Boron nitride quantum dots modified carbon-defects ultra-thin porous carbon nitride: double channels and quantum size effects facilitate photogenerated carrier migration and exciton dissociation

Graphite carbon nitride possesses great promise for visible photocatalysis, but the bulk carbon nitride prepared from nitrogen-rich precursors such as melamine has inherent drawbacks such as retarded photogenerated carrier migration and exciton effects, which limit its application. Herein, we constructed a novel Boron nitride quantum dots modified carbon-defects ultra-thin porous carbon nitride (BNQDs/Vc-UPCN). The double channels were constructed by carbon-defects structure and Boron nitride quantum effect to overcome its inherent drawbacks and applied to the photodegradation of common persistent organic pollutants (methylene blue). The structure, porosity, elemental composition, optical properties, photoelectrochemical properties, and photocatalytic properties of the prepared BNQDs/Vc-UPCN composites were investigated using various characterization methods. Meanwhile, the results of radical trapping experiments and electron spin resonance characterization demonstrated that BNQDs/Vc-UPCN promote molecular oxygen activation more than Vc-UPCN did. In terms of degradation effect, the best sample (BC-1) is 10 times more effective than the initial sample (BCN). This study proposes an effective mechanism for constructing novel visible-light-driven photocatalysts using carbon-defects ultra-thin structures and quantum dots, which can be used for the treatment of organic pollutants.


Introduction
Semiconductor photocatalysis technology, as a green and low-cost emerging technology for organic matter treatment, has gained considerable attention due to its potential as a feasible and cost-effective process for producing reactive oxygen species (ROS) [1][2][3][4]. ROS is a type of green oxidant, which has great potential for environmental governance, particularly in the degradation of persistent pollutants, due to its high efficiency in producing ·O 2 − and ·OH [5][6][7][8][9]. However, the technology for producing active oxygen by photocatalysis is usually limited by the visible light absorption capacity and the low separation efficiency of photogenerated carriers. Generally, the visible light absorption capacity and the low separation efficiency of photogenerated carriers are the main factors that restrict the production of active oxygen by photocatalysis [10][11][12][13]. Therefore, it is essential to identify an efficient, cost-effective photocatalyst with an appropriate bandgap. Graphitic carbon nitride (g-C 3 N 4 ) is considered to be an ideal photocatalyst for molecular oxygen activation due to its environmentally friendly, low cost, 2.7eV narrow bandgap, and tunable morphology [12][13][14][15][16][17][18][19][20]. However, the photoexcitation process of bulk carbon nitride prepared by nitrogen-containing precursors such as melamine generates a large number of Frenkel excitons [17,21]. The strong Coulomb force between Frenkel excitons in bulk g-C 3 N 4 leads to high exciton binding energy, resulting in slow exciton dissociation and severe charge recombination, which is unfortunately often overlooked [22]. Besides, the stacked sheet structure of bulk carbon nitride results in low specific surface area and insufficient reaction sites, which significantly impairs the photocatalytic performance of carbon nitride [12,16,23].
The construction of special morphologies has been proven as an effective strategy to overcome the original drawbacks of bulk carbon nitride, especially the ultra-thin porous two-dimensional structures [24][25][26]. Ultrathin and porous g-C 3 N 4 features have been observed to have a shorter interlayer distance and more reactive sites, which can enhance the exciton dissociation and free charge transfer processes, as well as provide additional sites for chemical reactions [27,28]. Currently, the construction of ultra-thin structures mainly involves both topdown and bottom-up approaches [17]. Gao et al [29] constructed porous ultra-thin carbon nitride with a carbon-defects structure (Vc-UPCN) by the one-pot method, which significantly enhanced the photocatalytic performance of carbon nitride. In particular, the introduction of carbon-defects causes the conduction band of Vc-UPCN to shift to a higher energy level, which is beneficial for activating reactive oxygen species. Moreover, the introduction of a carbon-defects structure not only reduces the Eg value, facilitating the activation of ROS, but can also act as an electron trap that captures photogenerated electrons, promoting photogenerated carrier separation [30,31].
However, photogenerated holes have a shorter lifetime and are more prone to compounding [32]. The carbon nitride with only a carbon-defects degradation effect is still not ideal. Therefore, it is necessary to further promote the reverse migration of the photogenerated electrons and holes. A built-in electric field is a feasible method. The presence of a built-in electric field provides an impetus for promoting the reverse migration of photogenerated electrons and holes [1,21,33]. The construction method mainly involves polarization field engineering [34] surface heterojunction formation [35] and doping with heteroatoms [36].
Boron Nitride Quantum Dots (BNQDs) are spherical particles with sizes ranging from 1 to 10 nm, featuring an abundance of negatively charged groups. These abundant negative charges are highly attractive to photogenerated holes, facilitating the separation of photogenerated electron-hole pairs [28,32,37,38]. Therefore, BNQDs can be utilized as a hole extractor to enhance exciton dissociation efficiency and facilitate charge separation.
In this study, 0D/2D-BNQDs/Vc-UPCN heterojunction photocatalyst materials were constructed, utilizing BNQDs to construct the built-in electric field, which cooperated with carbon-defects to boost the bidirectional separation of photogenerated electron-hole pairs, thereby enhancing exciton dissociation and improving the ability of molecular oxygen activation under visible light.

Experimental part
2.1. Synthesis of photocatalysts 2.1.1. Construction of carbon-defects ultra-thin porous carbon nitride (Vc-UPCN) According to Gao et al [29], Vc-UPCN was prepared via in situ thermal etching by solubilizing 20 g of urea in 20 ml of deionized water at 50°C. The mixed solution was then transferred into a crucible and calcined in an air atmosphere at 550°C for 4 h. The milk-white solid obtained by cooling was Vc-UPCN. Additionally, 20 g of urea was directly calcined to obtain bulk carbon nitride (BCN).

Synthesis of BNQDs
BNQDs are fabricated as follows [39]. First, 0.10 g of HBO 3 was solubilized in 10 ml of distilled water under magnetic stirring. Then, 0.034 g of melamine was added and agitated continuously for 20 min until a precipitate was formed. After, the suspension was shifted to a 25 ml reaction kettle and was maintained for 15 h at 200°C. After waiting for the reaction kettle to drop to room temperature, the large particles were filtered through a 0.22 μm filter membrane, and the solution obtained was BNQDs.

Synthesis 0D/2D BNQDs/Vc-UPCN heterojunction photocatalyst
BNQDs/Vc-UPCN (BC) composites were synthesized, as figure 1. First, 0.2 g of Vc-UPCN was added to 30 ml ethanol and ultrasonicated for 60 min to obtain a homogeneous mixture. Afterward, a certain volume of BNQDs solution was added dropwise to the above suspension under rapid stirring and continued for 24 h until the ethanol was completely evaporated. Finally, the above solid was transferred to a vacuum oven and kept at 50°C for 12h. Vc-UPCN with different BNQDs amounts is named BC-X, where 'X' denotes the added volume of the BNQDs solution (X = 0.5, 1, 1.5 ml).

Characterizations
Characterization of the sample crystal structure by x-ray diffraction (XRD, Empyrean, Cu target). Atomic force microscopy(AFM, Bruker Dimension Icon) was used to observe the thickness of nanosheets. Transmission electron microscopy (TEM) was performed using an FEI Tecnai F20 instrument at 200 Kv with mapping. Scanning electron microscopy (SEM) was performed using a Zeiss Sigma 300. The chemical compositions of the different samples were determined mainly by x-ray Photoelectron Spectroscopy (XPS, THermo Scientific) and Fourier Transform Infrared Measurements (FTIR, Thermo Scientific Nicolet iS20). The visible absorbance of the sample is mainly provided by UV-vis DRS (UV3600, 200-800 nm) Steady-state photoluminescence spectroscopy and time-resolved photoluminescence decay spectra of samples BCN, Vc-UPCN, and BC-1 were collected using an F-4600 FL Spectrophotometer and FLS 920 timeresolved Spectrofluoromter, respectively. Photoelectrochemical measurements were conducted on a Chenhua CHI 660E electrochemical workstation using a standard three-electrode model. The transient photocurrent responses of the prepared samples were monitored on an IVIUM Technologies electrochemical workstation using a standard three-electrode cell. A Pt foil and Ag/AgCl electrode (in saturated KCl) served as the counter and reference electrodes, respectively, and indium tin oxide (ITO) deposited with the catalyst was used as the working electrode. An aqueous Na 2 SO 4 (0.5 M) solution was used as the supporting electrolyte, and a 300 W Xe lamp was chosen as the light source. The working electrode was prepared as follows: approximately 15 mg of the as-prepared catalyst was suspended in 20 μl of a 5 wt% Nafion solution and 0.5 ml of ethanol. The mixture was then ground to make a slurry, which was then evenly spread as a thin film onto an ITO glass substrate with an active area of 0.8 cm 2 . The resultant coated ITO substrate was then dried at 80°C.

Photocatalyst degradation activity measurement
All photodegradation experiments were performed using a 500 W xenon lamp (with a 420 nm filter). In a typical experiment, 0.05 g of the synthesized catalyst was incorporated into 50 ml of a methylene blue solution (20 mg L −1 , M −1 B −1 ). Before irradiation, the composite was agitated in the dark for 0.5 h until adsorption-desorption was in equilibrium was reached at the ambient temperature. The samples were removed every 20 min during the illumination and oxidant processes and the concentration of MB in the solution was measured at 664 nm to determine the catalytic degradation efficiency. The errors in all graphs were calculated using standard deviations.
To investigate the effect of the major reactive radicals on methylene blue, N 2 , EDTA-2Na, and isopropanol (IPA) were used to capture ·O 2 − , h + , and ·OH, respectively.

Results and discussion
3.1. Characterization of photocatalysts X-ray diffraction (XRD) and Fourier-transform infrared (FTIR) spectroscopy were used to reveal the composition and chemical characteristics of the samples. In figure 2(a), both BCN and Vc-UPCN display two x-ray diffraction peaks, (100) and (002), which according to the tri-s-triazine unit and interlayer spacing of the carbon nitride, respectively. The (002) peak of Vc-UPCN is slightly shifted to the high-angle region compared to that of BCN, and the diffraction peak (002) becomes wider and weaker, indicating that the bulk C 3 N 4 framework is optionally mutilated and formed into short-range ordered graphite molecular debris. Therefore, the layer spacing becomes shorter. This effectively accelerates photogenerated carrier transfer and exciton dissociation, which also be observed using SEM and TEM. For BC-0.5, BC-1, and BC-1.5, the XRD peaks did not change significantly with Vc-UPCN, and there were no obvious BNQDs peaks, indicating that BNQDs were less abundant and evenly distributed. Moreover, FTIR spectra are in figure 2(b). Clearly, all samples have a comparable framework. Notably, the peaks located at 810 cm −1 , the group of peaks between 1200 and 1700 cm −1 , and 3000 to 3400 cm −1 are distinctive signatures of the triazine units, typical stretching vibration modality of C-N heterocyclic rings, -NH, and stretching vibrations, respectively. Compared to BCN, Vc-UPCN showed a larger peak area region corresponding to the NH x group within the range of 3100-3490 cm −1 . This could be due to the increased exposure of Vc-UPCN to N atoms, which improved its dispersibility in water [40]. For this purpose, EPR was used to observe the presence of carbon-defects structures, as shown in figures 2(c) and (d). BCN and Vc-UPCN showed a dominant EPR signal (g = 2.004), which was attributed to unpaired electrons on the carbon atom of the heptazine ring [41][42][43]. Notably, the EPR signal intensity of Vc-UPCN is significantly weaker than that of BCN, suggesting the construction of a carbon-defects structure in Vc-UPCN because carbon atom loss reduces the density of lone pairs of electrons. SEM, AFM and TEM were used to characterize the sample morphology. As illustrated in figure 3(a), the Vc-UPCN had a distinct lamellar structure. Further, an ultra-thin lamellar morphology was observed by TEM ( figure 3(b)). It is worth noting that the AFM shows that the thickness of Vc-UPCN nanosheets is 1.4 nm(figures 3(c), (d)). According to the TEM image ( figure 3(e)), BNQDs show a dispersed particle morphology with an average particle size of 3.5 nm. For BNQDs/Vc-UPCN ( figure 3(f)), the BNQDs were uniformly dispersed on the surfaces of the Vc-UPCN nanosheets. Meanwhile, the high resolution in the inset shows that BNQDs have a unique lattice spacing of approximately 0.21 nm, in correspondence with the (100) hexagonal pattern of Boron nitride. In addition, the elemental mapping of BC-1 ( figure 3(g)) also demonstrates the presence of B, C, and N, which indicates that a compositionally homogeneous complex was obtained. This information explicitly illustrates the intimate binding of the BNQDs of the BC-1 samples with Vc-UPCN.
The presence of some large pores is observed in figure 3(a), and for this reason, the pore structures of Vc-UPCN and BC-1 were analyzed by N 2 adsorption/desorption experiments ( figure 4). The pore volume and specific surface area of Vc-UPCN were computed to be ≈0.4 cm 3 g −1 and 142.21 m 2 g −1 , while that of BC-1 decreased to ≈0.3 cm 3 g −1 and 60.80 m 2 g −1 ), which could be attributed to the Vc-UPCN surface space occupied by BNQDs and pore blockage. However, the surface area of BC-1 is still much more than that of BCN (8.8 m 2 g −1 ) [32].
The surface chemical composition of the samples was investigated using the XPS. As depicted in figure 5(a), C, N, and O were all identified in Vc-UPCN and BC-1, with the difference that B was only detected in BC-1, which implies that BNQDs were successfully dispersed on the Vc-UPCN surface. The signal of the elemental O   tertiary carbon(N-C=N) and C-NH 2 groups, which is consistent with the tri-s-triazine structure of carbon nitride. Notably, the appearance of C-NH x at a binding energy of 287.1 eV implies partial rupture of the tri-striazine structure and the formation of a carbon-defects structure. In addition, the N 1s spectrum of Vc-UPCN in figure 5(c) can be divided into four peaks at 398.7, 400.2, 401.2, and 404.3 eV, respectively, which are assigned to the C=N-C structure, N-(C) 3 , C-NH groups, and π excitation effect in the heptazine rings. In contrast to Vc-UPCN, the new peaks at,398.1 and 399.7 eV of BC-1 represent N-B and C-N-B. Besides, in the B1s spectrum of BC-1 (figure 5(d)), the peak located at 191 eV corresponded to the B-N bond. Based on the above analysis, it can be concluded that BC-1 has been successfully prepared. Figure 6(a) shows the UV-vis DRS of BCN, Vc-UPCN, and BC-1, all of which show good visible-light capability. Compared to BCN, Vc-UPCN shows a significant blue shift and a decrease in visible light absorption, which is mainly ascribed to the quantum effect of the ultra-thin structure of Vc-UPCN. Moreover, the visible light absorption of Vc-UPCN after modification with BNQDs was enhanced, which indicates that BNQDs can improve the visible light absorption of Vc-UPCN [15]. The bandgap energies (E g ) of BCN and Vc-UPCN are evaluated by the equation αhν = A(hν-E g ) 1/2 , in which BCN and Vc-UPCN correspond to 2.5 eV and 2.69 eV(figures 6(b) and (c)). The valence bands (VB) of BCN and Vc-UPCN were obtained from the VB XPS spectra. As can be seen in figures 6(b) and (c), the VB edge of Vc-UPCN is located at approximately 1.8 eV, which is much more negative than that of BCN (2.2 eV). The negative shift of E VB is mainly due to the synergistic effect of carbon-defects ultra-thin structures, which promotes the reduction of electrons and greatly facilitates the activation with molecular oxygen [29]. The conduction band (CB) of BCN and Vc-UPCN can be calculated using E CB = E VB -E g . The value of CB potentials is −0.89 eV and −0.3 eV for Vc-UPCN and BCN ( figure 6(d)).

Exciton dissociation and charge transfer abilities
In organic semiconductors, owing to the strong Coulomb forces, photogenerated electrons, and holes tend to assemble to produce excitons, which is unfavorable for the production of ROS; therefore exciton dissociation is important. PL is commonly used to detect photogenerated electron-hole complexes. As illustrated in figure 7(a), BCN has the strongest emission peak at 460 nm due to the radiative complex effect of self-trapped excitons, while Vc-UPCN possesses a weaker peak intensity at 478 nm, indicating that carbon defects and ultra-thin nanostructures are favorable to inhibit the compounding of photogenerated carriers. What's more, BC-1 showed the weakest peak intensity, implying that the BNQDs acted as hole extractors, acting in coordination with electron traps (carbon-defects structure) to strongly prevent the compounding of photogenerated electrons with holes, ultimately significantly reducing the number of excitons.
To further study the changes in the photocarrier dynamics, time-resolved photoluminescence (TRPL) decay spectra were obtained for BCN, Vc-UPCN, and BC-1. Their fluorescence decay curves can be fitted well by the following double exponential function: where τ 1 and τ 2 are the fluorescence lifetimes and B 1 and B 2 are the respective PL amplitudes. The average radiative lifetime (τ ave ) was calculated according to the following equation As depicted in figure 7(b), the τ ave for BCN is 9.45 ns, compared with 9.68 and 12.78 ns for Vc-UPCN and BC-1, respectively. In contrast to BCN, the extended lifetime of Vc-UPCN is associated with the increased separation efficiency of photogenerated carriers due to the concerted effect of its ultra-thin structure and carbon defect structure. Furthermore, the separation efficiency was further enhanced after the attraction of holes by Boron nitride quantum dots. It is known that an extended lifetime increases the likelihood that photogenerated electrons and holes will be trapped by the active material and can generate more reactive oxygen species.
In addition, transient photocurrents show similar results, which are often employed to represent the interfacial charge transfer behavior. Figure 7(c) depicts the photocurrent time signatures of Vc-UPCN and BC-1 at several illumination switching cycles (λ > 420 nm). The construction of a dual channel transfer can remarkably promote the separation and migration of photogenerated carriers. Therefore, a higher current density is available in BC-1 than in the single-channel Vc-UPCN.

Photodegradation analysis
Characterization of photocatalytic activity of samples by visible light degradation of methylene blue solutions (20 mg L −1 , λ > 420 nm). Figure 8(a) depicts the variation in MB concentration (C t /C 0 ) with exposure time for the BCN, Vc-UPCN, and BC-1 photocatalysts under visible light. C 0 represents the original concentration of MB under dark conditions, and C t is the MB concentration after the photocatalytic reaction time t. It is clearly noticed in figure 8(b), the intensity of MB in the BC-1 solution on the 664 special absorption peak was significantly decreased and even close to 0 compared to other samples. In addition, the degradation rate of MB (DE,%) was computed by the following formula and is displayed in figure 8(c).
he degradation efficiencies of BCN and Vc-UPCN after 40 min of reaction were 43% and 88%, respectively. The degradation efficiency of BC-X samples was greater than that of both BCN and Vc-UPCN. With an increase in the BNQDs ratio, the photocatalytic degradation efficiency of BC-X showed an upward trend. The highest degradation efficiency (97%) was observed for MB in BC-1. However, the photocatalytic activity of BC-X decreased when the BNQDs surpassed the optimal ratio, and the BNQDs content in this nanocomposite was not proportional to the photocatalytic activity because the excess BNQDs masked the active sites of the Vc-UPCN. The degradation rate (k) of MB in BC samples can be expressed by this formula.
As shown in figure 8(d), the curves of ln(C 0 /C t ) and t were linear, indicating that the photocatalytic degradation of MB followed the pseudo-first-order model. Obtained by the slope, the value of k is BC-1 (0.107 min −1 ) > BC-0.5 (0.07 min −1 ) > BC-1.5 (0.059 min −1 ) > Vc-UPCN (0.053 min −1 ) > BCN (0.014 min −1 ). The BC-1 photocatalyst has the highest apparent rate constant of 0.107 min −1 , which was twice that of Vc-UPCN. The enhanced photocatalytic performance arises from the introduction of BNQDs, as BNQDs synergize with the carbon-defects structure to construct dual channels, facilitating the separation of photogenerated carriers and the activation of molecular oxygen.
The reuse of the catalyst is also an important characterization of the catalyst performance, as shown in figure 8(e), after three reuses, BC−1 still exhibited a good degradation effect.
In addition, to investigate the contribution of various active substances to MB degradation during photocatalysis, N 2 , EDTA-2Na, and IPA were used as ·O 2 − , h + , ·OH, and trapping agents, respectively. When large amounts of N 2 and 10 mM IPA were added to the photocatalytic system, the degradation efficiency of MB was significantly reduced from 97% to 65% and 73%, respectively. However, 10 mM EDTA-2Na resulted in a Based on the above results and analysis, the photodegradation mechanism of BC-1 is shown in figure 10. Under visible-light excitation, BCN generates photogenerated electron-hole pairs, but due to the strong Coulomb force, the photogenerated electrons and holes are very easy to compound and generate excitons. Thanks to the carbon-defects structure, as well as the ultra-thin structure, the migration rate of photogenerated electrons-holes is enhanced to suppress photogenerated carrier compounding; owing to the different typical compounding times of photogenerated electrons and holes, a large number of excitons are still generated. For this reason, BNQDs are added, and because of the negatively charged edge functional groups on BNQDs, not only are the double channels constructed in concert with the carbon-defects structure to promote the reverse migration of photogenerated electrons and holes, as well as the exciton dissociation. During photocatalytic degradation, photogenerated electrons (e − ) interact with oxygen in the water to form    Therefore, thanks to the dual charge transfer channel constructed by Boron nitride quantum dots and carbon-defects structure, not only is the dissociation of carbon nitride excitons promoted, but the migration rate of photogenerated carriers and the activation of molecular oxygen are enhanced.
Finally, we added a table about the activity of BNQDS/Vc-UPCN with other carbon nitride-based photocatalytic materials for the degradation of methylene blue (table 1). By comparison, it can be found that BNQDs/Vc-UPCN exhibit excellent catalytic activity. This may be because the composite uses ultra-thin carbon nitride as a carrier, which increases the contact area between the catalyst and methylene blue. Meanwhile, the Boron nitride quantum dots and the carbon-defective structure showed good synergistic effects during the catalytic reduction process.

Conclusion
In this study, Vc-UPCN was synthesized by a one-step calcination method, and BNQDs were introduced to construct dual channels to synergistically enhance the photocatalytic performance of Vc-C 3 N 4 . As a result, the photogenerated electrons and holes on the Vc-C 3 N 4 surface were directed toward the carbon-defects structure and BNQDs, respectively, due to the attraction of electrons by the carbon-defects structure and the trapping of holes by the BNQDs, thus extending the carrier lifetime and enhancing the ability of the photocatalyst to activate molecular oxygen. This study reveals how carbon-defects structures and Boron nitride quantum dots can synergistically enhance the photocatalytic performance of carbon nitride and also provides a strategy for designing low-cost dual-channel mechanism photocatalysts.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).