Band gap and photo charge carrier tailoring in zirconium doped carbon nitride using ZrCl₄–DMF–melamine for photocatalytic degradation of rhodamine B

Abstract

Improving the performance of carbon nitride (CN) photocatalysts in photocatalytic degradation applications involves optimizing their morphology, electronic properties, and optical characteristics. Zirconium-doped carbon nitride (Zr-doped CN) photocatalysts were synthesized using dimethylformamide (DMF) as a solvent to facilitate the formation of complex molecular structures for effective metal doping. By varying the concentration of the zirconium tetrachloride (ZrCl₄) precursor between 1 and 3 mmol, we observed significant enhancements in photocatalytic activity. Notably, controlling the ZrCl₄ concentration below 3 mmol prevented the formation of zirconium oxide phases, which could otherwise negatively affect the photocatalytic performance. Zr incorporation led to the morphological transformation of CN from a bulk structure into a hierarchical porous structure, increasing the surface area to 135 m² g⁻¹. Additionally, Zr doping changed the band energy and electronic properties, creating an optimal energy level for generating oxygen radicals in the photocatalytic water-splitting processes. The photocatalytic degradation of rhodamine B showed that the Zr-doped CN photocatalysts achieved 4.5-fold better performance than undoped CN. Moreover, a small amount of ethylenediaminetetraacetic acid (EDTA) significantly enhanced the photocatalytic efficiency of Zr-doped CN compared to that of undoped CN. These results indicate that combining Zr-doped CN with other materials to create Z-scheme or S-scheme structures could further enhance its performance, thus emphasizing the potential of increasing photocatalytic efficiency by optimizing energy band structures and forming heterostructured photocatalysts.

---

1 Introduction

Photocatalytic dye degradation using photocatalytic materials has attracted considerable interest as a promising solution for reducing pollution and contamination. The photocatalytic degradation of pollutants and organic dye molecules can be induced by oxidative radical species when light irradiates the surface of the photocatalyst, with the rate of degradation in the photocatalytic reaction determined by the optical, electronic, and morphological properties of the photocatalyst. Therefore, improving the efficiency of photocatalytic materials is an important research topic in this field. Among the various photocatalyst materials, metal and metal compound nanoclusters,^1,2 TiO₂, ZnO,^3–5 and carbon nitride (CN)^6–9 have attracted considerable interest. In particular, the CN structure, characterized by carbon and nitrogen atoms bonded through sp² hybridization forming π-conjugated graphitic planes with heptazine heterocyclic rings, provides several advantages, including a moderate bandgap (∼2.7 eV) for tri-s-triazine-based CN, sheet-like morphology that enhances surface modification, high stability, and environmental non-toxicity. However, despite these favorable intrinsic properties, CN faces significant limitations, such as a rapid charge carrier recombination rate, low specific surface area, and insufficient charge carrier mobility, all hindering its photocatalytic performance.

To overcome these challenges, recent studies have focused on optimizing the photocatalytic properties of CN by narrowing its bandgap, increasing the number of active sites for catalysis, enhancing the charge-carrier concentration, and improving the charge-separation efficiency under light excitation. Additionally, morphological modifications such as creating rod-like or highly porous structures have been explored to increase the specific surface area.^10,11 Furthermore, doping CN with various elements to modify its electronic and optical properties, including narrowing the bandgap, improving light absorption, and enhancing photoinduced charge separation, has emerged as a highly effective approach for significantly enhancing photocatalytic efficiency.

The doping of various atoms into the CN structure has been investigated as a strategy to improve photocatalytic performance, particularly for water-splitting reactions. Non-metal atoms such as S, O, Cl, P, and B have been incorporated to enhance the optical and physical properties of CN.^12–16 Moreover, the doping of metal atoms or ions has gained attention owing to its advantages of providing additional binding sites and active sites for separating photogenerated charge carriers.

Previous studies have explored the doping of several types of metal atoms or ions, including alkaline metals, such as Na and K,^17,18 and transition metals, such as Fe, Mn, Se, V, Zn, and Zr.^19–28 Doping with Zr atoms/ions has attracted interest because of its cost-effectiveness and high catalytic activity, making Zr a promising alternative to other transition metals for the modification of CN. Owing to its high work function, Zr facilitates electron transfer from the semiconductor to the Zr surface, promoting the efficient separation of photogenerated electron–hole pairs. Zr also serves as an electronic mediator in composite photocatalysts and promotes the formation of all-solid-state Z-scheme heterojunctions.²⁰ Additionally, the doping of Zr⁴⁺ ions can induce the mesoporosity of semiconductor materials such as TiO₂ and CeO₂.^29–32 To the best of our knowledge, only a few studies have been conducted on doping with Zr atoms or ions. For instance, Zr⁴⁺ ions have been introduced into the CN structure using sulfur (S)-containing precursors,¹⁹ or zirconium nitrate pentahydrate,^20,22 a dissociable salt, has been employed. Additionally, the dissociation of Zr⁴⁺ from Zr compounds can be induced by incorporating organic molecules that form complex molecular structures. Dimethylformamide (DMF) is widely used as a solvent in the synthesis of metal–organic frameworks (MOFs), including Zr-based materials such as UiO-66.³³ Other studies have investigated the molecular complexes formed between urea and DMF via recrystallization.³⁴ These complexes resulted in CN with an increased surface area owing to the hydrogen-bonding network. This enhancement improves light absorption over a broader range, boosting photocatalytic efficiency.

In this study, Zr⁴⁺ ions were doped into CN as complex molecular structures of Zr and melamine (ZrCl₄–melamine–DMF complex, ZMD complex). DMF acts as a Lewis base, with an oxygen group capable of donating electrons. In contrast, ZrCl₄ forms a linear polymeric structure, where each Zr atom adopts an octahedral geometry (coordinated with six chloride ions) and is connected via two chloride bridges, thus providing it with Lewis acid characteristics. The ability of DMF to dissolve the polymeric structure of ZrCl₄ also reduced the size and dissociation of Zr⁴⁺ ions. Moreover, coordination between the DMF molecules and Zr atoms promoted the formation of a linear polymeric structure. Given the strong oxophilic nature of the Zr atom, the polymeric structure breaks down when anhydrous ZrCl₄ is dissolved in DMF, with the chloride bridges replaced by coordinated DMF molecules.³⁵ By forming a ZrCl₄–DMF–melamine complex structure and optimizing the ZrCl₄ concentration, we successfully synthesized Zr-doped CN for photocatalytic dye degradation. Zr-doped CN exhibits a high specific surface area, an optimized band gap for photocatalytic reactions, and high photocarrier concentrations, with a 4.5-fold photocatalytic efficiency increase achieved for rhodamine B (RhB) degradation compared to CN without Zr doping.

2 Results and discussion

2.1 Characterization of the Zr-doped CN photocatalyst

An overview of the research process is presented in Fig. 1, along with photographs and SEM micrographs of the samples, illustrating the effects of variations in ZrCl₄ concentration on the morphological and catalytic properties of the photocatalysts. A precursor composed of melamine, DMF, and ZrCl₄ was used in this study. In path 1, melamine undergoes thermal polycondensation at high temperatures without Zr doping, resulting in CN formation, which exhibited a band gap of approximately 2.7 eV and low carrier density. Clearly, the properties of the synthesized CN were not affected by DMF during the precursor preparation process. The CN morphology retained a yellow powder and bulk flake structure, as shown in Fig. 1a and b. In path 2, introducing ZrCl₄ at optimal concentrations (<2 mmol) resulted in the successful synthesis of Zr-doped CN with an optimized band structure for use as a photocatalyst. These doped materials demonstrated enhanced optical properties, including a narrow bandgap and increased carrier density, compared to undoped variants. Additionally, the color of the photocatalyst powder changed, and the morphology changed to a hierarchical porous structure with a high specific surface area, as shown in Fig. 1c and d for Zr₂-CN and Fig. S1† for Zr₁-CN, indicating modifications in their structural and electronic properties. In path 3, Zr₃-CN, synthesized with a higher concentration of ZrCl_4, formed a ZrO₂ phase within the photocatalyst. This transition resulted in the appearance of a white powder and substantial changes in the optical properties, including an increased bandgap and reduced carrier density. Notably, introducing ZrO₂ significantly affects the overall photocatalytic efficiency. When the ZrCl₄ concentration exceeded 3 mmol, as illustrated in Fig. S2,† the CN structure became undetectable in both the X-ray diffraction (XRD) pattern and FTIR spectra, whereas the formation of ZrO₂ was observed. The implications of varying the ZrCl₄ concentration for the photocatalyst performance will be further discussed in the next section. This morphological change induced by the ZMD complex significantly enhanced the photocatalytic activity of Zr-doped CN, thereby improving its overall efficiency as a photocatalyst. Morphological analyses revealed that incorporating Zr into the semiconductor photocatalyst structure enhanced the mesoporosity of the CN framework.^29,30

Fig. 1 A schematic view to show the overall effect of ZrCl₄ concentration in DMF solvent on properties of the synthesized photocatalyst, SEM micrographs, the photograph (with the same amount of 25 mg), and optoelectronic band structure of carbon nitride (a–c), optimal Zr-doped carbon nitride (d–f), and Zr-doped carbon nitride with amorphous ZrO₂ (g–i).

The distribution of Zr in the CN structure was demonstrated using EDX elemental mapping. Zr was uniformly distributed across the surface of CN, as shown in Fig. S3,† indicating that Zr doping, facilitated by using DMF as a precursor to form the ZMD complex structure, effectively synthesized Zr-doped CN with the potential to exhibit good photocatalytic performance. Furthermore, the transmission electron microscopy (TEM) and selected area electron diffraction (SAED) images are presented in Fig. S4.† The TEM images indicate that the structure of CN transformed from a sheet-like morphology into a more fibrous porous structure when doped with Zr, as achieved by using the ZMD complex. The electron diffraction pattern of undoped CN shows high crystallinity, as evidenced by the intense diffraction peaks corresponding to the (200) crystal plane. The electron diffraction patterns of CN, Zr₁-CN, and Zr₂-CN exhibit similar characteristics. However, in the case of Zr₃-CN, the distinct electron diffraction patterns revealed the presence of a cubic crystalline structure of ZrO₂ in the (200) and (111) planes,^36,37 indicating the formation of a ZrO₂ phase owing to the increased concentration of ZrCl₄ as a result of excess doping.

X-ray photoelectron spectroscopy (XPS) was conducted to determine the elemental composition and bonding characteristics of the Zr-doped CN photocatalyst. The surface survey spectra confirmed the presence of C, N, and O in CN, whereas C, N, O, Cl, and Zr were detected in Zr-doped CN. Survey scans of all elements and high-resolution XPS spectra of Cl 2p are presented in Fig. S5 and S6.† Detailed deconvolution of the C 1s, N 1s, O 1s, and Zr 3d XPS peaks is summarized in Tables S1–S3.†

From Table S1† summarizing the atomic percentages of the CN photocatalysts and the effect of Zr doping, undoped CN had a carbon-to-nitrogen atomic ratio of 0.71, indicating a C₃N₄ structure. When Zr was doped into the CN structure, a slight increase in the carbon-to-nitrogen ratio was observed (Zr₁-CN and Zr₂-CN). However, a significant change was observed in the Zr₃-CN sample, attributed to the high concentration of ZrCl₄ interacting with DMF, which led to the breakdown of the CN structure. This breakdown was likely due to the decomposition of Cl during calcination and the subsequent formation of bonds with C and N from the DMF molecules within the CN framework. The increase in the atomic percentage of oxygen further indicates the formation of Zr–O bonds, likely due to the interactions between Zr and the oxygen atoms in DMF. Additionally, the atomic percentage of Zr showed a marked increase at higher ZrCl₄ precursor concentrations. Notably, in the Zr₃-CN sample, the atomic percentage of Zr increased significantly, indicating that at 3 mmol of ZrCl₄, the system reached saturation, which induced the saturation of the ZMD complex precursor. This result confirms the formation of ZrO₂ and the successful doping of Zr into the CN structure through complex interactions with the DMF molecules.

The C 1s XPS spectra (Fig. 2a) and the corresponding peaks summarized in Table S2† reveal primary peaks at binding energies of 288.0 eV and 284.7 eV, corresponding to N–C=N and C–C bonds within the CN framework, respectively.^19,20,22 When Zr ions are doped into the structure, the peak shifts slightly to higher binding energies, ranging from 288.10 to 288.15 eV, for the Zr₁-CN and Zr₂-CN samples, suggesting that the CN structure interacts with the doped Zr ion species.²⁰ However, for the Zr₃-CN sample, the N–C=N peak position shifted to a lower binding energy of 287.8 eV. This peak shift can be attributed to the formation of bonds between the Zr and N atoms in the CN structure, leading to changes in the electronic environment and local potential around the carbon atoms in the N–C=N configuration, which consequently reduced the binding energy. The XPS peak for the C–C bond observed at 284.7 eV may arise from the carbon atoms in the DMF molecule that bonded with melamine and ZrCl₄ during precursor preparation for CN synthesis. Additionally, when deconvoluting the XPS peaks for Zr₃-CN, small peaks corresponding to the C=O/C–O and C–C bonds appear at 285.8 eV and 284.6 eV, respectively, indicating that the CN structure incorporated C, N, and O atoms from the complex molecular structure of ZrCl₄ and DMF during the doping process.

Fig. 2 High resolution XPS spectra and deconvoluted peaks of (a) C 1s, (b) N 1s, (c) Zr 3d, and (d) O 1s of the Zr-doped CN powder.

Fig. 2b shows the N 1s XPS peak of the synthesized CN and the Zr-doped CN. The main components observed were pyridinic N, graphitic N, pyrrolic N, and π excitation. According to the summary of the peak positions and the pyridinic-to-graphitic and pyridinic-to-pyrrolic ratios shown in Table S3,† the peak positions for pyridinic and graphitic N in the Zr₁-CN and Zr₂-CN samples shifted slightly towards higher binding energies, attributed to the formation of the Zr ion dopant and N coordination bonds, thus indicating that the Zr ions were bonded to the N atoms of CN.^20,22 However, in the case of Zr₃-CN, a shift towards a lower binding energy was observed compared to that of undoped CN. The observed peak shift may be attributed to the formation of defects within the CN structure caused by the high concentration of ZrCl₄. Additionally, the decomposition of Cl may contribute to a reduction in the crystalline integrity and bonding. DMF molecules are considered to react with melamine and the CN structure during thermal polycondensation, leading to significant changes in the atomic ratios and bonding environments of the carbon and nitrogen atoms.

Fig. 2c illustrates the Zr 3d spectrum exhibiting two peaks at 184.8 and 182.4 eV, which correspond to the 3d_3/2 and 3d_5/2 electronic states of Zr, respectively. The energy difference between the 3d_3/2 and 3d_5/2 spin–orbit doublets confirmed the existence of Zr⁴⁺ species.^19,22,38 The higher binding energy peaks for the Zr⁴⁺ species can be attributed to the interactions between Zr⁴⁺ and nitrogen.^26,39,40 The 3d_3/2 and 3d_5/2 peaks are located at higher binding energies than those of bulk ZrO₂ (184.5 and 181.5 eV),¹⁹ suggesting the absence of the ZrO₂ phase in the Zr₁-CN and Zr₂-CN samples. However, the existence of ZrO₂ was observed in the case of Zr₃-CN by XRD and SAED micrograph as shown in Fig. 3a and S4h† due to a difference in volume of detection of the techniques. In the case of Zr₃-CN, XPS peaks of Zr 3d were not shifted to lower binding energy but the Zr 3d peaks were broadened, indicating increasing Zr ion dopant changes in electronic and atomic environments of Zr⁴⁺ ion dopants and crystallinity of the Zr-doped CN powder.

Fig. 3 Crystal structure and chemical bonding information of Zr-doped CN obtained by using (a) XRD patterns and (b) FT-IR spectra. Optical properties were measured by using (c) diffusion reflectance spectra and the Tauc plot to determine the optical band gap at the x-intercept and (d) PL spectra of Zr-doped CN with an excitation wavelength of 375 nm.

The O 1s XPS spectrum, shown in Fig. 2d and summarized in Table S3,† reveals two significant deconvoluted peaks at 532 eV, corresponding to the surface oxygen species, and at 530.4 eV, attributed to the Zr–O species.^19,41 The O 1s peak in Zr-doped CN is broader than that in CN, indicating the presence of additional Zr–O species alongside surface hydroxyl groups and adsorbed water molecules. The peak at 532 eV can be attributed to the Zr–O species. The XPS spectra further show that as the Zr concentration increases, the ratio of the Zr–O peak to the O–H peak increases from 0.48 to 1.74, indicating that Zr incorporation into the CN structure likely occurs through the formation of a complex molecular structure, where Zr⁴⁺ ions bond with oxygen atoms from the DMF molecules. During calcination with melamine, a thermal polycondensation reaction occurred, leading to the formation of a Zr-doped CN structure.

The crystallinity and chemical bonding of Zr-doped CN were further investigated using XRD and FTIR, with their optical characteristics illustrated in Fig. 3. All the photocatalyst samples exhibited similar XRD patterns and FITR spectra, indicating that Zr doping of the ZMD complex maintained the overall CN structure. Fig. 3a shows the XRD patterns of the Zr-doped CN photocatalysts with various amounts of Zr doping. Two characteristic diffraction peaks were observed in all the samples at 13.1° and 27.5°, assigned to the (100) and (002) planes, respectively. These peaks represent 2D in-plane packing and interlayer periodic stacking of the heptazine ring systems,^22,26,42 indicating that an increased concentration of the ZrCl₄ precursor resulted in a decrease in the peak intensity and broadening of the diffraction peaks, with no significant peak shifts observed. These results suggest a substantial reduction in the number of stacked layers, interplanar spacing, and overall crystallinity of CN, likely due to the incorporation of Zr ions and defect formation. This effect is attributed to the external doping of Zr ions, which hinders crystallization during thermal polycondensation.^22,26 The observed effect can be attributed to the presence of Zr atoms in the Zr⁴⁺ oxidation state within the carbon nitride crystal lattice, suggesting that Zr is integrated into the structure via interstitial substitution. This occurs due to electrostatic attraction with the electron cloud around the nitrogen atoms with lone electron pairs, resulting in decreased spacing between the carbon nitride stacked layers.

Notably, no distinct diffraction peaks corresponding to Zr were detected in the samples Zr₁-CN and Zr₂-CN, primarily owing to the low concentration and uniform distribution of the Zr ions. In contrast, for the Zr₃-CN sample, the XRD pattern exhibits significant peak broadening, along with shoulder peaks at approximately 30.26° and 34.9°, indicating the formation of cubic and tetragonal ZrO₂ phases.^36,37 The (002) peak of CN shifted from 27.5° to 27.8°, possibly attributed to a reduction in the strength of hydrogen bonding within the intralayer framework of CN. This result suggests that selectively breaking hydrogen bonds is feasible, thereby disrupting in-plane stacking.⁴³ The presence of a metal chloride during the thermal polycondensation of melamine suppresses the formation of hydrogen bonds between the heptazine ring units, thereby reducing the intralayer periodic ordering.^43–45 A similar effect was observed in the synthesis of alkali metal ion-doped CN using alkali halides as precursors,⁴⁵ suggesting that using metal halides as precursors can decrease the interlayer separation. These results indicated that exceeding the critical concentration of ZrCl₄ for the ZMD complex suppressed the formation of CN and induced the formation of ZrO₂. Moreover, the presence of amorphous ZrO₂ may negatively affect the photocatalytic efficiency of CN. In addition, during the thermal polycondensation reaction that forms the CN structure, the decomposition of Cl atoms from the molecular complex may interact with C and N atoms, resulting in their oxidation. This process leads to a crystalline structure with a high defect density.⁴⁶

The chemical bonds and molecular structure of Zr-doped CN were further investigated using attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy, as shown in Fig. 3b. The strong characteristic peaks of CN in the 1000–1700 cm⁻¹ range at 1236, 1315, 1407, 1455, 1557, and 1634 cm⁻¹ are attributed to the skeletal vibrations of the heptazine heterocyclic ring units and stretching vibrations of the tri-s-triazine ring units of CN. The small peaks at 814 cm⁻¹ and 890 cm⁻¹ and the broad peak in the 3000–3400 cm⁻¹ range correspond to the breathing mode of the heptazine units, bending modes of N-containing heterocycles, and N–H stretching modes, respectively.^38,47–49

The characteristic peaks in the 1000–1700 cm⁻¹ range become weaker with the introduction of Zr, suggesting that Zr doping induces defects and integrates into the CN structure. The characteristic peaks within the 1000–1700 cm⁻¹ range are significantly reduced after introducing Zr, implying that the Zr doping introduces structural defects and incorporates them into the CN matrix. However, the core structure of CN remained clearly identifiable, likely because of the controlled Zr doping levels. Further spectroscopic analysis did not detect any peaks associated with Zr bonds, such as Zr–O, typically found in the low-frequency region (600–800 cm⁻¹), suggesting that the concentration of these bonds is too minimal to be detected.^49,50

UV-vis diffuse reflectance spectra (UV-vis DRS mode) were obtained to measure the absorption band edge and optical band gap width of the Zr-doped CN pellets, as shown in Fig. 3c and S7a.† To explore the causes of the absorption changes, the optical band gap of Zr-doped CN was determined using Kubelka–Munk and Tauc plots. The bandgap was estimated by extrapolating the linear region of the plot of (F(R)hν)^1/2vs. photon energy (hν), which corresponds to the indirect bandgap of tri-s-triazine-based CN.⁵¹ The optical band gaps of CN, Zr₁-CN, Zr₂-CN, and Zr₃-CN are 2.76 eV, 2.43 eV, 2.26 eV, and 2.78 eV, respectively. As the Zr doping concentration increases, the band gap narrows from 2.76 eV to 2.26 eV, then increases to 2.78 eV. The initial reduction in bandgap aligns with observations from the Zr-doped CN materials prepared using sulfur-bonded precursors, as well as with other semiconductor materials such as CeO₂ and TiO₂.^19,29,30 In contrast, the increase in the optical bandgap of Zr₃-CN to 2.78 eV is attributed to the presence of ZrO₂, which has a large optical bandgap ranging from 3 to 5.2 eV within the CN structure.^48,49,52,53

In addition to its impact on the crystal structure, chemical composition, and optical bandgap, Zr doping influences photogenerated charge separation, as demonstrated by steady photoluminescence (PL) and time-resolved fluorescence decay (TRPL) measurements. The PL peaks of the photocatalyst are related to the recombination of photoexcited electron–hole pairs when the photocatalyst is irradiated with photons.^38,49,53 The steady-state PL spectra of Zr-doped CN, shown in Fig. 3d, reveal significant fluorescence quenching in Zr-doped samples compared to undoped CN, indicating that the photogenerated electrons and holes were trapped at the defect sites introduced by Zr doping within the CN structure. In addition, the PL peaks exhibited a red shift, suggesting that Zr doping modified the band structure of CN, resulting in a reduced bandgap. These observations imply that photocatalysts with a lower steady-state PL intensity and red-shifted emission are likely to exhibit enhanced catalytic performance for Zr-doped CN.

The carrier lifetimes of Zr-doped CN, measured using TRPL, show the average decay times of 1.74, 0.66, 0.42, and 0.38 ns for Zr_x-CN with x = 0, 1, 2, and 3, respectively, as shown in Fig. S7b.† The PL and TRPL results indicated fluorescence quenching, inhibition of photogenerated electron–hole recombination, and reduced carrier lifetimes upon Zr doping, suggesting rapid recombination processes involving band tails within the CN framework at defect sites.^43,54 These band tails act as shallow trap states for charge carriers, and an increase in the number of band tails facilitates faster PL processes. Introducing Zr defects into the CN framework enhances the band-to-tail charge transfer processes, suppressing the radiative recombination of photogenerated charge carriers, thus generating abundant shallow charge-trapping states—beneficial for photocatalytic reactions because they promote efficient charge separation.^19,20

2.2 Photoelectrochemical measurement of the Zr-doped CN photocatalyst

To evaluate the photocatalytic efficiency of Zr-doped CN, photoelectrochemical measurements of Zr-doped CN as a photoanode were performed using I–V linear sweep voltammetry (LSV), electrochemical impedance spectroscopy (EIS), transient photocurrent response measurements, and Mott–Schottky analysis, as shown in Fig. 4. LSV in the dark and under light illumination was performed by applying a potential, as illustrated in Fig. 4a. None of the samples showed any significant change in current density under dark conditions with an increase in potential. Furthermore, the photocurrent density under light conditions improved significantly with the applied potential, indicating that the photogenerated charge carriers changed the separation of the photocatalysts.

Fig. 4 Photoelectrochemical properties of Zr-doped CN on FTO in Na₂SO₄ electrolyte. (a) Scanning of current density versus potential in the range of 0.5 to 1.1 V under dark and light conditions, (b) EIS profile at a bias potential of 0.8 V under dark and light conditions, (c) transient photocurrent response at a bias potential of 0.8 V, and (d) Mott–Schottky plot in the presence of light at a fixed frequency of 1000 Hz.

The photogenerated charge-separation characteristics of the Zr-doped CN photoelectrodes were examined using EIS. The EIS profiles of the photoelectrodes were plotted as complex impedance Nyquist plots, as shown in Fig. 4b. The Nyquist plots of the photoelectrode consist of a semicircular arc, which indicates the charge transfer resistance (R_ct) between the photoelectrode and electrolyte; a lower arc diameter results in a lower R_ct and better charge transfer. Herein, the R_ct values of the photoelectrodes followed the order of Zr₂-CN < Zr₁-CN < Zr₃-CN < CN, implying that the lower charge transfer resistance of Zr-doped CN can be attributed to the higher charge carrier generation, charge separation, and diffusion to the interface of the electrolyte and electrode. The improvement in the charge-transfer ability of Zr-doped CN indicates that the band and electronic structures of CN were modified by Zr doping.

Fig. 4c presents the transient photocurrent density curves of the Zr-doped CN photocatalysts on the FTO substrate. The curves show an apparent increase in the photocurrent when exposed to visible light, followed by a return to the original state when the light was turned off, indicating that the electrons generated by light absorption create a photocurrent; this result suggests that the photocatalysts are active only when illuminated. The results indicate that the photocurrent densities follow the order of Zr₂-CN > Zr₁-CN > Zr₃-CN > CN. Additionally, the photocurrent for all the photocatalysts remained nearly constant throughout the entire cycle, demonstrating the good photostability of the catalysts under light. These results suggest that using an optimal amount of ZrCl₄ in the precursor leads to higher photocurrent density.

To clearly understand the band structure of Zr-doped CN, Mott Schottky analysis under light conditions was performed to evaluate the flat-band potential of the Zr-doped CN photoelectrode using the Mott Schottky function. As shown in Fig. 4d, all the samples exhibit a positive slope in the plot of 1/C²vs. the applied potential (V vs. the saturated calomel electrode (SCE)), indicating that all the samples exhibit characteristics of n-type semiconductors. The flat band potentials of the photocatalysts CN, Zr₁-CN, Zr₂-CN, and Zr₃-CN, determined from the x-intercepts, were −1.13 V, −1.08 V, −0.82 V, and −0.97 V (vs. SCE), respectively. Based on the slopes and assuming that the dielectric constant remains approximately constant, photocatalysts with lower slopes exhibit higher charge carrier densities, indicating that Zr₂-CN possesses the highest charge carrier density relative to the other Zr-doped samples and undoped CN. Owing to the n-type semiconductor properties, the flat band potential can be considered as the conduction band edge because the flat band level closely aligns with the conduction band.^55,56 Therefore, combining the optical band gap derived from diffuse reflectance spectroscopy (DRS) with the conduction band edge determined from the Mott Schottky analysis allows for the construction of the band structure diagram of Zr-doped CN relative to the normal hydrogen electrode (NHE), as illustrated in Fig. 5. For comparison with NHE, the potential measured against the SCE in this study was converted to NHE at pH 7 using the equation E_(NHE) = E_(SCE) + 0.24 V.

Fig. 5 The band structure diagram for Zr-doped CN was evaluated based on the flat band potential (from Mott–Schottky plots) and band gap (from DRS UV-vis spectra and the Tauc plot).

Based on the displayed band structure diagram, as the Zr doping concentration increased from 0 to 2 mmol, the band gap narrowed, while both the conduction band edge and calculated valence band edge shifted downward, approaching 0 V vs. NHE. However, ZrCl₄ precursor doping exceeding 2 mmol results in the formation of ZrO₂, degrading the crystallinity of CN. This process could be attributed to the formation of ZrO₂ and interaction of Cl during the thermal polycondensation process, which adversely affected the conduction and calculated valence band energy levels, causing them to shift to higher energy values. These results imply that photocatalysts with lower bandgaps and energy levels closer to the generation potential of radical species exhibit enhanced catalytic efficiencies for radical formation. Consequently, maintaining an optimal Zr doping concentration is crucial for achieving high photocatalytic performance.

2.3 Photocatalytic efficiency of Zr-doped CN for degradation of rhodamine B molecules

The photocatalytic activity of Zr-doped CN under irradiation by a commercial white LED lamp was examined using rhodamine B (RhB) dye molecules as the target pollutant, as shown in Fig. 6. The kinetic degradation curves of RhB using the Zr-doped CN photocatalyst were plotted as C/C₀vs. time, where C/C₀ is the ratio of the concentration at a certain time (t) to the initial concentration, which was measured by using the ratio of the absorbance at the highest intensity peak of the RhB solution using UV-vis spectroscopy, as shown in Fig. 6a and S9a.† The photocatalytic RhB degradation efficiency of Zr-doped CN was examined by fitting a kinetic curve to a pseudo-first-order reaction. The pseudo-first-order reaction is represented by the following function: −ln(C/C₀) = Kt, where K is the kinetic first-order rate constant and t is the irradiation time. As shown in Fig. 6b, the concentration of RhB decreased faster in the presence of Zr-doped CN than in the absence of Zr or the photocatalyst. The absorption–desorption before light irradiation, RhB degradation efficiency, and kinetic first-order rate constant after 540 min under light irradiation of Zr-doped CN are summarized in Table S4.†

Fig. 6 Photocatalytic degradation of 10 ppm RhB with the Zr-doped CN photocatalyst. (a) Time-dependent degradation kinetic curve (C/C₀–time). (b) Pseudo first order kinetic plot with time and a linear fitting line to determine the rate constant. (c) Photocatalytic degradation of RhB of Zr₂-CN in the presence of various scavengers. (d) The plot of the rate constant versus concentration of additive EDTA in mM of Zr-doped CN.

According to the pseudo-first-order kinetic plots and rate constants, increasing the Zr doping concentration from 0 to 2 mmol significantly enhanced the photocatalytic performance of the catalyst, with the degradation of RhB improving from 68% to 98% over 540 min. The rate constant (K) increased from 1.6 × 10⁻³ to 7.2 × 10⁻³ s⁻¹, indicating a 4.5-fold enhancement in the reaction rate compared to that of undoped CN; this improvement can be attributed to the reduced optical bandgap, PL quenching, and modification of the energy levels. The proposed photocatalytic mechanism elucidates the enhanced activity of Zr-doped CN, with the bandgap of CN reduced from 2.76 eV to 2.26 eV, thereby facilitating the absorption of visible light with improved efficiency. In this context, electrons in the valence band of Zr-doped CN can be readily excited to the conduction band, generating many electron–hole pairs. Moreover, at an optimal doping concentration, the incorporated Zr-ion doping enhanced the electronic conductivity, which markedly suppressed the recombination of photogenerated electron–hole pairs, thereby improving the photocatalytic efficiency. Additionally, the mesoporous structure and large specific surface area of Zr-doped CN promoted the efficient diffusion of photocarriers into the bulk surface, subsequently enabling rapid intraparticle molecular transfer rates. This structural advantage enhanced the photocatalytic performance of the Zr-doped CN photocatalyst. However, excess doping with Zr at concentrations of up to 3 mmol led to the formation of a ZrO₂ phase and a reduction in the crystallinity of CN, which acted as a recombination center and reduced the photocatalytic activity of Zr-doped CN. These changes also affect the photocatalytic properties by increasing the optical band gap, which reduces the ability of the material to absorb light in the visible spectrum, thus lowering the photocatalytic efficiency.^20,26

For practical applications, five rounds of recycling photocatalytic degradation of RhB were performed by using the photocatalyst Zr₂-CN to investigate its stability, as shown in Fig. S9b–d.† The degradation of RhB for each round of the recycling reaction approached 95–98% in 540 min, and the crystallinity and chemical bonding properties of the Zr₂-CN powder underwent only slight changes after the photocatalytic reaction, indicating that Zr₂-CN had sufficient stability and photocatalytic efficiency for practical dye degradation.

Radical scavenging tests are essential for understanding photocatalytic degradation, where photo-induced holes (h⁺), hydroxyl radicals (˙OH), and superoxide radicals (˙O₂⁻) play a crucial role in breaking down organic pollutants. Therefore, identifying the most reactive species responsible for the degradation of the dye molecules is crucial. The roles of the responsive and reactive species in the photocatalytic degradation of RhB using Zr₂-CN were investigated by adding various scavengers (Fig. 6c). The degradation of RhB was significantly suppressed in the presence of benzoquinone (BQ) and barely suppressed by the addition of isopropanol (IPA) and dimethyl sulfoxide (DMSO), which implied that superoxide radicals were the dominant reactive species.^6,57 Notably, the RhB degradation efficiency was significantly enhanced (up to 4.5-fold) with ethylenediaminetetraacetic acid (EDTA) compared to that without EDTA.

Further experiments were performed to investigate the effect of EDTA addition on the photocatalytic degradation of RhB (Fig. 6d). Zr-doped CN exhibited a significant increase in the photocatalytic activity for RhB degradation with the addition of EDTA, particularly at an optimal concentration of 1 mM. Under these conditions, the Zr₂-CN catalyst reached the highest photocatalytic efficiency. In comparison, undoped CN showed only a slight improvement in the photocatalytic performance when EDTA was added. The experimental results can be explained as follows: under visible-light irradiation, electrons in the valence band transition to the conduction band, resulting in the formation of electron–hole pairs. Excited electrons can recombine with holes through direct surface recombination, thereby suppressing the efficiency of the photocatalytic reactions. Adding EDTA as a scavenger significantly inhibited the recombination of electrons and holes by capturing the holes. Consequently, more electrons in the conduction band migrate to the surface of the photocatalyst, where they react with O₂ to generate oxygen radical species, which enhance the degradation of dyes by Zr-doped CN under visible light irradiation.^17,58

3 Conclusion

We successfully synthesized the Zr-doped CN photocatalyst using a ZMD complex. By optimizing the concentration of ZrCl₄ doping, the photocatalytic activity for generating oxygen radicals from the photocatalytic water splitting reaction was significantly enhanced, leading to a 4.5-fold increase in the degradation efficiency of RhB compared to that of undoped CN. The optimized doping level enhances the surface area-to-volume ratio resulting from hierarchical porous structure formation while achieving a suitable energy bandgap and optimal alignment of energy bands for photocatalytic processes, along with a higher concentration of photoinduced charge carriers. This study suggests that the photocatalytic properties of Zr-doped CN can be further optimized by combining it with other materials to form heterojunction photocatalysts. Moreover, this approach can be used to create Z-scheme or S-scheme photocatalytic structures by leveraging the tunable energy band levels achieved through Zr doping.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and ESI.†

Author contributions

Navaphun Kayunkid and Visittapong Yordsri: method, characterization and discussion of the results. Winadda Wongwiriyapan and Young Jae Song: supervised the project. All authors participated in the manuscript review.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work has been supported by the National Research Foundation of Korea (NRF) through a grant funded by the Korean government (MSIT) (Grant No. NRF-2021R1A2C2007141) and by the Institute for Basic Science (IBS-R036-D1). The authors are also thankful to the Nanotechnology and Material Analytical Instrument Service Unit (NMIS), College of Materials Innovation and Technology (CMIT), King Mongkut's Institute of Technology Ladkrabang (KMITL), for their valuable assistance with material characterization.

References

1. L. H. Gao, W. Y. Xiao, M. Y. Qi, J. Y. Li, C. L. Tan and Z. R. Tang, Mol. Catal., 2024, 113858,  DOI:10.1016/j.mcat.2024.113858.
2. X. Yan, J. H. Dong, J. Y. Zheng, Y. Wu and F. X. Xiao, Chem. Sci., 2024, 15, 2898–2913.
3. R. Ghamarpoor, A. Fallah and M. Jamshidi, ACS Omega, 2024, 9(24), 25457–25492.
4. Y. Sari, P. L. Gareso, B. Armynah and D. Tahir, Int. J. Hydrogen Energy, 2024, 55, 984–996.
5. A. R. Bhapkar and S. Bhame, J. Environ. Chem. Eng., 2024, 12(3), 112553.
6. A. Sudhaik, P. Raizada, P. Shandilya, D. Y. Jeong, J. H. Lim and P. Singh, J. Ind. Eng. Chem., 2018, 67, 28–51.
7. M. A. Ahmed, S. A. Mahmoud and A. A. Mohamed, RSC Adv., 2024, 14(35), 25629–25662.
8. Y. Wang, X. Wang and M. Antonietti, Angew. Chem., Int. Ed., 2012, 51(1), 68–89.
9. P. P. Singh and V. Srivastava, RSC Adv., 2022, 12, 18245–18265.
10. M. Tahir, N. Mahmood, J. Zhu, A. Mahmood, F. K. Butt, S. Rizwan, I. Aslam, M. Tanveer, F. Idrees, I. Shakir, C. Cao and Y. Hou, Sci. Rep., 2015, 12389,  DOI:10.1038/srep12389.
11. C. Zhou, R. Shi, L. Shang, L. Z. Wu, C. H. Tung and T. Zhang, Nano Res., 2018, 11, 3462–3468.
12. J. Fang, H. Fan, M. Li and C. Long, J. Mater. Chem. A, 2015, 3, 13819–13826.
13. Y. Wang, X. Tang, P. Huo, Y. Yan, Z. Zhu, J. Dai, Z. Liu, Z. Li and H. Xi, J. Phys. Chem. C, 2021, 125, 9646–9656.
14. E. X. Han, Y. Y. Li, Q. H. Wang, W. Q. Huang, L. Luo, W. Hu and G. F. Huang, J. Mater. Sci. Technol., 2019, 35, 2288–2296.
15. D. Zhao, Y. Wang, C. L. Dong, Y. C. Huang, J. Chen, F. Xue, S. Shen and L. Guo, Nat. Energy, 2021, 6, 388–397.
16. S. Guo, Z. Deng, M. Li, B. Jiang, C. Tian, Q. Pan and H. Fu, Angew. Chem., Int. Ed., 2016, 55, 1830–1834.
17. S. Hu, F. Li, Z. Fan, F. Wang, Y. Zhao and Z. Lv, Dalton Trans., 2014, 44, 1084–1092.
18. L. Zhang, N. Ding, M. Hashimoto, K. Iwasaki, N. Chikamori, K. Nakata, Y. Xu, J. Shi, H. Wu, Y. Luo, D. Li, A. Fujishima and Q. Meng, Nano Res., 2018, 11, 2295–2309.
19. A. Kumar, S. Samanta and R. Srivastava, ACS Appl. Nano Mater., 2021, 4, 6805–6820.
20. X. Guo, J. Duan, C. Li, Z. Zhang and W. Wang, Colloids Surf., A, 2022, 649, 129474.
21. T. Ma, Q. Shen, B. Z. J. Xue, R. Guan, X. Liu, H. Jia and B. Xu, Inorg. Chem. Commun., 2019, 107, 107451.
22. S. Zhu, X. Cao, X. Cao, Y. Feng, X. Lin, K. Han, X. Li and P. Deng, Mater. Des., 2021, 199, 109426.
23. Y. Rangraz, F. Nemati and A. Elhampour, Appl. Surf. Sci., 2020, 507, 145164.
24. I. N. Reddy, L. V. Reddy, N. Jayashree, C. V. Reddy, M. Cho, D. Kim and J. Shim, Chemosphere, 2021, 264, 128593.
25. J. C. Wang, C. X. Cui, Q. Q. Kong, C. Y. Ren, Z. Li, L. Qu, Y. Zhang and K. Jiang, ACS Sustain. Chem. Eng., 2018, 6, 8754–8761.
26. Y. Wang, Y. Wang, Y. Li, H. Shi, Y. Xu, H. Qin, X. Li, Y. Zuo, S. Kang and L. Cui, Catal. Commun., 2015, 72, 24–28.
27. X. Chen, J. Zhang, X. Fu, M. Antonietti and X. Wang, J. Am. Chem. Soc., 2009, 131, 11658–11659.
28. S. Tonda, S. Kumar, S. Kandula and V. Shanker, J. Mater. Chem. A, 2014, 2, 6772–6780.
29. J. Lukáč, M. Klementová, P. Bezdička, S. Bakardjieva, J. Šubrt, L. Szatmáry, Z. Bastl and J. Jirkovský, Appl. Catal., B, 2007, 74, 83–91.
30. X. Liu, J. Ding, X. Lin, R. Gao, Z. Li and W. L. Dai, Appl. Catal., A, 2015, 503, 117–123.
31. S. M. Chang and R. A. Doong, J. Phys. Chem. B, 2006, 110, 20808–20814.
32. X. Liu, W. Liu, X. Zhang, L. Han, C. Zhang and Y. Yang, Cryst. Res. Technol., 2014, 49, 383–392.
33. X. Zhang, Y. Yang, W. Huang, Y. Yang, Y. Wang, C. He, N. Liu, M. Wu and L. Tang, Mater. Res. Bull., 2018, 99, 349–358.
34. J. H. Zhang, M. J. Wei, Z. W. Wei, M. Pan and C. Y. Su, ACS Appl. Nano Mater., 2020, 3(2), 1010–1018.
35. M. Taddei, J. A. van Bokhoven and M. Ranocchiari, The influence of water in the synthesis of Zr-based metal-organic framework UiO-66: isolation and reactivity of the new species [ZrCl(OH)2(DMF)2]Cl, ChemRxiv, 2020, preprint,  DOI:10.26434/chemrxiv.11382165.v2.
36. S. Jayakumar, P. V. Ananthapadmanabhan, T. K. Thiyagarajan, K. Perumal, S. C. Mishra, G. Suresh, L. T. Su and A. I. Y. Tok, Mater. Chem. Phys., 2013, 140, 176–182.
37. F. Shi, Y. Li, H. Wang and Q. Zhang, Prog. Nat. Sci.:Mater. Int., 2012, 22, 15–20.
38. K. Zhang, M. Zhou, C. Yu, K. Yang, X. Li, W. Dai, J. Guan, Q. Shu and W. Huang, Dyes Pigm., 2020, 108525,  DOI:10.1016/j.dyepig.2020.108525.
39. S. Samanta, R. Yadav, A. Kumar, A. Kumar Sinha and R. Srivastava, Appl. Catal., B, 2019, 118054,  DOI:10.1016/j.apcatb.2019.118054.
40. T. Zhang, Y. Lu, W. Li, M. Su, T. Yang, A. Ogunbiyi and Y. Jin, Int. J. Hydrogen Energy, 2019, 44, 14527–14535.
41. X. Zhang, Y. Yang, W. Huang, Y. Yang, Y. Wang, C. He, N. Liu, M. Wu and L. Tang, Mater. Res. Bull., 2018, 99, 349–358.
42. Q. Wu, T. Jeong, S. H. Kim and Y. J. Song, J. Alloys Compd., 2022, 163310,  DOI:10.1016/j.jallcom.2021.163310.
43. Y. Kang, Y. Yang, L. C. Yin, X. Kang, L. Wang, G. Liu and H. M. Cheng, Adv. Mater., 2016, 28, 6471–6477.
44. Y. Kang, Y. Yang, L. C. Yin, X. Kang, G. Liu and H. M. Cheng, Adv. Mater., 2015, 27, 4572–4577.
45. Y. Xu, Y. Gong, H. Ren, W. Liu, L. Niu, C. Li and X. Liu, RSC Adv., 2017, 7, 32592–32600.
46. L. Liang, Y. Cong, L. Yao, F. Wang and L. Shi, Mater. Res. Express, 2018, 115510,  DOI:10.1088/2053-1591/aade38.
47. T. Narkbuakaew and P. Sujaridworakun, Top. Catal., 2020, 63, 1086–1096.
48. R. Wang, Y. Hu, J. Du, L. Xu and Y. Fu, J. Mater. Res., 2021, 36, 3086–3095.
49. M. Zarei, J. Bahrami and M. Zarei, Appl. Water Sci., 2019, 175,  DOI:10.1007/s13201-019-1076-8.
50. X. Bi, S. Yu, E. Liu, X. Yin, Y. Zhao and W. Xiong, RSC Adv., 2019, 10, 524–532.
51. G. S. Priyanga, G. Pransu and S. Sampath, Chem. Phys. Impact, 2024, 8, 100408.
52. N. Alebachew, H. C. A. Murthy, B. A. Gonfa, K. G. von Eschwege, E. H. G. Langner, E. Coetsee and T. B. Demissie, ACS Omega, 2023, 8, 13775–13790.
53. X. Bi, S. Yu, E. Liu, X. Yin, Y. Zhao and W. Xiong, RSC Adv., 2019, 10, 524–532.
54. H. Katsumata, F. Higashi, Y. Kobayashi, I. Tateishi, M. Furukawa and S. Kaneco, Sci. Rep., 2019, 9(1), 14873.
55. S. Samanta, B. Satpati and R. Srivastava, Nanoscale Adv., 2019, 1, 1395–1412.
56. J. Fu, B. Zhu, C. Jiang, B. Cheng, W. You and J. Yu, Small, 2017, 13, 1603938.
57. S. Ganesan, T. Kokulnathan, S. Sumathi and A. Palaniappan, Sci. Rep., 2024, 14(1), 2284.
58. T. Liu, L. Wang, X. Lu, J. Fan, X. Cai, B. Gao, R. Miao, J. Wang and Y. Lv, RSC Adv., 2017, 7, 12292–12300.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta09247c

‡ Chinathun Pinming and Qingshan Yang contributed equally to this work.

---