OH˙ was allowed to participate in deep reactions. In the case of H2O, no signal peak of DMPO-O2˙− appeared, attributed to O2˙− generated at the system undergoing a rapid disproportionation in aqueous solution to generate other ROS species (e.g., OH˙).45 The production of OH˙ was again demonstrated with the assistance of the UV-vis absorption spectrum of the indicator methylene blue (MB) (Fig. S8†).
Photo-induced electron transfer reversible addition–fragmentation chain transfer (PET-RAFT) polymerization is mild and controllable, and is a powerful means of preparing target polymers with desired molecular weight and molecular weight distribution. The fundamental solution to the problem of oxygen blocking polymerization is to consider oxygen as the feedstock for catalytic polymerization. Therefore, “water-/oxygen-fueled” PET-RAFT polymerization was examined (Fig. 4a). As shown in the polymerization kinetic curves (Fig. 4b, c, Tables S1 and 3†), Py-TPA-COF exhibits superior catalytic performance under the same experimental conditions with or without TEA addition. The obtained polymers all maintain a low molecular weight distribution (Mw/Mn < 1.5) and have molecular weight (Mn) close to the theoretical line (Fig. 4d). Simultaneously, the Mn, Mw/Mn and GPC elution trace (Fig. 4e and f) of the polymer almost overlap at each intermittent light start-stop operation revealing its high light dependence. Notably, the GPC curve shows a shoulder peak early in the polymerization (Fig. 4e and f), which is caused by the presence of multiple starting radicals at the initiation of polymerization. As seen from the LC-QTOF-MS spectral analysis (Fig. S12†), the peak distributions of multiple sequences are spaced at the exact mass of the HPMA repeat unit (144.08 g mol−1). The polymerization eventually goes into the RAFT cycle with increasing light time. In situ chain expansion experiments (Fig. 4g, h and S13†) demonstrate the high activity of the polymer chains, with the GPC curve shifting in the direction of increasing molecular weight with the addition of the second segment monomer, implying the retention of the thiocarbonyl sulfide group (–SCS–).
To dissect the photocatalytic reaction mechanism of Py-TPA-COF and Py-BPDA-COF, the main variables were modulated. When the O2˙− anion was quenched by a sufficient amount of BQ, the polymerization process did not occur in either case and was alleviated as the amount of BQ decreased (Fig. 5a, b, Tables S5 and 6†). The absence of O2˙− inevitably hinders the generation of OH˙ thereby inhibiting the polymerization reaction. Similarly, when another fuel water was neglected, the pathway of OH˙ generation was hindered (Fig. 5c). Control experiments (Table S2 entries 3–6 and Table S4 entries 3–6†) demonstrated that the polymerization process does not occur in non-aqueous solvents (DMAc/MeCN/EtOH/dioxane), due to the absence of water obstructing the pathway of OH˙ generation. Furthermore, it was found that the polymerization rate increased as the water content of the system increased in the compounding experiments of the good solvent water with the bad solvent dioxane, which further verified the above results (Fig. 5e–g; S8–10; Table S2 entries 1 and 2 and Table S4 entries 1 and 2†). Not surprisingly, the polymerization process did not occur when the OH˙ were scavenged by a sufficient amount of MB, and by diluting the amount of MB, the generation of OH˙ was allowed to initiate polymerization (Fig. 5h and i). The presence of oxygen limits the industrial scalability and wide application of controlled radical polymerization, whereas the strong oxidative properties of OH˙ enable controlled radical polymerization in an open subsurface environment. The feasibility of polymerization processing in an open environment provides an interesting strategy for advancing research in polymer science. Finally, we evaluated the cycling stability of Py-TPA-COF and Py-BPDA-COF (Fig. S14†). The chemical stability of Py-TPA-COF and Py-BPDA-COF was illustrated by XRD as well as FT-IR after 24 h immersion treatment under reaction system conditions. Py-TPA-COF and Py-BPDA-COF maintained high activity after five experimental cycles of photopolymerization, and the slight decrease in activity could be attributed to the decrease in crystallinity.
To further evidence that the εr is a promising descriptor of Eb, an additional pair of TFP-COFs with different air gap structures were investigated by considering the pore size as a single factor. Analogously, TFP-1Ben with smaller air gap structures has higher εr, lower Eb and more efficient high catalytic activity relative to TFP-2Ben (Fig. S15–19†). Experimental results suggest that COFs with smaller pore size have better photocatalytic performance when the air gap structure is regarded as a single influencing factor, revealing that the εr descriptor has the potential to facilitate reverse on-demand photocatalyst design, which will provide a new strategy for the design and development of more efficient photocatalytic systems.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3sc01719b |
This journal is © The Royal Society of Chemistry 2023 |