Oxygen doped graphitic carbon nitride nanosheets for the degradation of organic pollutants by activating hydrogen peroxide in the presence of bicarbonate in the dark

The development of novel wastewater treatment processes that use heterogeneous catalysts to activate hydrogen peroxide (H2O2) with bicarbonate (HCO3−) has been a subject of great interest in recent years; however, significant challenges remain, despite research into numerous metal-based catalysts. The work presented herein employed oxygen-doped graphitic carbon nitride (O/g-C3N4) as a non-metal catalyst for activating H2O2 in the presence of HCO3−, and this method represented the first system capable of removing organic pollutants in the dark, to our knowledge. The catalysts were characterized using several microscopic imaging, spectroscopic, electrochemical, and crystallographic techniques, as well as N2-physorption procedures. Analysis of the results revealed that the O/g-C3N4 catalyst possessed a high specific surface area and many defect sites. Various operational parameters, including the relative amounts of HCO3−, H2O2, and O/g-C3N4, were systemically investigated. A clear performance enhancement was observed in the degradation of organic contaminants when subjected to the HCO3−–H2O2–O/g-C3N4 system, and this result was ascribed to the synchronous adsorption and chemical oxidation processes. The novel system presented herein represented a new water treatment technology that was effective for removing organic contaminants.


Introduction
The bicarbonate anion, HCO 3 À , is relatively nontoxic and is common in natural water ecosystems. 1 For example, its concentration is known to reach 50-200 ppm in biological systems, and it may be presented at 14.7-25 mM in humans. It is worthwhile to note that HCO 3 À is always generated from CO 2 as a main product of advanced oxidation processes, which are typically impractical for certain applications. Directly introducing HCO 3 À into a solution containing H 2 O 2 oen inhibits the degradation reaction; however, a few cases demonstrate that this process can promote the removal of organic pollutants. [2][3][4][5][6][7][8][9][10][11] Despite this promising result, the huge required dosage is one of the main challenges associated with this approach because it signicantly increases the cost of the treatment. To reduce this input cost, various transition metal ions, such as cobalt(II) [12][13][14][15][16][17][18] and copper(II), 19,20 have been tested as homogeneous catalysts.
The results of those studies demonstrate that such metal catalysts accelerate the degradation of organic pollutants and distinctly reduce the required concentrations of HCO 3 À and H 2 O 2 . However, it is extremely difficult to recover the catalyst from the product solution, which is a signicant drawback of this approach. As an alternative, various heterogeneous catalysts have been carefully developed to avoid such problems. To date, these systems have involved metal-based materials, such as Co-based [21][22][23][24][25] and Cu-based catalysts. [26][27][28] For example, Guo et al. 23 reported an S-modied CoFe 2 O 4 catalyst, which degraded acid orange II in a reactive system containing HCO 3 À and H 2 O 2 .
Similarly, Pi et al. 25 showed that their Co x Mn 3Àx O 4 material was an effective catalyst for removing chlorophenols in the presence of a naturally-occurring concentration of HCO 3 À . Despite these advances, the leaching of toxic metal ions must be confronted during these reactions, because this phenomenon easily caused a second contamination and may reduce the stability of the catalyst. Therefore, other useful strategies have been explored. 29,30 For example, Pétrier et al. 29 adopted sonochemical technology to enhance the degradation of bisphenol in the presence of HCO 3 À . However, the required complex equipment input increased the cost. Therefore, in order to replace metalbased heterogeneous methods, there is a pressing need to develop simple non-metal-based heterogeneous catalysts that are inexpensive and green, and have excellent activity in combination with H 2 O 2 and HCO 3 À .
Among non-metal heterogeneous catalysts, graphitic carbon nitride is an advantageous choice because it is safety, low cost, and has high stability. Importantly, it exhibits excellent catalytic performance under visible light illumination. Therefore, this type of catalyst has been applied for various purposes, including hydrogen production, [31][32][33][34] but there are limited reports describing its activity for dark Fenton-like reaction. Cui et al. 32 reported the degradation of organic dyes using g-C 3 N 4 in the presence of H 2 O 2 ; however, this system showed hardly any catalytic activity in the dark. To attain the aforementioned target activity, it is necessary to rst explore the catalytic sites in g-C 3 N 4 to evaluate H 2 O 2 decomposition into reactive oxygen species (ROS). In general, this type of catalyst is characterized by extremely low specic surface area, which plays a vital role in the adsorption of organic pollutants. This factor signicantly hinders the broad application of such catalysts in industry. To overcome this challenge, our group 34 developed a novel g-C 3 N 4 catalyst using oxygen modication. This O/g-C 3 N 4 material possessed a huge specic surface area and exhibited some catalytic activity for the dark Fenton-like reaction. To further improve its catalytic performance, further system modication is necessary. As mentioned above, it can be expected that introducing HCO 3 À to the system should strengthen its capability to degrade organic pollutants in wastewater. This modi-cation creates a weakly alkaline environment, which is favorable for decomposing H 2 O 2 into ROS. 21 However, it may generate new radicals, such as CO 3 c À . The surface of g-C 3 N 4 has a negative charge under near-alkaline conditions; therefore, it generates a counteracting force whereby HCO 3 À leaves the surface of the catalyst, thus inevitably capturing ROS, such as cOH radicals. Once the cOH radicals enter the solution, they are immediately captured by HCO 3 À to produce CO 3 c À radicals, which have a longer lifetime. A literature survey 35,36 conrmed that CO 3 c À radicals could directly remove organic pollutants. Still, the relationships between the specic surface area, nitrogen defective sites, and catalytic activity must be further elucidated, especially in terms of the dark Fenton-like reaction. Herein, we describe a novel treatment technique for removing organic pollutants during dark Fenton-like reaction. To our knowledge, this is the rst time that O/g-C 3 N 4 has been used as a heterogeneous catalyst in a system containing HCO 3 À and H 2 O 2 . Various factors, including the specic surface area and nitrogen defective sites, are systemically studied and discussed in terms of how they inuence the overall catalytic activity. A clear enhancement in organic pollutant degradation is observed using this novel system.

Materials and reagents
All chemical reagents were purchased with analytical purity and used without any further purication. Deionized water (resistivity 18.2 MU) was ltered using a Millipore Milli-Q water purication system.

Preparation of O/g-C 3 N 4 catalyst
The synthesis of the O/g-C 3 N 4 catalyst has been reported previously by our group. 34 Specically, melamine powder was placed in a crucible with a cover, and then calcinated in a static air atmosphere at 550 C for 4 h. The obtained yellow sample was denoted as g-C 3 N 4 . This g-C 3 N 4 was added to deionized water, then transferred into a Teon-sealed autoclave and maintained at 180 C for 4 h. Aer cooling to room temperature, the obtained sample was washed with more deionized water and dried. Finally, the dried sample was placed in a crucible with a cover and calcinated for 4 h at different temperatures (350, 450, and 550 C) in a static air atmosphere. The resulting samples were denoted as O/g-C 3 N 4 -T (where T ¼ 350, 450, or 550 C). Unless otherwise stated, the sample at 550 C was denoted as simply, O/g-C 3 N 4 . Additionally, the Na-O/g-C 3 N 4 variant was synthesized using a wet impregnation method. Essentially, the O/g-C 3 N 4 catalyst was added to a solution containing NaHCO 3 , which was stirred vigorously for 24 h at room temperature. Next, they were dried. Finally, the sample was placed in a crucible with a cover, and calcinated for 4 h at 300 C in a static air atmosphere (the theory value of Na + is 10 wt%).

Characterization
X-ray diffraction (XRD) spectroscopy was carried out with a Bruker D8-Advance X-ray diffraction instrument; N 2 -physisorption was conducted on a Quantachrome Autosorb-1 instrument at liquid-N 2 temperature; scanning electron microscopy (SEM) was carried out on a JEOL JSM 6700 F operating at an accelerating voltage of 10 kV; transmission electron microscope (TEM) was conducted on TALOS F200 X instrument operating at an accelerating voltage of 200 kV; electron paramagnetic resonance (EPR) was conducted on Bruker A300; X-ray photoelectron spectra (XPS) was examined on Thermo Fisher Scientic using Al Ka; zeta potential was carried out on Zetasizer Nano ZSP instrument.

Degradation of organic pollutants in the dark and under illumination
The degradation reactions were performed in a ask at 25 C, at atmospheric pressure. The reaction solution contained a certain amount of catalyst, NaHCO 3 , H 2 O 2 , and organic pollutants. The reaction mixture was stirred vigorously in the dark unless otherwise stated. For the photocatalytic reactions, the reaction mixture was placed under LED illumination, using all other reaction conditions identical to the experiments conducted in the dark (i.e., same catalyst dosages and concentrations of H 2 O 2 and HCO 3 À ). In all cases, a certain aliquot of each reaction solution was extracted at xed intervals and then separated. The liquid was collected, and they were analyzed using a UV-vis spectrophotometer to quantify the concentration of residual organic pollutants.

Characterization
Fig . 1 displayed the XRD patterns of the g-C 3 N 4 and O/g-C 3 N 4 catalysts. The diffraction peaks at 2q ¼ 13.0 and 27.5 were observed for the g-C 3 N 4 species, and these were assigned to the repeating tri-s-triazine units within the g-C 3 N 4 unit layer (100), and the inter-planar stacking of unit layers (002), 32 respectively. Besides the characteristic peaks associated with g-C 3 N 4 , the O/g-C 3 N 4 catalyst didn't exhibit any diffraction peaks corresponding to other phases, indicating that the original structure of g-C 3 N 4 was largely retained following the oxygen modication. However, the strongest diffraction peaks for the O/g-C 3 N 4 catalyst shied toward higher angles (i.e., toward 2q ¼ 27.8 ), relative to those observed for g-C 3 N 4 , representing the decreased distances within the layered structure. 11 This change was attributed to the doping effect of oxygen atoms and the resulting distortion of the graphite structure. The images of the g-C 3 N 4 and O/g-C 3 N 4 catalysts captured using SEM characterization were shown in Fig. S1. † It is clear from these images that the g-C 3 N 4 catalyst contained agglomerated particles, and its surface appeared relatively rough. Compared to the bulk g-C 3 N 4 , a cotton-like morphology was visible in the O/g-C 3 N 4 image. This is likely because the nanosheets of g-C 3 N 4 were partially decomposed following the hydrothermal-calcination treatment. This difference was further conrmed through characterization using TEM characterization ( Fig. S2 †). Table S1 † listed the contents of carbon, nitrogen, and oxygen atoms, which were determined by elemental analysis of the g-C 3 N 4 and O/g-C 3 N 4 samples from SEM characterization. It is observed that O/g-C 3 N 4 contained lower nitrogen atom content but a higher atomic ratio of carbon to nitrogen, relative to g-C 3 N 4 . These results indicated that delamination and depolymerization processes were happened, causing loss of nitrogen atoms and creation of nitrogen defect sites. To conrm the existence of such defect sites, EPR measurements were carried out, and the results were presented in Fig. S3. † There were six large positive peaks observed in the spectrum of the g-C 3 N 4 catalyst, but only four positive peaks in the O/g-C 3 N 4 catalyst's spectrum. These results suggested that the latter had a greater quantity of unpaired electrons than the former. This can be justied based on the fact that some nitrogen atoms have been removed and other nitrogen atoms were replaced by oxygen atoms, in agreement with the XRD and elemental analysis. The analysis of all of these results led to the conclusion that the nitrogen content, and especially the morphology, of O/g-C 3 N 4 was altered relative to those of g-C 3 N 4 .
The textural properties of g-C 3 N 4 and the series of O/g-C 3 N 4 -T catalysts, which were determined from N 2 adsorption-desorption experiments, were compiled in Table 1. Relative to the g-C 3 N 4 , and within the series of O/g-C 3 N 4 -T catalysts, the specic surface area (S BET ), rst slightly increased and then increased remarkably, up to >23 times. Simultaneously, the total pore volume (V total ) rst increased, and then signicantly increased, up to approximately 10 times. It is well-known that g-C 3 N 4 produced from various precursors can be easily obtained via high-temperature calcination, but the resulting S BET was typically below 10 m 2 g À1 . This is because the interactions between the layers of g-C 3 N 4 were too strong, owing to the van der Waals forces and/or hydrogen bonds in the material, which led to serious aggregation of the particles. Aer the hydrothermal treatment, these interactions were weakened due to the attacks from water molecules at high temperature and pressure. The resulting exposed nanosheets were further attacked by oxygen during the high-temperature calcination, thus creating a new morphology and increasing the S BET . Based on the SEM, TEM, and pore size distribution results (see Fig. S4 †), we determined that the increased S BET was mainly attributed to the change in the morphology of the catalyst. In general, the large S BET is an important factor for strengthening catalytic performance, because this represents a greater proportion of exposed adsorption and catalytic active sites. Fig. 2 displayed the binding energies of the N 1s in g-C 3 N 4 and the series of O/g-C 3 N 4 -T samples, which were determined based on XPS. In the bulk g-C 3 N 4 , four different peaks at 398.9, 399.8, 401.5, and 404.6 eV were observed, corresponding to the sp 2 -hybridized nitrogen (N-C]N; denoted as sp 2 (N)), the sp 3hybridized nitrogen (N-[C] 3 ; denoted as sp 3 (N)), the N-H bonds, and the p band localized in heterocycles, respectively. 10,11,[14][15][16][17][18] The N 1s spectra associated with the series of O/g-C 3 N 4 -T samples contained similar peaks to the bulk g-C 3 N 4 . However, the ratio of sp 2 (N) to sp 3 (N) rst increased (see Table S2 †) and then decreased with increasing temperature in the O/g-C 3 N 4 -T series (i.e., 350 C vs. 450 C vs. 550 C), and as compared to g-    degradation rate was reduced, although it remained prominently higher than that measured in the absence of HCO 3 À (i.e.,

when [HCO 3
À ] > 10 mM, the reaction rate was still enhanced). These results showed that the concentration range for HCO 3 À was very broad under these experimental conditions. Literature reports 16,19 demonstrated that the impact of HCO 3 À concentration on the organic pollutant degradation varied depending on the system. For example, Cheng et al. 19 found that the degradation of acid orange II using a Cu(II)-H 2 O 2 system was restrained when the concentration of HCO 3 À was greater than 5 mM. This indicated that another reaction may have taken place in solution, besides that involving the catalyst. Surprisingly, the solution pH was found to increase slightly ($1.0 pH unit) aer the reaction, as shown in Fig. S6. † Lei et al. 28 observed the same phenomenon in their catalytic system containing a mixture of CuO-FeO and persulfate. According to eqn (1)-(6), singlet oxygen species can be obtained on the basis of pH changes aer the reaction.
(ii) Effect of the H 2 O 2 concentration. Fig. 4  concentration further did not lead to any change in the efficiency of RhB degradation. As mentioned above, the ROS were generated from the decomposition of H 2 O 2 , so it followed that, when the concentration of H 2 O 2 was relatively lower, fewer ROS could be formed. Therefore, by increasing the concentration of H 2 O 2 , the efficiency of RhB degradation was enhanced. However, the H 2 O 2 concentration was increased further, the negligible inuence on the RhB degradation indicated that the overall reaction predominately relied on the action of the catalyst.
(iii) Effect of the O/g-C 3 N 4 catalyst. Fig. 5 illustrated the impact of the O/g-C 3 N 4 catalyst on RhB degradation in the dark Fentonlike reaction. The degradation of RhB clearly increased with increasing catalyst loading, reaching a maximum value at a catalyst concentration of 0.8 g L À1 . The rate remained stable when additional catalyst was added. It is evident that the removal of RhB was closely related to the catalyst dosage, as shown in Fig. S7. † This is because a larger quantity of catalyst   offered more adsorption sites, thus enabling more RhB removal. However, addition of too much catalyst led to agglomeration and increased counter forces between the particles, ultimately reducing RhB adsorption. In contrast, an insufficient amount of catalyst was introduced, there would not have enough available sites to activate H 2 O 2 for producing ROS, so the RhB degradation rate stabilized or may decrease slightly.
Based on the described systematic screening experiments, the optimal reaction conditions were determined: [NaHCO 3 ] ¼ 10 mM, [H 2 O 2 ] ¼ 15 mM, and [O/g-C 3 N 4 ] ¼ 0.8 g L À1 . Applying these conditions, other important parameters were investigated. Fig. S8 † showed the inuence of metal ions (with Cl À as the counter anion) on RhB degradation in the dark Fenton-like reaction. Based on this gure, there appeared to be no clear difference in RhB degradation among the tested metal ions, indicating that these ions had no appreciable effect on the RhB degradation reaction. This is because they were not involved the catalytic reaction, but rather, only impact solubility. species, which then adsorbed H 2 O 2 . As a result, the cOH radicals were immediately generated, and attacked the RhB molecules on the surface of the catalyst. These processes led to a remarkable enhancement in RhB degradation. In contrast to transition metal ions, the metal ions studied herein were widely distributed in natural water environments, and they did not form M [HCO 3 À -H 2 O 2 ] 2 RhB complexes. We additionally studied the inuence of several anions (using Na + as the metal ion), and the results were displayed in Fig. S9. † It is determined that most anions, except for CO 3 À and CH 3 COO À , positively impacted the RhB degradation. These results demonstrated that the studied catalytic system was resilient in terms of metal ions and anions, so it represented great potential for applications in industry. The removal of other organic pollutants was also investigated, and the results were shown in Fig. S10. † All of the tested species were removed effectively using this system, further verifying that it was favorable for degrading various organic contaminants. Fig. S11 † displayed the impact of darkness versus LED illumination on the degradation of RhB. It is evident that the system containing O/g-C 3 N 4 , HCO 3 À , and H 2 O 2 exhibited better catalytic activity under LED illumination, relative to in the dark. This is because additional light-induced reactions occurred simultaneously with the dark Fenton-like reaction, as described by eqn (7)-(9); these represented a typical photo-Fenton-like reaction process. Specically, under LED light, the photogenerated hole (h + ) directly reacted with RhB, and the photogenerated electron (e À ) interacted with the dissolved oxygen in aqueous solution to form the O 2 c À species. These species could remove RhB, so the degradation of RhB was enhanced under LED illumination. The results of these experiments clearly showed that this catalytic degradation process could proceed under both light and dark conditions. This is a particularly useful nding for this new non-metal catalytic system, which is capable of eliminating organic pollutants, thus reducing environmental contamination as much as possible.
On the contrary, traditional carbon-based materials that promote Fenton-like reactions doesn't happen under light illumination, so the system developed herein clearly has an advantage in its ability to function in the dark.
Catalyst + hv / h + + e À The stability of the O/g-C 3 N 4 catalyst was investigated, and the results were presented in Fig. 6. Cyclic RhB degradation experiments were conducted under dark conditions. Specically, aer the reaction, the catalyst was carefully washed with deionized water and ethanol, respectively, and then dried. Aerward, it was employed for the next cycle following the addition of a fresh reaction solution containing HCO 3 À and H 2 O 2 , wherein the concentration of these components remained the same as in the rst reaction. As shown in Fig. 6, the RhB degradation rate hardly decreased, even aer several reaction cycles, thereby verifying the high stability of the system and repeatability of the process. The degradation of organic dyes in wastewater was completed using the system containing HCO 3 À and H 2 O 2 , and the results were compiled in Table 2. Based on these reactions, the HCO 3 À and H 2 O 2 consumptions were signicant in the absence of catalyst. When Co(II) or Cu(II) was added into the reaction solution, signicant degradation of the organic dyes was always observed. The main drawbacks associated with these catalysts included their cost, high toxicity, and difficult recovery. In order to nd a suitable substitute for homogeneous catalysts, one current method involved using heterogeneous catalysts. For example, diatomite-supported cobalt, 21 CoMgAl-Na-Y, 22 and S/CoFe 2 O 4 (ref. 23 ) have been applied for removing organic pollutants; however, these options also faced problems, such as their cost and complex preparation. Importantly, leaching of metal ions from these heterogeneous catalysts was inevitable and caused secondary contamination. These factors constituted the main challenges that must be overcome to allow their industrial application. Relative to such heterogeneous variants, the HCO 3 À -H 2 O 2 -O/g-C 3 N 4 catalytic system was green, inexpensive, and easy to operate. These advantages conrmed that the novel technology presented herein represented high potential for practical applications in the removal of organic pollutants from wastewater.
3.2.2. The mechanism of organic pollutant removal. To identify whether various ROS were produced in the system containing HCO 3 À and H 2 O 2 with the O/g-C 3 N 4 catalyst, we attempted to detect selected ROS, including cOH, O 2 c À , and 1 O 2 , using different radical scavengers. 21 First, a common radical scavenger for cOH, ascorbic acid, was introduced into the reaction mixture solution. As observed in Fig. S12, † addition of ascorbic acid caused the degradation rate to decrease. However, it did not completely quench the reaction, suggesting the existence of a radical-mediated degradation pathway, except for the adsorption process. Ethanol was a relatively more powerful radical scavenger, and Fig. S13 † displayed the inuence of ethanol on RhB degradation. It is clear that the addition of ethanol signicantly decreased RhB degradation, and this inhibition effect was concentration-dependent over a wide range. Therefore, we concluded that the cOH species was involved this reaction, but it was not the only ROS. In order to probe the production of O 2 c À along the reaction pathway, the inuence of different concentrations of the scavenger, Trion, on the RhB degradation was investigated. It is clear from Fig. S14 † that the RhB degradation in the presence of 5 mM Trion decreased remarkably at rst and then stabilized with increasing reaction time. This trend was also observed at higher concentrations (50 and 200 mM) of Trion. These results suggested that the O 2 c À radical was an important ROS. Another radical scavenger, benzoquinone, was also studied in this reaction, and the results were shown in Fig. S15. † It is determined that the degradation of RhB decreased with the addition of 3 and 6 mM of benzoquinone, respectively. This conrmed that the O 2 c À radical had an important role in the removal of organic pollutants in this system. Fig. S16 †shows the effect of different concentrations of NaN 3 (used for singlet oxygen detection) on the RhB degradation. In this case, the change toward the RhB degradation was not signicance at two different NaN 3 concentrations (4 and 10 mM). Therefore, we determined that 1 O 2 was not a major ROS in this degradation mechanism. In order to support this conclusion, another experiment was carried out by adding furfuryl alcohol (FFA), rather than NaN 3 , into the reaction, and the results were shown in Fig. S17. † Under these conditions, the destruction of the RhB structure. The structural change was further conrmed based on UV-vis results, as shown in Fig. S18. † It is worth noting that there was a large difference in the Na + content in the NaOH and NaHCO 3 systems (negligible (mM) vs. 10 mM, respectively), and its role was not thoroughly taken into consideration. To certify this assumption, preliminary experiments were conducted using Na-O/g-C 3 N 4 as the catalyst, and the results were displayed in Fig. S19, † clearly showing that there was little inuence on the removal of RhB. In summary, the high degradation capability of the HCO 3 À -H 2 O 2 -O/g-C 3 N 4 system was mainly ascribed to the synergistic effect between adsorption and chemical oxidation. The HCO 3 À and O/g-C 3 N 4 components were responsible for activating H 2 O 2 and removing RhB. Specically, the O/g-C 3 N 4 catalyst acted as a bifunctional material for dispelling RhB, thus contributing to the simultaneous adsorption and chemical oxidation processes. The huge specic surface area of the O/g-C 3 N 4 catalyst allowed signicant adsorption via a non-radical pathway, but the chemical oxidation was carried out through a radical route, which relied on the defective sites of O/g-C 3 N 4 for activating H 2 O 2 to ROS in the presence of HCO 3 À . Chemical oxidation also occurred in the reaction solution itself, and this mechanistic pathway was outlined in Scheme 1.
3.2.4. The relationship between the catalyst's structure and its catalytic activity. To rmly establish the relationship between the structure of the catalyst and its ability to degrade RhB, the effect of the calcination temperature when preparing O/g-C 3 N 4 -T was investigated. It is known that the calcination temperature inuenced the specic surface area and defective sites of the resulting O/g-C 3 N 4 . Fig. S20 † illustrated the effect of the calcination temperature across the series of O/g-C 3 N 4 -T catalysts on their RhB degradation performances. The O/g-C 3 N 4 -550 catalyst exhibited the maximum RhB degradation efficiency among all variants tested under identical conditions; the catalysts prepared with lower calcination temperatures demonstrated less RhB degradation efficiency (especially below 450 C). This behavior was likely due to their relatively lower specic surface areas and intact polymerization. Additionally, a higher calcination temperature induced a tremendous change to the g-C 3 N 4 structure, particularly at 550 C. In general, the higher the calcination temperature is employed, the greater the separation between nanosheets of g-C 3 N 4 is happened. It can also easily decompose to generate small gaseous products, which diffuse between the layers of nanosheets. In this case, the morphology of the catalyst is greatly altered, leading to a higher specic surface area, as conrmed by SEM and TEM characterizations. Simultaneously, numerous defect sites, such as vacancies and edges, are generated. The resulting increased specic surface area is favorable for molecular adsorption, because it offers more catalytic active sites (the defect sites can activate H 2 O 2 toward ROS in situ). [39][40][41][42][43][44] However, no interesting products are obtained if the calcination temperature is too high (>550 C). To test this hypothesis, the relationships between the S BET , the content and ratio of N(sp 2 ) and N(sp 3 ) (denoted as defect sites), and the catalytic activity were evaluated, and the results were compiled in Table 1 and S2. † The S BET increased gradually with applied temperatures below 550 C and then signicantly improved at 550 C, clearly indicating that the S BET was closely related to the catalytic activity. Accordingly, increasing the S BET of the catalyst led to increased RhB degradation. Meanwhile, the N(sp 2 ) and N(sp 3 ) ratio rst increased, when changing from 350 C to 450 C, but then began to decrease at 550 C. This behavior suggested that the defect sites corresponding to N(sp 3 ) can to some extent act as catalytic active sites. Combined with the analysis of the degradation efficiency, we concluded that the S BET and the defective sites from N(sp 3 ) were two important factors for enhancing the catalytic activity of the developed system.

Conclusions
This report discussed a newly-developed strategy for treating wastewater polluted with organic contaminants. The employed system consisted of HCO 3 À , H 2 O 2 , and an O/g-C 3 N 4 catalyst, which together demonstrated efficient degradation of organic compounds. Various inuencing factors were investigated, and the reaction conditions were optimized in terms of the catalyst loading and preparation, concentrations of HCO 3 À and H 2 O 2 , and light vs. dark operation. The catalytic system retained high stability and exhibited sufficient reproducibility aer running several reaction cycles. Mechanistic probing revealed that the degradation pathway was dominated by the synergistic effect between adsorption and chemical oxidation processes. The adsorption was closely related to the tunable specic surface area of the catalyst, and the chemical oxidation was achieved by various ROS. Overall, this work presented a novel, widelyapplicable method for removing organic pollutants from wastewater.

Conflicts of interest
There are no conicts to declare.