1. Introduction
In the past hundred years, traditional fossil energy has brought great improvement to human material life, but it has also brought a series of inevitable energy and environmental problems [
1,
2]. The effective use of clean energy and protection of the living environment are necessary guarantees for the sustainable coexistence of all species on the earth [
3]. In the current research, photocatalysis technology using clean solar energy has aroused great attention [
4,
5]. Among various solar photocatalytic technologies, photocatalytic H
2O
2 production is a viable solution for dealing with energy and environmental challenges, because of its minimal energy usage, security, and green credentials [
6]. In addition, H
2O
2 has a significant market potential as an industrial effluent treatment involving sterilization and pulp bleaching [
7]. Notably, the performance of photocatalysts determines the conversion and storage efficiency of photocatalytic technology, which is the key issue affecting the development of photocatalysis technology [
8]. Since Fujishima and Honda discovered in 1972 that titanium dioxide as an electrode can decompose water to produce hydrogen under light [
9], researchers have been working hard to improve the solar energy conversion efficiency of photocatalytic materials, and continue to explore the space configuration, electronic structure, interfacial reaction, and ultrafast charge dynamics theory of photocatalytic materials [
10,
11,
12]. However, the recombination rate of photogenerated carriers of these materials is still very high, resulting a low light quantum efficiency [
3,
13,
14,
15,
16]. Fortunately, organic semiconductors, as a new parallel substitute material for photocatalysis, have attracted great attention and the interest of researchers by the virtue of controllable structure, broad spectral response, designability, and flexibility [
11,
17,
18,
19]. However, a poor charge separation and transportation (CST) process has undermined photocatalytic efficiency in most organic semiconductor-based photocatalysts [
2,
17,
18,
19,
20,
21]. Therefore, there is an urgent need to explore the preparation of new high-efficiency organic semiconductor photocatalytic materials.
Perylene diimides (PDI) and their derivatives are one of the best n-type organic semiconductors [
22,
23]. They demonstrate very unique photoelectrochemical properties and excellent light and thermal stability [
24,
25,
26]. In recent years, with the development of photocatalytic materials, PDI and their derivatives have gradually entered the attention of photocatalysis researchers because of their ultra-wide spectral response and excellent carrier transport properties [
27]. PDI-composite catalytic materials with different central catalysts have shown wide application prospects in the field of environment and energy [
28]. These include PDI-Ni, PDI-TiO
2, PDI-C
3N
4, and other photocatalytic catalysts used for hydrogen production [
29,
30,
31,
32,
33]. Although PDI and their derivatives have been reported in the field of photocatalysis, while PDI exist in the forms of phototrapping/photosensitizer in these catalyst systems, they are not the main catalytic material. As far as we know, there are few studies on PDI as a main photocatalyst, and many problems still need to be solved, such as the effects of conjugation degree and symmetry on the photocatalytic performance of perylene imide, and its regulation mechanism is not very clear.
Herein, we propose a new strategy aiming to construct efficient perylene imide-based photocatalysts by regulation of its monomer symmetry. Accordingly, a novel polymer photocatalyst, named PDI-1,5NDA, was successfully constructed through a simple polymerization method. Furthermore, through a series of experiments, we adjusted the conjugation degree and symmetry of perylene imide by changing the linker and measured the photocatalytic performance data under different conditions. The structural characteristics and photocatalytic performance of the obtained perylene imide are discussed in detail, and the mechanisms of influence of conjugation degree and symmetry on its photocatalytic performance are proposed.
2. Results and Discussion
During the synthesis, 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) was selected as the monomer with perylene groups, while the 1,5-diaminonaphthalene was used as the linker to construct the polymer photocatalyst, and the prepared polymer was named PDI-1,5NDA (
Figure 1a). For purposes of comparison, p-phenylenediamine was used to explore the effect of conjugation, and 1,4-diaminonaphthalene was used to compare the influence of monomer symmetry. The PTCDA was reacted with the diamine monomer through imide condensation reaction under the catalysis of a typical Lewis acid, zinc acetate. In order to verify the chemical structure of the synthesized products, two kinds of samples were measured by Fourier transform infrared spectroscopy (FT-IR). As shown in
Figure 1b, the stretching vibration peak of the -C=O bond appears at around 1640 cm
−1 in all three kinds of PDI samples, while the stretching vibration signals of -O-H/-N-H disappear. Additionally, the absorption peaks at around 1020–1350 cm
−1 verified the generation of a -C-N bond in three PDI. These absorption peaks clearly indicated that the carbonyl bond of PDI is formed by polycondensation. The peaks in the range of 1600 to 1450 cm
−1 come from the carbon atoms in the central nucleus of conjugated PDI of the prepared samples. The FT-IR results show that the synthesized PDI have been imidized completely; there is no unreacted monomer residue, and all of them are consistent with the target products.
Additionally, solid-state Cross-Polarization Magic Angle Spinning Carbon-13 Nuclear Magnetic Resonance (CP/MAS-NMR) was used to determine the chemical environments. In order to exclude the effects of sidebands, cross polarization/total sideband suppression (CP-TOSS) NMR spectra were used to analyze the chemical structure of PDI samples. As shown in
Figure 1c–e, the strong peak at 164 ppm in the low field is typical imine carbons, and various peaks at 110–150 ppm were assigned to carbons on the aromatic motif of PDI and benzene/naphthalene moieties. The above-mentioned information confirmed the formation of the targeted PDI unambiguously.
The chemical structures of PDI were analyzed by XPS technologies. XPS survey spectra show that all PDI are composed of C, N, and O, and no zinc residue was detected (
Figure 2a), which indicated that the PDI are completely metal-free polymers and the Zn salts are completely removed. The solo peak around 399.0 eV in high-resolution scans of the N1s in all the spectra further confirmed the formation of the imine bonds during the Lewis acid-catalyzed poly-condensation reaction (
Figure 2b). High-resolution scans of the O1s peak could be deconvoluted into the main peak at 530.5 eV and a smaller peak at 532.1 eV, which corresponds to C=O bonds in PDI moieties and oxygen-containing species adsorbed on the surface, respectively (
Figure 2c). As shown in
Figure 3d, the key C1s of carbonyl carbon in 1,5-diaminonaphthalene was detected at around 287.3 eV, while that in PDI was at 285.4 eV, and the analyses of another two kinds of PDI were approximately the same as PDI-1,5NDA. The above-mentioned information about valence bond confirmed the accuracy of the chemical structures of PDI unambiguously.
In order to distinguish the crystal structures of three PDI, the powder X-ray diffractometer (PXRD) was used to characterize the samples. From the X-ray diffraction spectra, it can be inferred that all three kinds of materials have a certain degree of crystallinity, and the crystallinity of PDI-1,5NDA is the best among them, followed by those of the other two materials (
Figure 3a–c). Le Bail fitting of the PXRD patterns of PDI-1,5NDA (
Figure 3a) indicated that PDI-1,5NDA had a monoclinic structure with lattice parameters of a = 9.395 Å, b = 7.007 Å, c = 29.134 Å, and β = 124.4° (space group: Pm). The results indicated that the achieved PDI products remain as highly crystalline materials, which can be attributed to the high level of conjugation. Meanwhile, the lower symmetry of the monomer could not destroy the high level of crystallinity of the PDI polymeric materials.
The morphologies of the PDI were subsequently characterized by transmission electronic microscopy (TEM). As presented in
Figure 3d, the highly crystalline PDI-1,5NDA shows the typical nanosheet morphology. The high-resolution transmission electronic microscopy (HRTEM) results for PDI-1,5NDA present clear lattice fringes with a lattice spacing of 0.449 nm, which perfectly matches the peak at 19.73° in the PXRD pattern (
Figure 3a). This result further proves the high level of crystallinity of PDI-1,5NDA. Theoretically, the high level of crystallinity and good π-π stacking would lead to the regular overlap of delocalized π electrons in multiple layers, and the inter-layer electron transfer of PDI-1,5NDA can be greatly promoted through it, by means of which the sample would obtain relatively good catalytic activity. PDI-1,4NDA and PDI-PDA also show multi-layer nanosheet morphology in the images from TEM, but it is difficult to observe their lattice fringes (
Figure 3e,f), indicating that their levels of crystallinity might be lower than that of the PDI-1,5NDA.
In general, the photocatalytic activity of a semiconductor photocatalyst is largely determined by the photon absorption process and the efficient separation and transportation process of the photogenerated charge. The light harvest capacity and band position of the photocatalyst determine the thermodynamically feasible photocatalytic reaction. The light-absorption of the three PDI samples was tested by UV-Vis diffuse reflectance spectroscopy (UV-DRS). Three PDI with different crystallinity showed similar good visible-light capture abilities (
Figure 4a). All of the prepared materials were able to achieve full spectral absorption in the visible region. Notably, both the PDI-1,5NDA sample and the PDI-1,4NDA sample exhibited significant light-absorption in the range of 600–800 nm, which could be ascribed to the large conjugate systems. The band gaps of PDI with different levels of crystallinity are calculated by constructing Tauc Plots (
Figure 4b). Meanwhile, combining the results of XPS valence band spectra and Mott–Schottky measurements, the energy band positions of three kinds of PDI are determined. As shown in the XPS valence band spectra (
Figure 4c), the maximum valence band values (VB max) of PDI-1,5NDA; PDI-1,4NDA; and PDI-PDA are found to be 1.01eV, 0.92 eV, and 0.81 eV, respectively. Thereafter, the energy band structures can be obtained, a process which is shown schematically in
Figure 4d. By using the band gap energy, the conduction band (CB) positions were calculated to be −1.01 eV, −1.12 eV, and −1.24 eV, respectively.
The positive slopes of the Mott–Schottky curves demonstrate that all three PDI are typical n-type semiconductors, and Mott–Schottky tests at three different frequencies pointed to the same flat band potential, which confirmed the accuracy of the experiments (
Figure 5). The flat band potentials of the three kinds of PDI match well with the CB potentials calculated from XPS valence spectra and Tauc Plots. As a reference, the generation potentials of typical reactive oxygen species in photocatalytic water disinfection are −0.33 V (O
2/
•O
2−), 0.28 V (O
2/H
2O
2), 0.82 V (H
2O/O
2), and 1.1–1.99 V (H
2O/
•OH, depending on pH values). Compared with the CB potentials and the VB potentials obtained by the experiments, it can be inferred that the electronic band structures of the PDI are beneficial to the production of reactive oxygen species (ROS) in photocatalytic water experiments, and the kinetic factors of PDI will play an important role in their photocatalytic performance, since the band structures of PDI are quite similar.
The separation and migration of photogenerated carriers are the decisive kinetic factors for the photocatalytic reaction, and the strong internal electric field provides a driving force for the process of charge separation and transportation. The photoluminescence spectra at 570 nm excitation wavelength (
Figure 6a) showed that, compared to the PDI-1,4NDA and PDI-PDA samples with lower crystallinity, the PDI-1,5NDA, with a high level of crystallinity, tends to have lower photoluminescence intensity and a lower fluorescence response, which confirms that the recombination degree of its photogenerated carriers is lower; these experimental results indicated that PDI-1,5NDA has the highest photoquantum efficiency among the three materials.
Electrochemical impedance spectroscopy (EIS) shows that PDI-1, 5NDA appears to have a smaller semicircular Nyquist curve radius than other PDI under the visible light illumination (
Figure 6b), which indicates that the charge transfer resistance of PDI-1, 5NDA is lower. In the photocurrent test, the PDI-1,5NDA, with a high level of crystallinity, also showed higher instantaneous photocurrent (
Figure 6c), indicating that the photogenerated electron-hole separation efficiency of PDI-1,5NDA was the highest, which further proved that the strong internal electric field greatly promoted the separation and transmission efficiency of photogenerated carriers in PDI-1,5NDA.
Since the highly crystalline PDI-1,5NDA has a stronger internal electric field, which promotes the separation and transport of photogenerated carriers, we boldly predict that the visible light-driven photocatalytic performance of PDI-1,5NDA will be stronger than those of the other two PDI samples. We carried out the photocatalytic hydrogen peroxide experiment to verify our prediction.
The experiment used 10% isopropanol solution as a sacrificial agent. Meanwhile, the 300W xenon lamp and the 420 nm cut-off filter were used for visible light irradiation. The photocatalytic activity levels of the three kinds of PDI samples were evaluated with the yield of hydrogen peroxide as the index. The concentration of hydrogen peroxide was determined by iodometry. From the test results (
Figure 7), it can be seen that the photocatalytic activity of PDI-PDA is definitely poor, as only 20.218 μmol·L
−1 H
2O
2 has been produced in one hour. The catalytic activity of perylene imide increased with the addition of a benzene ring to the connecting molecule, and the rate at which hydrogen peroxide was produced by 1 mol PDI-1,5NDA reached 45.394 μmol·L
−1 per hour. These results also confirm that among the three PDI samples, PDI-1,5NDA can produce higher electron concentration and charge mobility. The charge separation ability of PDI-1,5NDA is more efficient than those of the other two PDI. Also, PDI-1,5NDA has the possibility of inhibiting carrier recombination.
In order to further understand the photocatalytic reaction mechanism of PDI-1,5NDA, we explored the time process and rate of photocatalytic hydrogen peroxide production of PDI-1,5NDA under different concentrations of dissolved oxygen by controlling the concentration of dissolved oxygen in 10% isopropanol solution. The results clearly showed that in oxygen-saturated isopropanol solution, the photocatalytic rate of hydrogen peroxide catalyzed by PDI-1,5NDA is the highest, reaching 75.008 μmol·L
−1·h
−1, while the yield of H
2O
2 in argon is extremely low, at only 7.597 μmol·L
−1·h
−1. It can be inferred that the oxygen reduction reaction (ORR) is dominant in the photocatalytic process of hydrogen peroxide production by PDI-1,5NDA, and when ORR is inhibited due to hypoxia, another pathway, which is known as the water oxidation reaction (WOR), encounters difficulty in producing hydrogen peroxide. The asymmetric structural features of PDI-1,5NDA enhance the polarity of the polymerization unit, which facilitates oxygen adsorption. Based on the above analysis, we speculate that the efficient photocatalytic hydrogen peroxide production activity of PDI-1,5NDA is attributed to the asymmetric structural features that enhance the polarity of the polymerization unit, which is conducive to oxygen adsorption. Then, the absorbed O
2 is reduced by the light-generated electron to produce H
2O
2. In addition, the structural unit of the imide bond is very conducive to the generation of hydrogen peroxide, which is also confirmed by previous reports in the literature [
34,
35]. Furthermore, the obtained H
2O
2 could be an excellent broad-spectrum bactericide without risk of increased drug resistance [
36].
3. Experimental Procedures
3.1. Preparation of PDI-1,5NDA Polymer Photocatalyst
The mixture of 780 mg (1 eq, 2.0 mmol) 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA), 306 mg (1 eq, 2.0 mmol) 1,5-diaminonaphthalene, 439 mg (1 eq, 2.0 mmol) zinc acetate dihydrate, and 30 g 1H-imidazole was stirred in an oil bath at 140 °C for 24 h. After the reaction, the mixture was cooled down to 80 °C, 100 mL deionized water was added, and then the solvent was removed by centrifugation. The obtained dark red solid was washed thoroughly with ethanol to remove the residual monomer, then washed with 0.1 mol/L sodium bicarbonate solution to remove excess zinc acetate, and finally, the solvent residue was removed by freeze-drying.
3.2. Preparation of PDI-PDA Polymer Photocatalyst
The preparation of PDI-PDA is the same as the procedures described for the preparation of PDI-1,5NDA. The mixture of 780 mg (1 eq, 2.0 mmol) 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA), 216 mg (1 eq, 2.0 mmol) p-phenylenediamine (PDA), 439 mg (1 eq, 2.0 mmol) zinc acetate dihydrate, and 30 g 1H-imidazole was stirred in an oil bath at 140 °C for 24 h. After the reaction the mixture was cooled down to 80 °C, 100 mL deionized water was added and then the solvent was removed by centrifugation. The obtained dark red solid was washed thoroughly with ethanol to remove the residual monomer, then washed with 0.1 mol/L sodium bicarbonate solution to remove excess zinc acetate, and finally, the solvent residue was removed by freeze-drying.
3.3. Preparation of PDI-1,4NDA Polymer Photocatalyst
Firstly, the 1,4-diaminonaphthalene was prepared as follows: 1 g 4-nitro-1-naphthylamine was dissolved in 50 mL ethanol, and then 10 mg commercial Pd/C was mixed with the solution as catalyst. The NaBH4 powder was added slowly into the solution under stirring. More than 500 mg NaBH4 was added, for a total of 4 additions, until no bubbles came out, resulting in the 4-nitro-1-naphthylamine being completely reduced. The reaction process was tested by Liquid Chromatogram (LC). After the reaction was completed, the catalysts were removed by centrifugation and the solution was concentrated to 5 mL by vacuum distillation. A quantity of 20 mL deionized water was added into the concentrated solution to precipitate the 1,4-diaminonaphthalene out as the product. The achieved 1,4-diaminonaphthalene was further cleaned by deionized water 5 times to remove the salt, and dried in vacuum at 40 °C. Finally, 570 mg 1,4-diaminonaphthalene was prepared.
Secondly, the experimental procedures of the preparation of PDI-1,5NDA were repeated. The 780 mg (1 eq, 2.0 mmol) 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) and 306 mg (1 eq, 2.0 mmo1) 1,4-diaminonaphthalene were mixed with 439 mg (1 eq, 2.0 mmol) zinc acetate dihydrate, and then the mixture was added into 30 g 1H-imidazole and stirred for 24 h in an oil bath at 140 °C. After the reaction, the mixture was cooled down to 80 °C, 100 mL deionized water was added, and then the solvent was removed by centrifugation. The obtained dark red solid was washed thoroughly with ethanol to remove the residual monomer, then washed with 0.1 mol/L sodium bicarbonate solution to remove excess zinc acetate, and finally, the solvent residue was removed by freeze-drying.
3.4. Physicochemical Characterization
FT-IR analysis was carried out on an Agilent (Santa Clara, CA, USA) Cary 630 Fourier Transform Infrared Spectrophotometer, in the spectral range of 4000–400 cm−1. X-ray photoelectron spectra (XPS) were obtained on an ESCALAB250 X-ray photoelectron spectrometer with Al Kα radiation. The binding energy of C1s was used to calibrate the charge effect. The 13C NMR spectra of the solid state were acquired on a Bruker (Billerica, MA, USA) Avance 400 NMR. The powder X-ray diffraction spectra (PXRD) were analyzed by using Cu Kα radiation in the range of 2θ = 5–70 degrees on a Rigaku (Cedar Park, TX, USA) D/MAX-2500PC. The images from a transmission electron microscope (TEM) operating at 200 kV were recorded on an F20 G2 transmission electron microscope (FEI Tecnai). Scanning electron microscope images (SEM) were obtained on an Apreo field emission scanning electron microscope (FEI, Hillsboro, OR, USA). The UV-visible diffuse reflectance spectra (DRS) were recorded by using a Shimadzu (Tokyo, Japan) UV-2600 UV-Vis spectrophotometer and BaSO4 as the reflection standard. The steady-state photoluminescence (PL) spectra of the samples were collected by a Shimadzu RF 6000 fluorescence spectrophotometer. The transient fluorescence (PL) spectra were obtained at room temperature with an excitation laser at 525 nm on a fluorescence spectrophotometer (FLSP-920, Edinburgh instrument, Edinburgh, UK).
3.5. Photocatalytic H2O2 Evolution
The 20 mg photocatalyst was dissolved in 50 mL of 10% isopropyl alcohol (IPA) solution. Then, a 300W xenon lamp (Beijing Perfectlight Technology Co., Ltd., Beijing, China, Microsolar300) equipped with a 420 nm cut-off filter was used to provide visible light irradiation. The effective irradiation area was about 12.6 cm2. Samples comprising 1.5 mL of the reaction solution were taken at regular intervals and centrifuge to get the supernatant. The concentration of H2O2 was analyzed by iodometry. A quantity of 1.0 mL of 0.1 mol∙L−1 potassium hydrogen phthalate (KHC8H4O4) aqueous solution and 1.0 mL of 0.4 mol∙L−1 potassium iodide (KI) aqueous solution was added to the obtained solution, which was then kept for 30 min. The H2O2 molecules reacted with iodide anions (I−) under acidic conditions (H2O2 + 3I− + 2H+ → I3− + 2H2O) to produce triiodide anions (I3−), which can possess a strong absorption at around 350 nm. The concentration of I3− was determined by using a Shimadzu UV-2600 UV-Vis spectrophotometer on the basis of an absorbance of 350 nm, so that the concentration of H2O2 produced during each reaction and at each time period could be estimated by this method. All of the experiments were carried out at room temperature.
3.6. Electrochemical Analysis
Electrochemical and photoelectrochemical measurements were performed on an electrochemical workstation (CHI-660B, Shanghai, China) by using a standard three-electrode cell at room temperature. The working electrode is the PDI-1,5NDA sample loaded on the fluorine-doped tin oxide (FTO) transparent glass, and a platinum wire was used as the counter electrode, while the standard calomel electrode (SCE) was used as the reference electrode. The electrolyte is 0.2 M Na2SO4. The potentials were given with reference to SCE. The transient photocurrent response of the photocatalysts was measured at 0.0 V by turning the light on and off.
The preparation method for the working electrode is as follows: 8 mg catalyst was dispersed in 1.0 mL PVDF standard solution by ultrasonic means. Then, 20 μL of the mixed suspension was dripped onto the surface of the fluorine-doped tin oxide (FTO) transparent glass and dried in an oven at 60 °C.