Potassium-doped g-C3N4 enables efficient visible-light-driven dye degradation

Element doping is recognized as an efficient method to boost the photocatalytic performance of photocatalysts. Here, a new potassium ion-doped precursor, potassium sorbate, was employed in melamine configuration during calcination process to prepare the potassium-doped g-C3N4 (KCN). By various characterization techniques and electrochemical measurements, the doping of K in g-C3N4 can efficiently modify the band structure to enhance the light absorption and greatly increase the conductivity to accelerate charge transfer and photogenerated carrier separation, ultimately achieving an excellent photodegradation of the organic pollutant (methylene blue, MB). These results have demonstrated that the approach of potassium incorporation in g-C3N4 has potential in fabricating high-performance photocatalysts for organic pollutant removal.


Introduction
In the past few decades, the fast development of industry and extensive use of fossil fuels have caused serious harm to our environment, especially the irreversible water pollution by various organic compounds, inorganic toxic, and heavy metal pollutants (McGlade and Ekins 2015;Schrope 2011). Photocatalytic technique has been recognized as an effective environmental detoxification approach, since it can convert the non-biodegradable and toxic materials into CO 2 , H 2 O, and other harmless salts (Djellabi et al. 2021;Li and Li 2002;Zhang et al. 2018). Hence, seeking and developing high-performance photocatalysts are of significant importance.
Over the past few decades, various semiconductors have been reported and utilized for water purification, such as the metal-based photocatalysts TiO 2 (Ye et al. 2017), CdS (Lin et al. 2021), and ZnO (Rokhsat and Akhavan 2016). Nevertheless, these materials always present wide band gap, which greatly limits their practical application (Wang et al. 2010). Other photocatalysts like Cu 2−x Se (Liu et al. 2019) and ZnIn 2 S 4 (Mao et al. 2021) also have been widely researched, whereas their slow photogenerate charge-hole separation rate still hinders their further development. Graphitic carbon nitride (g-C 3 N 4 ), a metal-free photocatalyst, has been intensively studied in water splitting (Cheng et al. 2018) and water purification (Akulinkin et al. 2021;Groenewolt and Antonietti 2005;Yan et al. 2009). g-C 3 N 4 is composed of sp 2 -hybridized C and N forms a π-conjugated system, favoring the adjustment of nanosheet-based materials in electronics and optoelectronics (Cui et al. 2011;Yang et al. 2015). Additionally, g-C 3 N 4 presents suitable energy gap of 2.7 eV, indicating it can absorb and utilize visible light with wavelengths < 470 nm (Li et al. 2019). To further expand the visible-light absorption range of g-C 3 N 4 , and  (Xu et al 2022;Yan et al. 2018), doping non-metallic elements (Bellardita et al. 2018), introducing defects or vacancy (Cao et al. 2019a), and building composites or heterojunctions (Ali et al. 2022;Zhao et al. 2018). Among these strategies, the element doping can change the band structure, expand the light absorption range, and increase the conductivity to boost the charge transfer, thus boosting carriers' separation and increasing the reaction kinetics (An and Onishi 2015;Wang et al. 2020).
Herein, a potassium-doped g-C 3 N 4 (KCN) has been prepared via calcining a mixture of melamine and potassium sorbate, and the as-prepared catalyst has been employed to degrade the methylene blue (MB) under the visible light irradiation. The effects of K doping on phase structure, chemical composition, and band gap are identified via various characterizations. Furthermore, the degradation performance and main reactive radical for degradation have been investigated.

Chemicals
Melamine and potassium sorbate were purchased from Aldrich Reagent Co., Ltd. (Shanghai, China).

Material preparation
Typically, 2 g of melamine and 0.25 g potassium sorbate were uniformly mixed in a mortar for half hour, and the mixture was transferred in muffle furnace to 550 °C (5 °C min −1 ) and kept for 3 h. Finally, the K-doped g-C 3 N 4 was prepared. The g-C 3 N 4 (CN) was fabricated by directly heating melamine without potassium sorbate.

Photocatalytic degradation experiments
The photocatalytic performance of CN and KCN was investigated via the MB degradation under visible light (λ ≥ 400 nm). Generally, 30 mg sample was dispersed in 50 mL 10 mg L −1 MB solution. After adsorption for 30 min, a 300-W Xenon lamp with 400 nm cutoff filter was employed as the light source, and 1 mL solution was collected at 5-min interval. The MB concentration was detected via a UV-vis absorption spectra (Thermo Scientific, USA) after centrifugation.

Results and discussion
The structure information of CN and KCN was identified via the XRD, as shown in Fig. 1a; all the diffraction signals of CN and KCN are well matched with the standard pattern of JCPDS87-1526 (Wang et al. 2020), and peaks located at 12.9 and 27.8° are ascribed to the (100) and (002) planes, respectively. Notably, the characteristic peak of the (002) plane is slightly shifted to lower angle, indicating that the interlayer distance of KCN is enlarged and demonstrating that the small-sized K atoms have entered into the CN structure (Fig. 1b). Furthermore, the weakened (002) peak intensity can be corresponded to the incomplete polymerization induced by the K incorporation during the condensation process. In addition, the chemical structure information of CN and KCN was analyzed by FT-IR, as displayed in Fig. 1c; the characteristic signals at 1200-1650 and 815 cm −1 are related to the aromatic heterocyclic ring stretching and the heterocyclic ring condensation of triazine unit in g-C 3 N 4 (Li et al. 2007;Zhou et al. 2016). The apparent peaks of -OH and uncondensed ammonium groups (NH) are observed at 3000-4000 cm -1 (Bai et al. 2013). Impressively, there is no K-related vibration found in the FT-IR spectra, indicating that the introduction of K atoms in CN will not affect the internal chemical deposition.
The XPS spectra of CN and KCN are further fitted to understand the chemical valence states and element composition. As depicted in Fig. 1d, a pair of apparent K 2p peaks can be fitted in K 2p XPS single, whereas there is no K 2p single in CN sample, indicating the successful introduction of K. Besides, the C 1 s fine spectra of CN and KCN (Fig. 1e) can be divided into two obvious peaks, and the peak located in 284.5 eV and 284.5 eV are ascribed to the sp 2 hybridization of C-C and sp 2 hybridization of the carbon (N-C = N) (Li et al. 2014). In the Fig. 1f, three peaks located at 400.8, 399.0, and 398.5 eV in N 1 s spectrum can be attributed to N-H of the uncondensed amino groups, tertiary nitrogen in N-(C) 3 group, and C-N = C group (Gu et al. 2015;Lin and Wang 2013). Note that the intensity ratio of sp 2 N (C-N = C)/sp 3 N (N-(C) 3 ) in KCN is 3.95, which is much lower than that in CN (5.79), implying that the introduction of K atoms in KCN will cause more structural distortion and defects, as sp 2 N becomes sp 3 N when the C-N plane is distorted (Wang et al. 2020). As presented by SEM and energy dispersive spectroscopy (EDS) images, the KCN demonstrates the typical morphology of g-C 3 N 4 composed of highly condensed sheets (Fig. 2a). The EDS mapping demonstrates that the C, N, and K elements are uniformly distributed on the surface of KCN (Fig. 2b-d), further confirming the K atoms are successfully inserted into the g-C 3 N 4 .
The optical properties of CN and KCN have been investigated by DRS, PL, and photogenerate current measurements. As shown in Fig. 3a, CN sample presents a visible light absorption and the edge is located at 478 nm, whereas the KCN shows an apparent shift to larger absorption wavelength at 492 nm, suggesting that the electronic structure and band gap are significantly modified in KCN. Further, the energy band gap of CN and KCN is calculated via the Tauc/David-Mott model to be 2.8 and 2.7 eV, respectively (Ding et al. 2018), demonstrating that the photocatalytic performance of KCN can be greatly improved via increasing the visible light absorption. Besides, the separation rate of photogenerated electrons/holes in CN and KCN is studied by PL measurement, as depicted in Fig. 3b; the peak at 450 nm can be related to the direct recombination of holes and electrons. The CN sample displays the lower intensity when comparing with the KCN, indicating the K doping can efficiently depress the recombination of holes and electrons. In addition, the charge separation efficiency can be further confirmed via the photoelectrochemical measurements. As depicted in the photogenerate current Fig. 1 a XRD pattern of CN  and KCN. b Schematic illustration of KCN. c FT-IR, d K 2p, e C 2 s, and f N 1 s spectra of CN and KCN response single (conduct at the open circuit potential), it can be found that the photocurrent density of KCN is much higher than that of pure CN and almost two times improved (Fig. 3c). The EIS plots of CN and KCN are recorded in Fig. 3d. It is obvious to see that the arc radius of KCN is smaller than that of CN. These results are consistent with photocurrent response and PL results, all implying the K doping in KCN can enhance the visible light absorption range and accelerate the photogenerated carriers transfer, thus improving the photocatalytic activity. To verify the photocatalytic activity of KCN and CN, the degrading methylene blue dye (MB) efficiency under the visible light were compared. To achieve adsorption-desorption equilibration, the configurations with catalysts and pollutants were stirred for 30 min in the dark to achieve adsorption-desorption equilibration (Fig. S1). As shown in Fig. 4a, the utilization of CN brought 61% removal in 60 min, and the MB can be directly photolyzed by 20% within 60 min. In a sharp contrast, the photodegradation of MB was greatly improved in KCN and achieved a superior removal efficiency of 99% after 60 min, which can be ascribed to the increased specific surface area after K doping (Fig. S2). Furthermore, the photodegradation results were fitted with the fist-order kinetic in Fig. 4b; the rate constant of MB degradation in KCN was about 3.6 times that of CN, and the degradation rate constants normalized by the surface area of KCN also show a higher value (Table S1). The comparison of the activity of KCN with other CN-based materials has been listed in Table. S2. To further investigated the positive effect of K incorporation in CN, the KCN with different K content were also conducted the photocatalytic activity. As displayed in Fig. S3, the KCN with 0.1 g potassium sorbate precursor also shows a little enhancement in degradation, whereas the excessive K doping leads to the degradation of catalytic performance in KCN-0.5, which is ascribed to the excessive doping that will cause the lattice collapse and affect their performance. After the degradation, the used catalyst was collected and washed and then reused to check the cycling stability of KCN. In Fig. 4c, the MB removal by KCN could still be maintained above 95% after 5 cycles, indicating an excellent stability of KCN photocatalyst. In addition, to reveal photocatalytic mechanism and reactive species in the degradation system, the ESR spectrums in different configurations were conducted. In ESR spectrum of Fig. 4d, the O 2 •− intensity in KCN was much stronger than that in CN, which means that there was more O 2 •− generated during the light irradiation, which can be related to the improved conductivity and enhanced photogenerated carrier separation rate. Therefore, more electrons will react with the dissolved O 2 to form more O 2 •− in KCN system (Cao et al. 2019b), thus accelerating the degradation of MB.

Conclusion
In summary, to increase the visible light utilization and the separation of photogenerated carriers in the g-C 3 N 4 , potassium incorporating g-C 3 N 4 (KCN) is proposed to promote effective MB degradation. KCN was fabricated by a facile thermal condensation, and the K doping in g-C 3 N 4 was identified via XRD, XPS, and EDS mapping. The characterizations and electrochemical measurements revealed that the K incorporation in g-C 3 N 4 could not only alter the band structure to enhance the visible light absorption but also improve the conductivity to accelerate the separation of photogenerated electrons/ holes, thus achieving a superior MB degradation efficiency, and the degradation kinetics reached 0.119 min −1 . This work presented a new potassium ion-doped precursor to synthesize high-performance photocatalyst, thus offing a new pathway for designing and developing high-efficiency photocatalysts.