3.1. Structure and Morphological Analysis
The morphology, elemental composition and distribution of the prepared PCN/BiOCl-0 composite photocatalyst were characterized via SEM (
Figure 2). As shown in
Figure 2a,b, both PCN and BiOCl monomers have a loose and porous layered structure. PCN/BiOCl-0 (
Figure 2c) exhibits a layered structure, with small nano-sized BiOCl sheets loaded on top of large PCN sheets.
Figure 2d is the elemental distribution map of PCN/BiOCl-0, which further proves that BiOCl is uniformly dispersed on PCN. PCN/BiOCl-10 (
Figure 2e) also shows a stacked layer structure, but here the nano-sized BiOCl sheets tend to aggregate. The BiOCl in PCN/BiOCl-20 (
Figure 2f) presents a nano-flower ball structure. The EDS was used to detect and analyze the elemental distribution of the samples, and their molar ratio was calculated according to the mass fraction of Bi and Cl displayed in
Table S1. The molar ratio of Bi and Cl in both the monomeric BiOCl and all PCN/BiOCl composite materials is greater than the molar ratio of chlorine and bismuth elements in standard bismuth oxychloride, indicating the presence of chlorine defects in these photocatalytic materials. Among them, the value of the molar ratio of Bi and Cl in the PCN/BiOCl-0 composite material is the highest (5.71), indicating that PCN/BiOCl-0 has the most chlorine defects. According to the literature, the improved activity of the photocatalyst is due to the presence of defects that enhance the light absorption capacity of the photocatalyst [
13].
To further investigate the microstructure of the PCN/BiOCl-0 photocatalyst, TEM characterization of PCN/BiOCl-0 was performed (
Figure 3). As shown in
Figure 3a, small nanoscale BiOCl sheets (in the red cycle) were observed to be dispersed on large PCN layers. In the HRTEM image at 5 nm (
Figure 3b), the clear lattice fringes demonstrate the good crystallinity of the photocatalytic material. The spacing of the lattice fringes was determined to be 0.75 nm and 0.29 nm, corresponding to the (001) and (110) crystal planes of BiOCl [
14], which further confirms the successful loading of BiOCl onto the PCN surface.
The crystal lattice configuration and phase composition of PCN/BiOCl-X were determined via XRD (
Figure 4a). In the XRD pattern of pristine PCN, there are two diffraction peaks at around 13.8° (100) and 27.6° (002), respectively, which are attributed to the stacking of the structural units within the PCN plane and the interlayer spacing. The results are consistent with previous reports on nitrogen–carbon materials [
15]. In the XRD pattern of pure BiOCl, multiple diffraction peaks were observed, corresponding to the peak positions of the standard card (JCPDS 06-0249), indicating the successful synthesis of BiOCl [
16]. All PCN/BiOCl-X composites showed characteristic peaks of PCN and BiOCl, which confirmed the successful integration of these two materials, consistent with the TEM characterization results. Additionally, the infrared spectra of the prepared samples are shown in
Figure 4b. The appearance of a peak at 814 cm
−1 in the PCN monomer spectrum can be ascribed to the bending vibration of the 3-s-triazine ring [
17]. The presence of peaks between 1190 and 1681 cm
−1, including 1241, 1319, 1415, 1461 and 1636 cm
−1, were characteristic of the stretching modes associated with the C-N=C heterocyclic ring in PCN [
18]. The broad peak centered at 3237 cm
−1 was assigned to the uncoupled amino group and the water molecule that was adsorbed on the surface [
19]. Furthermore, the peak shifted to a lower wavenumber, which may be due to the hydrogen bonds between the hydroxyl and amino groups that may be present in the composite. In the IR spectrum of the BiOCl monomer, the peak at 529 cm
−1 was associated with the Bi–O stretching vibrational mode [
20,
21]. The presence of the peak at 1654 cm
−1 can be explained from the stretching of Cl–O bonds [
22]. The PCN/BiOCl composite materials showed characteristic absorption peaks of PCN; however, the characteristic absorption peaks of BiOCl were not clear. The result was attributed to the low content of BiOCl in the samples and the strong overlap of the BiOCl peak with the PCN peak, making it difficult to distinguish. The specific surface area of the samples was detected using an automatic physical adsorption instrument and recorded in
Table S2. The
SBET value of PCN/BiOCl-0 (65.86 m
2·g
−1) is the highest, indicating that the reaction’s active sites of PCN/BiOCl-0 are more fully exposed and conducive to come in contact with the substrate compared to those of PCN (30.04 m
2·g
−1), PCN/BiOCl-10 (58.97 m
2·g
−1) and PCN/BiOCl-20 (39.67 m
2·g
−1). The increase in the
SBET of PCN/BiOCl-0 may be due to two possible factors: (1) the BiOCl in the PCN/BiOCl-0 has a nanosheet structure, which has a larger
SBET than the flower ball-like BiOCl; (2) the presence of chlorine defects may lead to the formation of some pores [
23].
3.2. Chemical State Analysis
The surface electronic states and elemental compositions of the synthesized samples were analyzed via XPS (
Figure 5).
Figure 5a shows the full spectra of PCN, BiOCl and PCN/BiOCl-0, and all the signature peaks attributed to both PCN and BiOCl are observed in the PCN/BiOCl-0 composite material, indicating the successful synthesis of PCN/BiOCl-0 composite [
15].
Figure 5b shows the high-resolution C 1s spectra of PCN and PCN/BiOCl-0. The PCN and PCN/BiOCl-0 have characteristic peaks at 284.8 eV and 288.1 eV, which are attributed to the standard reference carbon (the relative content of the standard reference carbon was not equal for different samples) and the N-C=N functional group in the aromatic ring with N, respectively [
24,
25]. Compared with monomeric PCN, the N-C=N functional group peak (278.98 eV) in the high-resolution C 1s spectrum of PCN/BiOCl-0 moves towards a lower binding energy. Both PCN and PCN/BiOCl-0 show three distinct peaks in N 1s high-resolution spectra (
Figure 5c), which can be attributed to the sp
2-hybridized C-N=C nitrogen, sp
3-hybridized N-(C)
3 nitrogen and the free amino nitrogen, indicating the presence of these chemical functionalities in both cases [
26]. The N 1s high-resolution spectra of the PCN/BiOCl-0 composite material shows characteristic peaks at a lower binding energy than those of PCN, indicating the shift in the position of these peaks towards the left. As shown in
Figure 5d, the high-resolution XPS spectrum of Bi 4f exhibits peaks at 158.4 and 163.8 eV, which originate from the Bi 4f
7/2 and Bi 4f
5/2 orbitals of BiOCl, respectively. These findings provide strong evidence for the existence of Bi
3+ in the studied sample [
27]. Compared with BiOCl, the high-resolution Bi 4f spectrum of the composite material moves towards a higher binding energy.
Figure 5e shows the high-resolution O 1s spectra of BiOCl and PCN/BiOCl-0. According to the O 1s spectrum of BiOCl, two characteristic peaks situated at 529.2 and 530.8 eV can be ascribed to the Bi–O bond in BiOCl and surface-including water (-OH) species [
28].
Figure S1 shows the O 1s spectra of PCN; the signal was very noisy due to the lower O elemental content, and the contribution of O to PCN/BiOCl can be negligible. Compared to pure BiOCl, the O 1s spectrum of PCN/BiOCl-0 shifts towards a higher binding energy.
Figure 5f shows the high-resolution Cl 2p spectra, where BiOCl has characteristic peaks at 197.1 and 198.7 eV binding energy, corresponding to Cl 2p
3/2 and Cl 2p
1/2 of BiOCl, respectively, proving the existence of Cl in the material [
29]. Compared with monomeric BiOCl, the peak strength of the Cl 2p signal of PCN/BiOCl-0 shifts towards a higher binding energy. In summary, after the formation of the complex material, a decrease in binding energy is observed in the C and N 1s spectra, while the Bi 4f, O 1s and Cl 2p spectra shift towards higher binding energies. The binding energy of elements tends to decrease as their electron density increases, and the electron density on the surface of BiOCl is lower than that on the surface of PCN. This result implies the occurrence of an electron transfer process from BiOCl to PCN, demonstrating the successful formation of a heterojunction structure between PCN and BiOCl [
19]. Additionally, the Cl 2p spectrum of PCN/BiOCl-0 has poor signal-to-noise ratio and weak signal peak intensity, which is attributed to the low absolute content of chlorine, further proving the existence of chlorine defects in the composite.
3.3. Photoelectrochemical and Optical Properties
The ability of the samples to absorb light was measured through UV-Vis, and the corresponding results are illustrated in
Figure 6a. Compared to PCN/BiOCl-10 and PCN/BiOCl-20, PCN/BiOCl-0 has a higher light absorption intensity, which may be attributed to the chlorine defects in PCN/BiOCl-0 that cause the emergence of defect states between the conduction and valence bands of BiOCl. Previous studies indicated that the charge excitation from the defect states into the conduction band can improve the light absorption capability of the material [
13]. The high light absorption capacity of PCN/BiOCl-0 is beneficial for promoting photocatalytic reactions. The absorption edge of PCN can be observed at 443 nm in
Figure S2a. After loading with BiOCl, the absorption edge of PCN in the composite material undergoes blue shift (the absorption edge of BiOCl is located at 341 nm) (
Figure S2b). The combination of PCN and BiOCl causes the absorption edge of the composite material to shift towards the midpoint of the two monomers, consistent with many previous studies [
30]. The bandgap values of the photocatalysts were evaluated through the application of the Kubelka–Munk equation, and the bandgap values of PCN and BiOCl were 2.80 eV and 3.60 eV, respectively. Compared to PCN monomers, the bandgap values of PCN in all composite materials (Eg ≈ 3 eV) increased (
Figure 6b), which may be due to the interaction between PCN and BiOCl, further confirming the successful combination of PCN and BiOCl.
By measuring the fluorescence intensity, the rate of electron-hole recombination in the sample can be detected (
Figure 7). Under illumination, part of the energy generated by the photo-induced electron-hole recombination in the photocatalyst is transferred to fluorescence [
31]. The quenching of fluorescence indicates that the rate of electron-hole recombination is effectively suppressed. Thus, the weaker the PL emission peak of the photocatalyst, the higher rate of electron-hole recombination, indicating the photocatalytic performance was stronger.
Figure 7 shows that the fluorescence emission peaks of the samples from high to low are PCN > PCN/BiOCl-10 > PCN/BiOCl-20 > PCN/BiOCl-0. Compared to the PCN monomer, the PL emission peak of the composite materials is weaker due to the formation of a heterojunction between PCN and BiOCl, which increases the efficiency of charge separation. The PL emission peak of PCN/BiOCl-0 is the weakest, indicating the best separation efficiency of its photogenerated charge carriers.
The photocurrent response of the samples was studied using an electrochemical workstation. A stronger photocurrent response of a semiconductor photocatalyst indicates better charge separation efficiency [
32].
Figure S3a shows the photocurrent of the samples under five intermittent light irradiations. The composite materials have a stronger photocurrent than the PCN and BiOCl monomers. Among them, the PCN/BiOCl-0 sample shows the strongest photocurrent response, implies that it has the highest mobility of photogenerated charge carriers and effectively suppresses the recombination of e–h
+ pairs. This result is in line with the result of photoluminescence analysis.
The separation and recombination behavior of the photogenerated charge carriers was further investigated through EIS to analyze the conductivity of the samples under light and dark conditions.
Figure S3b shows the impedance of the samples under illumination. The Nyquist plot of the samples’ arc radii from high to low is PCN > BiOCl > PCN/BiOCl-10 > PCN/BiOCl-20 > PCN/BiOCl-0. PCN/BiOCl-0 displays the smallest Nyquist plot arc radius, indicating the highest charge carrier mobility.
Figure S3c shows the dark impedance of the samples, and the order of the arc radius size is the same as that of the light impedance. These results demonstrate that the charge carrier transport efficiency of PCN/BiOCl-0 composite material is superior, in agreement with the photocurrent testing results.
3.4. Photocatalytic Activity
The photocatalytic performance of the catalysts was evaluated through the photocatalytic coupling reaction of benzylamine, and the experimental results are listed in
Table 1. All the catalysts showed excellent selectivity (~99.9%) for the photocatalytic coupling reaction of benzylamine. Compared with the monomer PCN (40.1%) and BiOCl (54.5%), the PCN/BiOCl-X composite materials showed significantly enhanced photocatalytic performance in the photocatalytic coupling reaction of benzylamine, which can be attributed to the formation of the heterojunction between PCN and BiOCl. The photocatalytic performance of the three composite materials was as follows: PCN/BiOCl-10 (72.7%) < PCN/BiOCl-20 (76.1%) < PCN/BiOCl-0 (88.0%) (
Table 1, entries 3–5). Among them, PCN/BiOCl-0 had the highest photocatalytic activity, which may be attributed to two main factors: (1) a higher specific surface area allows the active sites on the catalyst to come in contact with substrate more completely; (2) the presence of chlorine defects makes the material have high light absorption intensity, which is beneficial for enhancing the photocatalytic performance.
In order to explore the perfect conditions for the photocatalytic oxidation coupling reaction of phenylethylamine, optimization was carried out from two aspects, reaction time and catalyst dosage, and the experimental results are shown in
Table 2. It was found that the PCN/BiOCl-0 photocatalyst had the best catalytic effect based on
Table 1. Therefore, all subsequent reactions were catalyzed using PCN/BiOCl-0. Firstly, the optimization of the catalyst dosage was carried out (
Table 2, entries 1–4). When the amount of the photocatalyst was 0 g, the conversion rate of benzylamine was only 2.7%, indicating that the photocatalyst was crucial in the reaction. Generally, when the catalyst dosage was increased, the conversion rate of the photocatalytic reaction also increased correspondingly. However, it can be observed from
Table 2 that the amounts of the photocatalysts were 0.02 g and 0.03 g and the yields of the product were 68.3% and 66.2%, respectively. This result indicates that increasing the catalyst dosage does not significantly increase the yield of products. From the perspective of reaction results and economics, the optimal catalyst dosage was 0.02 g.
When the catalyst dosage was 0.02 g, the optimization of the irradiation time was carried out (
Table 2, entries 5–8). With the increase in irradiation time, the conversion rate of phenylethylamine also increased. When the irradiation time was 6 h, the conversion rate of phenylethylamine reached 97.2%. Under the condition without light (
Table 2, entry 5), the conversion rate of the phenylethylamine was only 2.2%, which proves that light is an important condition for catalyzing the oxidation coupling reaction of benzylamine.
The scope of the reactants was expanded under the optimal reaction conditions, and the corresponding experimental results are documented in
Table 3. The photocatalytic oxidation of amine derivatives gives rise to the formation of the corresponding imines, with impressive selectivity (~99%). Compared to electron-withdrawing groups, electron-donating groups were more reactive (
Table 3, entries 2–4). By using the benzylamine derivatives substituted with a methyl group to study the steric hindrance effect, the conversion rates of the benzylamine derivatives were as follows: 4-methylbenzylamine (100%) > 3-methylbenzylamine (83.9%) > 2-methylbenzylamine (76.6%) (
Table 3, entries 4–6). The results of the experiment indicated that the greater the steric hindrance of the reaction substrate, the more restricted the reaction proceeded. Furthermore, for the reaction of heterocyclic amines (
Table 3, entries 7–9), good conversion rates (around 80–90%) were obtained within a given reaction time.
In order to broaden the application of the photocatalysts, a photocatalytic oxidation–dehydrogenation reaction was carried out on secondary amines. Using the oxidation–dehydrogenation of dibenzylamine as a template reaction, the photocatalytic activities of different catalysts were determined. The photocatalytic activity of the catalysts can be evaluated according to the conversion rate of dibenzylamine in
Table 4, which suggests the following order: PCN (25.7%) < BiOCl (36.0%) < PCN/BiOCl-10 (39.9%) < PCN/BiOCl-20 (45.2%) < PCN/Bi
2O
3 (52.0%) < PCN/BiOCl-0 (59.1%). The PCN/BiOCl-0 exhibited the best photocatalytic activity, which may be attributed to the successfully constructed heterojunction structure, abundant chlorine defects and high specific surface area. This rule was consistent with the photocatalytic coupling reaction of benzylamine oxidation.
To further optimize the reaction conditions, the reaction was carried out with different amounts of catalyst and irradiation times. To be consistent with the coupling reaction of benzylamine, the amount of reactant was increased to 0.25 mmol. The different amounts of catalyst are compared in
Table 5 (entries 1–4); the appropriate amount of catalyst was 0.02 g based on the conversion rate and selectivity of the photocatalytic oxidation–dehydrogenation of dibenzylamine. By comparing the reaction results at different irradiation times (
Table 5, entries 5–9), when the irradiation time was increased from 3 h to 7 h, the conversion rate of dibenzylamine gradually increased, but the selectivity decreased. Side reactions may occur after a certain reaction time and lead to the formation of by-products. Taking all factors into account, the optimized irradiation time was 6 h.
A range of substrates were explored under the optical reaction conditions, and the experimental outcomes are listed in
Table 6. For symmetric and asymmetric secondary amines (
Table 6, entries 1–4), high conversion rates and selectivity were achieved and a small amount of benzaldehyde was detected. For 1,2,3,4-tetrahydroquinoline, high selectivity (99%) for quinoline was obtained by extending the reaction time (
Table 6, entry 5). For 1,2,3,4-tetrahydroisoquinoline, one molecule of hydrogen was removed in the reaction, and the conversion rate reached 91.2% within 2 h (
Table 6, entry 6). However, the further oxidation of 3,4-dihydroisoquinoline to isoquinoline during the reaction reduced the selectivity.
In order to study the stability of the PCN/BiOCl-0 photocatalyst, cyclic experiments were carried out.
Figure 8 shows the cyclic reaction tests carried out using benzylamine (
Figure 8a) and dibenzylamine (
Figure 8b) as substrates. The selectivity and conversion rate remained stable after five cycles, and a slight decrease was observed. The used catalyst in
Figure 8b was collected, and the stability of the catalysts was investigated using XRD, UV-Vis DRS, FT-IR and XPS.
Figure S4a–c are the XRD, UV-Vis DRS and FT-IR spectra;
Figure S5 contains the XPS spectra of before and after PCN/BiOCl-0, respectively. The positions of the main characteristic peaks had not changed after the reaction, indicating that the structures of catalysts were stable after the reaction. The catalytic activity of the photocatalyst was decreased, possibly due to the mechanical damage during the cyclic process. Overall, the PCN/BiOCl-0 photocatalyst shows excellent chemical stability, making it a viable candidate for recycling.
3.5. Mechanism Analysis
To further investigate the main active species in the reaction, isopropanol (IPA), silver nitrate (AgNO
3), p-benzoquinone (BQ) and potassium iodide (KI) were added to the PCN/BiOCl-0 photocatalytic amine oxidation reaction as ·OH, e
−, ·O
2− and h
+ scavengers, respectively (
Figure 9). In the light of these experimental results, it may be proposed that ·OH has a minimal impact on the reaction, based on the slight reduction in conversion rate observed upon IPA addition. The addition of BQ and AgNO
3 leads to a decrease in the conversion rates of benzylamine and dibenzylamine, indicating a certain involvement of ·O
2− and e
− in the reaction process. The conversion rates of benzylamine and dibenzylamine were decreased significantly after the addition of KI, demonstrating the primary involvement of h
+ in the reaction process. In addition, the yield of product decreased significantly when the reaction was carried out under N
2 atmosphere, further indicating the minor role of O
2− in the catalytic process.
According to characterization data obtained from UV-Vis diffuse reflectance spectroscopy and the Kubelka–Munk equation, the band gap values (E
g) of BiOCl and PCN were found to be 3.60 eV and 2.80 eV, respectively. The potential energy values of the valence band (VB) and conduction band (CB) of the photocatalysts were determined using Equations (1) and (2):
Herein, χ is used to denote the absolute electronegativity value of the sample, while E
e refers to the energy level of the free electrons relative to hydrogen (~4.5 eV) [
33]. Based on the characterization results, a possible Z-scheme electron transfer mechanism is formulated and displayed in
Figure 10. Under the light illumination, PCN and BiOCl produce photogenerated h
+ and e
−, respectively. The photogenerated h
+ on the VB of PCN and e
− on the CB of BiOCl were recombined, while the photogenerated e
− on the CB of PCN reacted with O
2 to form ·O
2−. The generated h
+ and ·O
2− play a crucial role in the amine oxidation reaction. In the reaction mechanism of benzylamine oxidation, benzylamine forms a free radical cation by reacting with photoinduced h
+ captured at the VB of PCN; the benzylamine then reacts with ·O
2− to produce imine and hydrogen peroxide in the further reaction. The benzyl imine produced in the process undergoes a nucleophilic addition reaction with benzylamine to form N-phenylethylene-1-phenylmethylamine [
33]. For the reaction mechanism of diphenylamine, there is a possible reaction pathway: diphenylamine undergoes direct oxidation and dehydrogenation to produce N-benzylidene-1-phenylmethanamine.