NC State
Wu, J., Zhao, L., and Song, G. (2024). “Carbon dots prepared from waste wood and residual adhesive and their use as catalysts for hydrogen production,” BioResources 19(2), 2417-2435.



Large quantities of waste wood with residual resin adhesives are not recycled efficiently. To address this issue, waste wood with residual resin adhesives was synthesized into carbon dots (CDs) via a facile hydrothermal self-assembly method to enhance the H2 evolution performance of graphite C3N4 (g-C3N4), a metal-free photocatalyst. Among all samples, the most significant enhanced sample was MCN-UF-3.5, which has an H2 evolution of 22.1 mmol·g-1·h-1, which is 3.91 times that of unmodified g-C3N4. The band gap and recombination of photogenerated charges were both improved by the doping of CDs. Meanwhile, the DFT calculation showed that adding CDs, especially with the -NH2 group, can significantly deform the structure and destroy the symmetry. This consequence implies an enhancement in the activity of the hydrogen evolution reaction, confirming the feasibility of the modification.

Download PDF

Full Article

Carbon Dots Prepared from Waste Wood and Residual Adhesive and Their Use as Catalysts for Hydrogen Production

Ji Wu,a Liang Zhao,a,* and Guohui Song b

Large quantities of waste wood with residual resin adhesives are not recycled efficiently. To address this issue, waste wood with residual resin adhesives was synthesized into carbon dots (CDs) via a facile hydrothermal self-assembly method to enhance the H2 evolution performance of graphite C3N4 (g-C3N4), a metal-free photocatalyst. Among all samples, the most significant enhanced sample was MCN-UF-3.5, which has an H2 evolution of 22.1 mmol·g-1·h-1, which is 3.91 times that of unmodified g-C3N4. The band gap and recombination of photogenerated charges were both improved by the doping of CDs. Meanwhile, the DFT calculation showed that adding CDs, especially with the -NH2 group, can significantly deform the structure and destroy the symmetry. This consequence implies an enhancement in the activity of the hydrogen evolution reaction, confirming the feasibility of the modification.

DOI: 10.15376/biores.19.2.2417-2435

Keywords: Waste wood; Residual adhesive; Carbon dots; g-C3N4; Photocatalytic; DFT

Contact information: a: Co-Innovation Center of Efficient Processing and Utilization of Forest Resources, College of Materials Science and Engineering, Nanjing Forestry University, Nanjing, 210037, China; b: School of Energy and Power Engineering, Nanjing Institute of Technology, Nanjing, 211167, China;

* Corresponding author:;



A screenshot of a computer Description automatically generated


The proliferation of waste wood-based panels is a problem, but these materials could be used as valuable resources (Farjana et al. 2023). Due to the chemical pollutants and resin adhesives, the traditional disposal methods, such as landfilling and burning, are low-efficiency and high cost (Girods et al. 2008; Kim and Song 2014). These residual resins cannot be eliminated easily (Zhong et al. 2017). The residual resins hinder the modulus of rupture, modulus of elasticity, and internal bonding strength of remanufactured particleboards (Czarnecki et al. 2003). The residual resin and other compounds that are not conducive for recycling are difficult to handle efficiently and cost-effectively in the current common treatments. Therefore, there is an urgent need for a better way to utilize waste wood and its residual adhesive compounds.

Carbon dots (CDs) derived from biomass materials are emerging carbon nanomaterials with the advantages of advances in synthesis (Bian et al. 2016), hydrophilicity (Ehtesabi and Massah 2021), tunability (Ma et al. 2016), and optical properties (Liu et al. 2013), as well as a wide variety of raw materials, especially biomass (Li et al. 2020), whose sources are extensive and low-cost. CDs are commonly used in ion detection (Yang et al. 2022b), fluorescence imaging (Liu et al. 2017), photocatalysis (Sbacchi et al. 2023), and other applications. The main structure consists of a core composed mainly of sp2 hybridized regions, and a shell composed of sp3 hybridized regions (Zhu et al. 2015). The shell composed of sp3 endows CDs with the sites to bind functional groups such as -OH, -C=O, -NH2, etc., which makes them well tunable (Xia et al. 2019). The sp2-enriched nucleus of CDs is essential for photocatalytic reactions because of its large specific area and outstanding electron conductivity (Li et al. 2018).

Photocatalytic reactions include photocatalytic degradation and hydrogen evolution. The main principle of the photocatalytic hydrogen evolution reaction (Jafari et al. 2016) is the use of a catalyst to absorb photons to produce electron-hole pairs. Subsequently, H+ formed at the catalytic site of the catalyst reacts with the electrons to produce hydrogen. Compared to metal catalysts, which are expensive and prone to secondary pollution, the two-dimensional material with rich sp2 hybridization g-C3N4 is a more accessible and environmentally advantageous option. The g-C3N4 has the advantages of suitable band gap in visible light region, non-toxic feature, high chemical stability, and good elastic and tensile strength (Han et al. 2020), which means that it is suitable for photocatalytic hydrogen evolution (Zhao et al. 2015). However, the photocatalytic efficiency of the pristine g-C3N4 is unsatisfactory due to small specific surface area and easy complexation of photogenerated electrons (Mamba and Mishra 2016). At present, the main improvement directions for photocatalysts include (Fajrina and Tahir 2019) doping with heteroatoms, which can change the band gap structure, to enhance the absorption capacity in various wavenumber regions. Another option is to utilize the microstructure of nanoparticles with good mechanical properties and compatibility (Jia et al. 2023), such as SiC, as skeleton loading photocatalysts reduces the recombination of photogenerated electron-hole pairs (Wang et al. 2017).

Wood is made up of cellulose, semi-cellulose, and lignin. Cellulose was studied to be a good carbon source to prepare CDs (Souza et al. 2018). Lignin can offer O atoms and functional groups to enhance CDs performance. In addition, CDs prepared from waste wood can reduce preprocessing steps. So, the work in this study focused on providing a new route for the reuse of waste wood with residual adhesives. Carbon dots, which were prepared from waste wood, was used to modify g-C3N4 in order to enhance its photocatalytic H2 evolution performance. Meanwhile, adding different kinds and ratio of adhesives to adjust photocatalysts performance, including phenolic (PF) resins, urea-formaldehyde (UF) resins, and melamine-formaldehyde (MF) resins. First principle calculation is used to emulate surface of photocatalysts to verify the mechanistic feasibility of the modification in this study.



Melamine (≥ 99.6%, CP) was purchased by Lingfeng Chemical Reagent Co. Ltd. (Shanghai, China). Triethanolamine (TEOA, ≥ 99.0%, AR) was purchased by Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). Chloroplatinic acid hexahydrate (≥ 98.0%, AR) was purchased by Rhawn Reagent Co. Ltd. (Shanghai, China). Waste poplar was purchased by Maoyou Timber Co. Ltd. (Jiangsu, China). Urea-formaldehyde resin (UF), phenolic resin (PF) and melamine-formaldehyde resin (MF) were purchased by Zhongcheng Plastic Co. Ltd. (Zhejiang, China).

Synthesis Methods

The preparation method of CDs followed and was adapted from the hydrothermal method in previous studies (Wareing et al. 2021). The holistic synthesis process of CDs, g-C3N4 and photocatalyst is shown in Fig. 1(A). Specifically, 80 mesh poplar wood powder was mixed with different types of resins to form a mixture with a total weight of 2.5 g. The types of resins used in this study were urea-formaldehyde (UF), phenol-formaldehyde (PF), and melamine-formaldehyde (MF). The weight ratio of the resins in the mixture was 2.5%, 3.5%, and 4.5%, respectively. The mixture was dispersed in 50 mL deionized water. Then, after thorough stirring, the solution was transferred to a 100 mL Teflon-lined autoclave and heated at 180 °C for 8 h. Moreover, the suspension was centrifuged at 10000 rpm for 15 min to remove the bulk particles, leaving only the supernatant. Subsequently, the supernatant was dialyzed for 24 h using a dialysis bag with a molecular weight cut-off of 1000. Finally, the solution was lyophilized for 48 h and the obtained powder was collected and stored in a drying dish.

A total of 10 g of melamine was placed in a crucible of alumina and calcined in a tube furnace at 550 °C by 10 °C·min-1 heat rate for 4 h and annealed at 520 °C for 2 h. The samples were allowed to cool naturally to room temperature, and ground into 150-mesh powder.

For the preparation of photocatalysts obtained by functionalization of g-C3N4 coupled CDs, 100 mg of g-C3N4 was mixed with 10 mg of different kinds of CDs powder in 50 mL deionized water, which was then performed 2 h ultrasonic treatment and magnetically stirred for 2 h. Thereafter, the solution was transferred to a Teflon-lined autoclave and heated at 180 °C for 2 h. Following, the suspension was centrifuged at 10000 rpm for 2 min to obtain the lower precipitate after solid-liquid separation. Collected precipitate was lyophilized for 12 h. The final obtained powder is the photocatalyst denoted as MCN-x-y (modified g-C3N4). The x represents the kinds of resins added in preparation of CDs, and y represents the ratios of the CDs system. Samples without the CDs powder are named as g-C3N4, and ones with added unmixed CDs powder are named as MCN-PP (poplar powder-derived CDs).


SEM (Reguluss8100, Hitachi), BET (BSD-PS(M), BEISHIDE), Raman spectrum (DXR532, Themor), FT-IR spectroscopy (VERTEX 80V, Bruker) and XPS (AXIS UltraDLD, Shimadzu) were used to obtain the morphology, microstructure, and surface elements chemical states of the samples. The optical and photoelectrochemical properties were characterized by a UV-vis diffuse reflectance spectra (Lambda 950, PE) and a photoluminescence (PL) emission spectra (FluoroMax-4, HORIBA). Hydrogen production was analyzed using a gas chromatograph (GC9790 Plus, Fuli).

Photocatalytic Hydrogen Evolution Experiment

For the photocatalytic H2 evolution experiments, a 350 W cold xenon lamp with a light intensity set to 1100 W·m-2 was used as a light source to simulate solar irradiation. The reaction solution consisted of deionized water, 10 mL of TEOA, and H2PtCl6 at 3wt% to make up a total of 100 mL of the system. Next, 100 mg of catalyst was added to the solution, followed by purging with high-purity argon gas at a flow rate of 50 mL·min-1 and magnetic stirring for 60 min in a dark environment. After turning on the light source, the reaction gas products were collected using an aluminum foil gas bag, which was changed every 10 min. The gas products were passed into the GC for analysis.

DFT Calculation Details

The Cambridge Serial Total Energy Package (CASTEP) modules in BIOVIA Material Studios were used to emulate photocatalyst structures (Segall et al. 2002). All theoretical simulation calculations adopted Perdew-Burke-Ernzerhof (PBE) (Perdew et al. 1996a) functional within the generalized gradient approximation (GGA) (Perdew et al. 1996b). The ultra-soft pseudopotential and the norm-conserving pseudopotential were chosen to complete different parts of the calculations. The former has looser convergence conditions and was used for geometry optimization. The latter has stricter convergence and can obtain more accurate results. The properties of photocatalyst structures, such as band structure, the density of states (DOS), and potential, were calculated by the norm-conserving pseudopotential. Considering the Van der Waals interactions, the TS custom method was used for DFT+D to correct (Tkatchenko and Scheffler 2009). The cut-off energy was set with 400 eV for geometry optimization and 500 eV for properties (Yang et al. 2022a). Kohn-Sham wave functions were taken as a relativistic treatment. The parameters of convergence tolerance (Yang et al. 2022a): energy, max force, and max displacement were set as 1×10-5 eV·atom-1, 3×10-2 eV·A-1, and 1×10-3 A respectively, and SCF tolerance was set as 1×10-6 eV·atom-1. In addition, given that hybrid functions have better performance in the band structure and DOS, HSE06 (Heyd et al. 2006), a hybrid function, was separately adopted in corresponding calculations.


Morphology and Structure Characterization

SEM is commonly used to detect the microstructure and morphology of micrometer-sized catalysts. As displayed in Fig. 1(B), the major component of MCN-UF-3.5 photocatalyst was bulk g-C3N4, and the diameter was 15 to 20 μm.

Fig. 1. (A) Synthesis of CDs, g-C3N4 and photocatalyst; (B, C) SEM images of MCN-UF-3.5

Numerous microparticle CDs were dispersed and attached to the surface of bulk g-C3N4. As shown in Fig. 1(C), the major component of the photocatalyst had lamellar features, composed of multiple ultrathin paper-fold sheets, which is the typical structure of g-C3N4. The morphology of the synthesized photocatalysts coincided with that reported by Cui et al. (2020). The morphology of CDs distributed on the surface was predominantly sphere-like or granular, matching the features in Singh’s study (Singh et al. 2023). To better demonstrate the features of CDs, TEM, and other microscopy analysis will be carried out in the coming research. By measuring the particles in Fig. 1(C), it was known that their diameter was concentrated in the range of 1.4 to 0.2 μm or even smaller. These results suggested that the photocatalyst had the fundamental g-C3N4 structure, which could absorb radiation and emit electrons (Zhang et al. 2021), and the combination of CDs modified the surface.

The BET specific surface area of g-C3N4 and its modified samples were characterized by N2 adsorption-desorption experiments. The isothermal curves are plotted in Fig. 2(A). As depicted in Fig. 2(A), both g-C3N4 and CDs doped samples exhibited type IV isothermal curves that did not reach the saturation plateau (Brunauer et al. 1938), indicating the presence of mesoporous architecture. The data revealed that MCN-UF-3.5, with the most significant specific surface area of 38.6 m2·g-1. The surface area of MCN-UF-3.5 was larger than that of g-C3N4 (10.3 m2·g-1) by 3.74 times. The previous studies reported that increased specific surface area is helpful to enhance photocatalytic performance (Alshammari et al. 2023). Therefore, the doping of CDs increased photocatalysts’ specific surface area, making the 2-dimensional structure turn into the 3-dimensional structure (Zhu et al. 2017). This means that adding CDs improved the efficiency of charges transferring across the interface, positively affecting photocatalytic H2 evolution.

Raman scattering is often used to obtain vibrational information specific to the chemical bonds in molecules. The Raman spectra in the work were excited at 780 nm using a dual laser system. Figure 2(B) shows Raman spectra of g-C3N4 and MCN-x-3.5 (x=PP, UF, MF, PF). MCN-UF-3.5 outperformed all other samples in terms of holistic intensity. The most substantial peak of samples in Fig. 2(B) was the peak located at 700 cm-1. It is generally noted by the prior studies that this 700 cm-1 peak is a typical one caused by triazine ring vibrations with bent surface (Li et al. 2015). There was a weaker peak at 750 cm-1, which was not universally presented in g-C3N4. Zinin et al. (2009) found that the 750 cm-1 peak mainly relates to the properties of raw materials used to synthesize. According to the report (Zinin et al. 2009), one typical peak of melamine, the raw material of the catalyst, was located around 750 cm-1, which fits the regularity. Besides, the peak at 975 cm-1 had discrepancies with the simulation report results (Gonze et al., 2002) but was identical with Zinin’s report (Zinin et al. 2009). As has been stated (Deifallah et al. 2008; Kroke et al. 2002), it was ascribed to the s-triazine breathing mode or polymorphs. As for the polymorphs, it was probably formed due to the destruction of the original structure of g-C3N4 by the ultrasonic treatment during synthesis. The peak with high intensity near 1340 cm-1 was a typical D band because of defects on the catalyst surface (Saraswat and Yadav 2020). These defects were caused by the surface modification by CDs, enhancing the sp3 hybridization of the region. The cause of the peak formation at 1410 cm-1 was similar to the peak at 750 cm-1. It was attributed to breathing and stretching of the ring (Meier et al. 1995). Furthermore, the undulating peak at 1590 cm-1 was the G band, mainly related to the sp2 hybridization of C atoms (Grimm 2008). The prominent intensity of the peak at 1590 cm-1 of MCN-UF-3.5 clarified the higher electron conductivity than others (Luo et al. 2019). It may be deduced from the phenomenon that MCN-UF-3.5 possessed a better capacity for H2 evolution. The curve of g-C3N4 and MCN-PP had a more apparent G peak at 1410 cm-1 than other samples doped CDs. This phenomenon can be interpreted as related to ultrasonic treatment and doping of CDs. They did not destroy the structure of the triazine ring and s-triazine ring but accounted for the low intensity of the G peak.

Fig. 2. (A) Nitrogen adsorption-desorption isotherms of samples; (B) Raman spectra (785 nm) of photocatalysts after background and offset corrections; (C) FT-IR spectra of g-C3N4 and different doped samples; (D) XPS survey spectra of samples; (E, F) High-resolution N1s and C1s spectra of photocatalysts

Analysis of information on molecule vibrations in FT-IR spectra referred to identifying the material molecules and functional groups on the surface. The FT-IR spectra of g-C3N4 and other samples doped with CDs are exhibited in Fig. 2(C). The key messages in Fig. 2(C) were concentrated in three sections, which were a section located in the range of 3255 to 3050 cm-1, a section ranging from 1650 to 1200 cm-1, and a section within 900 to 800 cm-1 wavenumber range (Li et al. 2021). The first consisted of three scattered weak peaks: 3255, 3165, and 3082 cm-1. Spectra in this area were analyzed to illustrate the information of the stretching vibrational characteristic of the substance. The peaks of three samples doped with diverse CDs had a higher intensity than the others. In particular, the peak at 3255 cm-1, thought to be caused by the stretching of the O-H group (Cui et al. 2020; Li et al. 2021), was too weak to recognize in the pristine g-C3N4 but was notable in doped samples. Adding CDs synthesized by poplar powder and resins brought more O atoms into the photocatalyst, making the peak visible. The peak at 3165 cm-1 was regarded as the result of N-H stretching vibrations in uncondensed amino groups (Majdoub et al. 2020). There was an apparent disparity in peak intensity between MCN-UF-3.5 and the pristine g-C3N4. The stretching vibration of C-H bonds generally forms the peak located around 3040 cm-1 (Coates 2000). However, when the C atoms were located at the end of the molecular structure, C-H2 bonds, whose stretching vibrational peak was found at 3085 cm-1 nearby (Ogita et al. 2004), would be formed. It can be inferred that C atoms at the edge positions in the structure destroyed by ultrasonic treatment reformed C-H2 bonds during hydrothermal synthesis. The section, which was brought about by C-N heterocyclic vibrations ranging from 1650 to 1200 cm-1, was subdivided into six peaks located at 1637, 1562, 1463, 1402, 1325, and 1236 cm-1. The peak at 1637 cm-1 corresponded to the C=C bond of the aromatic ring skeleton, and the peak at 1469 cm-1 referred to the deformation of -CH2 or the antisymmetric deformation of -CH3 (Li et al. 2021). The last section, covering 900 to 800 cm-1 and composed of two peaks, represented the typical breath mode of triazine and s-triazine ring units, the skeleton of g-C3N4. The above analysis elaborated that the doping of CDs maintained the skeleton of g-C3N4. On this foundation, the treatment inducted more O atoms and reinforced the sp3 hybridization of the surface.

The surface elemental chemical states of g-C3N4 and samples doped with CDs were detected using XPS. Figure 2(D) shows a survey scan spectra of five samples, which could be divided into three elemental solid components: O1s, N1s, and C1s. Among them, the intensity of O1s peak on g-C3N4 was weaker than others, and its peak area only occupied 4.6% of the total peak area of g-C3N4. The doping of CDs significantly increased the intensity of O1s, especially MCN-PF-3.5. The O1s peak area of MCN-PF-3.5 was up to 18.7% of the total peak area, owing to rich oxygenated functional groups on phenolic resin. The characteristics of the N1s spectra of four CDs doped samples were consistent. As presented in Fig. 2(E), five peaks located at 398.2, 398.9, 399.6, 400.8, and 404.2 eV were associated with sp2 hybridization in C-N=C, pyridine N, N-(C)3, N-H and charge effects triggered by stacking of π-bonds, respectively (Cheng et al. 2017; Fang et al. 2017). For the unmodified g-C3N4, the lower O content than the others blurred the boundary between the pyridine N peak and the N-(C)3 peak, resulting in a distinct peak at 399.3 eV. This distinction meant that there was a weak modification of CDs adjusting the structure of g-C3N4, thus promoting photocatalytic performance. The C1s peaks of g-C3N4 and MCN-x-3.5 were divided into four and five sections, individually. The shared segments were located at binding energies of 284.2 and 285.0 eV, on behalf of the sp2 hybridization part of carbon and the sp3 hybridization part of C-C bonds (Chernyak et al. 2020). The trend demonstrated in Fig. 2(F) could be explained by the fact that these two peaks of CDs doped samples had higher intensity than the pristine g-C3N4. Oxygenated functional groups were attached to the surface broadened these peaks, increasing the difficulty of separating peaks. The peak located at 286.4 eV was not present in the unmodified g-C3N4 but in modified samples, indicating the C-O bonds (Li et al. 2021). There were two peaks at 287.9 and 288.2 eV. The former was related to the C=O bond, and the latter meant the sp2 hybridization of N-C=N (Wang et al. 2018). There was a noticeable difference in the area of the peak located at 287.9 eV between samples of modified CDs with adhesives and others. O atoms generally were present in cured resins mixed with wood powder in the configuration of C-O-C, C-O-N, and O-H bonds. Although the solution was not in acidic conditions, the hydrothermal synthesis of waste wood with residual adhesives was reacted at higher temperatures and pressures than in the study of Liu et al. (2018). Due to the hydrolysis of cured resins, the original chemical bonds were broken, and C=O bonds were formed in such conditions. As described by Lukowsky (2002), ether bridges (HNCH2-O-CH2NH) in MF resins are more stable than methylene bridges (NHCH2-NH). It was hard for O atoms to be restricted to the ether bridges to form C=O bonds. This concept was able to account for the lower peak intensity at 287.9 eV of MCN-MF-3.5 than samples modified by UF and PF. The peak unique to the unmodified g-C3N4 and MCN-PP had higher separation than others located at 287.6 eV, denoting the existence of the C=N combination. It could be concluded that the analysis of XPS spectra for each sample coincided with the results of FT-IR spectra. They both elaborated that the doping of CDs had modified the combination of chemical bonds on the surface without destroying the classical structure of the pristine g-C3N4, which enhanced the hydrogen evolution capacity.

Optical and Photoelectrochemical Properties

The photocatalyst, which occurred photon-excitation, caused the transition to excited states. Then, the emission of light or luminescence formed by the energy or photons released through the relaxation process is photoluminescence (PL). Thus the PL emission spectra were studied to understand the separation of photogenerated electron-hole pairs and recombination to study the optical properties of photocatalysts further. As shown in Fig. 3(A), the intensity of the unmodified g-C3N4 was much higher than that of others. The higher PL intensity of the photocatalyst represented more energy released through the recombination of electron-hole pairs than modified samples. This component energy was wasted and not transmitted to the photocatalyst surface for the hydrogen evolution reaction. The PL intensity of the unmodified g-C3N4 indicated the lower hydrogen evolution performance than others. Remarkable results relative to the problem of recombination of photogenerated electron-hole achieved by doping CDs could be deduced from Fig. 3(A). The data of time-resolved transient PL decay measured the lifetime of charge carriers. A shorter lifetime of charge carriers meant a faster rate of photogenerated electrons transferred to the photocatalyst surface.

Equation 1 calculates the final average lifetime of PL (Zhang et al. 2022),


where τ1 and τ2 are the first-order fitting and the second-order fitting lifetime, and B1 and B2 correspond respectively to their normalized amplitudes (Chauhan et al. 2016). As shown by data listed in Table 1, modified samples obtained a shorter average PL lifetime than the unmodified g-C3N4, especially MCN-PF-3.5. Comparing MCN-PF-3.5 with the unmodified g-C3N4, the reduction of the PL lifetime was up to 31.2%. This result manifested that the time for the recombination of charge carriers became much less, and due to the increase in the utilization of photogenerated electrons, the number of unrecombined electron-hole pairs increased.

Table 1. Average Lifetime of Samples after Second-Order Fitting

As a standard method to analyze the optical properties, UV-Vis DRS was detected to further investigate the electronic transition between energy levels g-C3N4 and samples doped with CDs. On the basis of the intensity that emerged in Fig. 3(B), it can be inferred that the doping of CDs obviously raised the absorbance of related samples in the vicinity of 275 nm and the visible light region. A recognizable peak appeared at 375 nm of all samples, adding CDs with residual resins. The peak was located in the near-ultraviolet region, symbolizing the transition of n-π*, which was contributed by chromophores, e.g., C=O, C=N (Fang et al. 2017; Selmi et al. 2023). The feature of UV-vis DRS spectra of samples containing adhesives conformed to that of C1s spectra. In addition, the presence of chromophores made red shift occur on the highest characteristic peak due to the promotion of the conjugation effect. Consistent with Shibayama et al. (2019), the better absorbance performance of the photocatalyst in the visible light region can drive photocatalytic evolution better because of the more efficient utilization of irradiation. The curves of Tauc plots (Davis and Mott 1970), whose data was solved by Eq. 2 (Chen et al. 2016) for the g-C3N4 direct transition photocatalysts, are plotted in Fig. 3(C),


where Eg is band gap energy, h is Planck’s constant, ν represents light frequency, and A means a constant. As seen in Fig. 3(C), the band gap energy of several samples were 2.79, 2.74, 2.73, 2.65, and 2.78 eV, corresponding to the unmodified g-C3N4, MCN-PP, MCN-MF-3.5, MCN-UF-3.5, and MCN-PF-3.5, respectively. It was easier for the photocatalyst with a narrower band gap to absorb the charge with lower energy and react. Based on the results, the doping of CDs, particularly with UF resins, significantly narrowed the band gap and made photocatalytic evolution react more efficiently.