Increasing water pollution is one of the greatest challenges to the environment. Industries based on textiles, leather, food and cosmetics use large amount of dyes, and the organic pollutants from these industries is a major environmental anxiety. Methyl orange (MO) is one of the main sources of water pollution, which is an intensely colored compound used in dyeing and printing textiles. Therefore, seeking efficient, greenand safe purification technologies and materials which can effectively remove organic pollutants from wastewater are of great significance to environmental protection. Semiconductor materials have great potential in environmental and energetic applications, because they can eject an electron from the valence band to the conduction band under irradiation at the proper wavelength, where it is mobile and can be used for photocatalytic degradation reactions^[1-3].

Since the early report of graphitic carbon nitride (g-C₃N₄) photocatalysts in the H₂ evolution in 2009, the metal-free conjugated polymeric semiconductor has gained considerable interdisciplinary interest^[4]. Owing to its visible light response, high chemical and thermal stability, a relatively narrow band gap of 2.7 eV, reasonable cost, "earth-abundant" nature, flexible super molecular networks and environmentally benign characteristics, g-C₃N₄ has been considered as the next ge- neration photocatalyst in the research communities^[5-8]. However, the intrinsic drawbacks of graphitic carbon nitride, especially the high recombination of photogenerated charge carriers, the limited active sites, low specific surface area and the serious aggregation during the photocatalytic process, seriously restrict the cataly- tic vigor^[9-10]. Thus, to promote the separation of photogenerated electrons (e^-) and holes (h⁺) pairs would be a key factor for the photocatalytic degradation of organic contaminants^[11]. Many studies have been devo- ted to improve the separation rate of charge carriers by modified g-C₃N₄ using metal and non-metal heteroatom^[12-13], fabricating heterojunctions with other semiconductors^[14], and nanostructure modulations^[15-16], etc. However, most of these methods are complicated, time consuming, and unfriendly to the environment. Therefore, an easy way to increase the activity of carbon nitride is urgently needed. Liu et al. reported that persulfate (PS) captured photogenerated e^- and then reduced O₂ into ·O₂^-, thus boosted the separation of photogenerated e^-/ h⁺ pairs and led to the increase of the g-C₃N₄ photocayalytic performance^[17]. So, it is a simple and effective way to improve the photocatalytic performance of g-C₃N₄ by adding an e^- or h⁺ trapping agent to promote the separation of photogenerated e^-/ h⁺ pairs.

We all know that EDTA-2Na is commonly used as an h⁺-trapping agent in capture experiments. However, an interesting experimental phenomenon that the photocatalytic degradation rate was remarkably increased in the presence of EDTA-2Na has been observed. This phenomenon has also been found in other literatures^[18-21]. What's more, we found that EDTA promoted the photocatalytic degradation of MO by g-C₃N₄ at a higher rate than EDTA-2Na. A possible reaction mechanism is proposed by the capture experiments. These results indicated that acidic conditions favor the degradation of MO, h⁺ is trapped and the separation of photogenerated e^-/h⁺ pairs is promoted when EDTA is added. So, the photocatalytic activity of g-C₃N₄ remar- kably improved. This work could inspire an idea of adding some kind of agents which could react with e^- or h⁺ to reduce the recombination of photo-generated carriers, then, achieved a highly efficient photocataly- tic reaction.

1 Materials and methods 1.1 Synthesis of g-C₃N₄

The g-C₃N₄ sample was prepared by a previously reported thermal polymerization method^[22]. The g-C₃N₄ powder was synthesized typically by heating 5 g of melamine in a semi-closed quartz crucible with a cover in a nitrogen atmosphere. The heating program was as follows: heating to 550 ℃ at the rate of 5 ℃/min and holding for 4 h, then naturally cooling to room temperature. The product was collected and ground into powder.

1.2 Catalyst characterization

Fourier transform infrared (FTIR) spectroscopy was gained by a Themo Scientific Nicolet-6700 FTIR spectrophotometer. X-ray diffraction (XRD) patterns were recorded by a Rigaku Dmax/Ultima Ⅳ diffractometer with monochrommatized Cu kα radiation (k = 1.5418 Å). Scanning electron microscopy (SEM) was taken with a JEOL JSM-6510LV scanning electron microscopy. And UV-vis diffuse reflectance spectra (DRS) was recorded on a Shimazu UV-3600 spectrophotometer equipped with diffuse reflectance accessories, using BaSO₄ as the reference sample. X-Ray photoelectron spectroscopy (XPS) measurements were carried outusing an ESCALAB 250 Xi with a high-performance Al monochromatic source (hv = 1486.6 eV, 150 W). All binding energies were referenced to the C 1s peak at 284.8 eV of surface adventitious carbon, and the elemental compositions were determined from peak area ratios after correction for the sensitivity factor for each element. Brunauer-Emmett-Teller (BET) surface area measurements were conducted using the N₂ adsorption-desorption isotherms obtained at 77 K using Quantachrome Instruments version 3.0.

1.3 Photocatalytic activity

The photocatalytic activities of the prepared sample were assessed by the photodegradation of MO under the UV-visible light irradiation by using a 70 W metal halide at ambient temperature in air with magnetic stirring^[23-25]. In a typical experiment, 25 mg of g-C₃N₄ and 1 mmol/L Ethylenediaminetetraacetic acid (EDTA) were dispersed into 50 mL MO aqueous solution (10 mg/L) at room temperature. Before the light irradiation, the suspension was firstly sonicated for 10 min and then magnetically stirred for 20 min in the dark to obtain an adsorption-desorption equilibrium. Then the light was turned on and a 3 mL sample was taken out at certain time intervals from the reaction system and centrifuged to remove the photocatalyst powders for analysis. The concentration of the target pollutant was analyzed by using a Shimadzu UV-2550 UV-vis spectrophotometer. The relative concentration (C/C₀) of the MO solution was calculated at 463 nm, in which C₀ and C are the concentrations of MO at the beginning of light irradiation and at time t respectively.

1.4 Active species detection

Radical capturing experiments were conducted to identify the possible photocatalytic reaction mechanism of this reaction system. The detection process was similar to the photodegradation experimental process. Three radical scavengers, isopropanol (IPA) as hydroxyl radical (·OH) scavenger, ethylenediamineteraacetic acid disodium salt (EDTA-2Na) as h⁺ radical scavenger and p-benzoquinone (BQ) as superoxide radical (·O₂^-) scavenger, were selected to investigate the role of radicals in the photocatalytic degradation of MO^[24-25].

2 Results and discussions 2.1 Structural characteristics

The as-prepared g-C₃N₄ was characterized by a series of methods (XRD, FTIR and SEM). Fig. 1a shows the XRD pattern of as-prepared g-C₃N₄. The catalyst exhibits two peaks at 12.8° and 27.4°, corresponding to (100) and (002) crystal planes of g-C₃N₄. The characteristic peak at 27.4° corresponds to the interlayer stacking reflection of conjugated aromatic systems. The weak peak at 12.8° can be attributed to in-plane repeating units of continuous heptazine framework. The chemical structures of the photocatalysts were analyzed by FTIR spectrum, as shown in Fig. 1b. Three strong absorption peaks at 3650~3300 cm^-1, 1240~1650 cm^-1 and 810 cm^-1 present in the sample. The broad peak at 3650~3300 cm^-1 is ascribed to the stretching vibration of N-H and the stretching vibration of O-H of the surface adsorbed water molecules. The peaks in the range of 1240~1650 cm^-1 belongs to the stretching vibrations of tri-s-trizaine skeleton ring, and the bending mode of trizaine units at 810 cm^-1. The morphology and microstructure of the g-C₃N₄ was characterized by SEM. As can be seen in Fig. 1c, the g-C₃N₄ composes a large number of irregular particles. As shown in Fig. 1, the results are consistent well with reported in the literatures^[26-31].

2.2 Effect of EDTA on degradation of MO

Fig. 2a illustrates the influence of the addition of EDTA on the degradation of MO. As shown in Fig. 2a, negligible removal of MO was found in the presence of g-C₃N₄ (CN) after 20 min of irradiation, but 97.59% MO was degraded in the addition of EDTA under the same condition. This result indicated that the addition of EDTA greatly promotes the photocatalytic degradation of MO.

As shown in Fig. 2b, due to the addition of EDTA, the maximum absorption wavelength of MO is red-shifted from 463 to 501 nm, and the intensity of the absorption peak increased before the light irradiation. For MO, it has quinoid and azo structures under acidic and basic conditions, respectively. The chemical structure of MO changes following the equation in Fig. 3^[32]. The addition of EDTA increased the acidity of the MO solution, made the structure of MO changed from azo to quinoid, and the color of the solution changed from yellow to orange-red. Therefore, the maximum absorption wavelength of MO is red-shifted and the peak intensity is changed in the UV-Vis absorption spectrum. So, the relative concentration (C/C₀) of the MO solution is calculated at 463 and 510 nm, in which C₀ and C are the concentrations of MO at the beginning of light irradiation and at time t respectively. From the calculation results, the value of C/C₀ is greater than 1 during the dark.

2.3 Mechanism for the photocatalysis

Herein, a series of tests were designed to probe the photocatalytic degradation process. Firstly, to identify the main active species in the photocatalytic degradation process, the trapping experiments by adding various scavengers were conducted. As we all know, BQ can scavenge ·O₂^- from the reduction reaction of O₂ by photoexcited electrons on conduction band (CB), IPA can quench ·OH which is generated by the oxidation of water by h⁺ and EDTA-2Na can trap photogenerated h⁺ on valence band (VB). From Fig. 4a, the photocatalytic activity of g-C₃N₄ is seriously suppressed by the addition of BQ, indicating that ·O₂^- is the main active specie in the photocatalytic reaction. Negligible effect of IPA on the degradation of MO is observed, revealed that ·OH radical is not the domina- ting active species for MO photodegradation. However, in the presence of EDTA-2Na, the photocatalytic degradation was improved to a certain extent. This interes- ting phenomenon inspired us to further study.

As we known, EDTA, EDTA-2Na and EDTA-4Na all contain four carboxyl groups, but their aqueous solutions are different in acidity and basicity. So, under the same experimental conditions, the effects of EDTA, EDTA-2Na and EDTA-4Na on the photocatalytic degradation of MO over g-C₃N₄ were compared with the additives were all at the same concentration (1 mmol/L). The pH values of EDTA, EDTA-2Na and EDTA-4Na solutions are 2.35, 4.31, and 9.24, respectively. The experimental results are shown in Fig. 4b. The photodegradation activities of MO were improved after adding the three reagents respectively, but the effect of EDTA on the promotion was the most obvious that led the MO almost completely degrade after 20 min of light irradiation. EDTA-4Na also had a certain increase of the degradation, but it was not as good as EDTA and EDTA-2Na. MO degraded only 24.8% with EDTA-4Na addition, while EDTA-2Na can reach 62.08% after only 1 hour of light irradiation. The main differences of these three substances are the acidity, it is obvious that the H⁺ concentration of EDTA aqueous solution is the largest and the smallest is EDTA-4Na. Under acidic condition, MO dye had higher photocatalytic degradation efficiency based on these experimental results. So, we may infer that its quinoid structure is more prone to photocatalytic degradation than the azo structure.

The pH of the EDTA, EDTA-2Na, and EDTA-4Na systems was adjusted to be the same and degraded under illumination (pH =2.35). As shown in Fig. 4c, the changes in these three systems are almost negligible, further demonstrated the addition of EDTA, EDTA-2Na and EDTA-4Na mainly changed the acidity of the reaction solution.

To verify the influence of carboxylate anions on the degradation of MO by EDTA, the photocatalytic performance of g-C₃N₄ photocatalyzed degradation MO was compared by using HCl and EDTA as additives in the same pH (3.03), respectively. As shown in Fig. 5a, after 20 min of light irradiation, 33% and 98% of MO is degraded in the presence of HCl and EDTA, respectively, EDTA have a superior efficiency than HCl. The effect of EDTA is greater than that of HCl under the same pH, indicating that the H⁺ is not the only factor for the degradation of MO, whereas carboxyl group also plays an important role. According to the 'photo-Kolbe' reaction^[33], carboxylate anions can react with h⁺ to generate organic radicals after removal of CO₂. Photogenerated h⁺ are constantly accepted by carboxylate anions in EDTA, reducing the recombination of e^- and h⁺, resulting in an increase in the amount of ·O₂^- which is formed by the reduction reaction of O₂ by photoexcited e^-. Finally, MO is degraded by ·O₂^-.

According to the result of Fig. 4a, ·O₂^- is the main active species in the photocatalytic degradation process. EDTA as an efficient h⁺ accepter is conducive to the charge separation and lowering the recombination rate of photogenerated e^-/h⁺ pairs then bring about higher photocatalytic degradation rate. To further verify the role of ·O₂^-, formed by the reduction reaction of O₂ by photoexcited e^-, the photocatalytic degradation of MO were compared with and without O₂^[34]. As shown in Fig. 5b, it can be clearly observed that the photodegradation reaction is rather weaker without O₂. When the air was isolated the formation of ·O₂^- was reduced, therefore, the rate of photodegradation decreased.

To identify the change of electronic structure and photoelectric property of g-C₃N₄ with the addition of EDTA, UV-vis diffuse reflectance spectroscopy (DRS) (Fig. 6a) were used to characterize the g-C₃N₄ and the mixture of g-C₃N₄ and EDTA after sonicated and stirred (CN-EDTA). As shown in Fig. 6a, there is no obvious change in the absorption edge and derived electronic band gaps. So, the addition of EDTA has not changed the electronic structure and photoelectric pro- perty of g-C₃N₄. The Brunauer-Emmett-Teller (BET) surface areas of CN and CN-EDTA were measured by N₂ adsorption-desorption measurements at 77.4 K. As shown in Fig. 6b, the BET surface area of CN-EDTA is a little larger than CN. Thus, we suspect that EDTA promotes g-C₃N₄ degradation of MO may also be due to an increase of CN surface area.

To further research the CB and VB of g-C₃N₄, the valence band X-ray photoelectron spectroscopy (VB XPS) was tested. From Fig. 7, it can be seen that the VB potential of CN is 2.06 eV (vs. NHE). Combined with the UV-vis DRS result, the CB potential of CN is -0.67 eV (vs. NHE). From Fig. 7, we see that the CB edge potential of CN is more negative than the standard redox potential O₂/·O₂^- (-0.33 V vs. NHE) to reduce the molecular oxygen to yield ·O₂^-. But the VB potential of CN is small difference with the standard redox potential of OH^-/·OH (+1.99 V vs. NHE)^[35]. Consequently, we might infer the h⁺ cannot efficiently oxidize OH^- to generate ·OH radicals. The above conclusions are consistent with the results of the trapping experiments.

Based on the above results, a plausible mechanism of EDTA promoted photodegradation of MO by g-C₃N₄ was proposed:

$ g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}}{\rm{ + }}h\nu \to {\rm{}}g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}}{\rm{ - }}{{\rm{h}}^{\rm{ + }}}{\rm{ + }}g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}}{\rm{ - }}{{\rm{e}}^{\rm{ - }}} $

(1)

$ g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}}{\rm{ - }}{{\rm{e}}^{\rm{ - }}}{\rm{ + }}{{\rm{O}}_{\rm{2}}} \to {\rm{ \cdot}}{{\rm{O}}_{\rm{2}}}^{\rm{ - }}{\rm{ + }}g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}} $

(2)

$ {\rm{MO + EDTA}} \to {\rm{ MO }}\left( {{\rm{quinoid}}} \right){\rm{ + RCO}}{{\rm{O}}^{\rm{ - }}} $

(3)

${\rm{RCO}}{{\rm{O}}^{{\rm{ - }}}}{\rm{ + }}g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}}{\rm{ - }}{{\rm{h}}^{\rm{ + }}}{\rm{}} \to {\rm{ RCOO\cdot + }}g{\rm{ - }}{{\rm{C}}_{\rm{3}}}{{\rm{N}}_{\rm{4}}} $

(4)

$ {\rm{RCOO\cdot}} \to {\rm{ R\cdot + C}}{{\rm{O}}_{\rm{2}}} $

(5)

$ {\rm{MO }}\left( {{\rm{quinoid}}} \right){\rm{ + \cdot}}{{\rm{O}}_{\rm{2}}}^{\rm{ - }} \to {\rm{Photoproducts}} $

(6)

As depicted in Fig. 7, the photogenerated e^- and h⁺ are produced in the conduction band (CB) and valence band (VB) of g-C₃N₄ under irradiation (Eq.1). The photogenerated e^- can induce the adsorbed O₂ into ·O₂^- (Eq.2) because of its sufficient reductive ability. Meanwhile, the addition of EDTA makes the aqueous solution acidic and the structure of MO at this time is quinoid (Eq.3). Carboxylate anions capture photogenerated h⁺ which inhibits the recombination of photogenerated e^- and h⁺ pairs and generate organic radicals after removal of CO₂ (Eq.4 and 5). EDTA promotes the separation of e^- and h⁺ pairs, increasing the amount of ·O₂^- which could oxidize MO (quinoid) to products (Eq.6). Finally, g-C₃N₄ photocatalytic degradation of MO greatly increased.

2.4 Effect of EDTA on degradation of other pollutant

To confirm the universality of the method, we have studied the photocatalytic degradation of rhodamine B (RhB). Under the same conditions, this reaction system used 25 mg g-C₃N₄, 1 mmol/L Ethylenediaminetetraacetic acid (EDTA), 50 mL RhB aqueous solution (10 mg/L), and the relative concentration (C/C₀) of the RhB solution was calculated at 554 nm. As shown in Fig. 8, 99% RhB is degraded in the addition of EDTA after 20 min of irradiation, while only 16.37% RhB is degraded in the presence of g-C₃N₄. The result also verified the excellent ability of the addition of EDTA to remove organic pollutants in water.

3 Conclusions

In this work, we demonstrated an efficient approach to improve the photocatalytic activity of g-C₃N₄ by adding EDTA. According to the experimental results, EDTA trapped the h⁺ and effectively promoted the separation rate of the e^-/h⁺ pairs, which is the main reason for the raised photocatalytic activity of g-C₃N₄. What's more, the addition of EDTA has not changed the electronic structure and photoelectric property of g-C₃N₄. The trapping experiments verified that ·O₂^- is the major oxide specie for the photocatalytic degradation of MO, then, the plausible mechanism was proposed. This work offers a new promising route for the rapid degradation of organic pollutants by adding some kind of agents to inhibit the recombination of e^-/h⁺ pairs of the photocatalysts.