An efficient eco advanced oxidation process for phenol mineralization using a 2D/3D nanocomposite photocatalyst and visible light irradiations

Nanocomposites (CNTi) with different mass ratios of carbon nitride (C3N4) and TiO2 nanoparticles were prepared hydrothermally. Different characterization techniques were used including X-ray diffraction (XRD), UV-Vis diffuse reflectance spectroscopy (DRS), X-ray photoelectron spectroscopy (XPS), transmission electron spectroscopy (TEM) and Brunauer-Emmett-Teller (BET). UV-Vis DRS demonstrated that the CNTi nanocomposites exhibited absorption in the visible light range. A sun light - simulated photoexcitation source was used to study the kinetics of phenol degradation and its intermediates in presence of the as-prepared nanocomposite photocatalysts. These results were compared with studies when TiO2 nanoparticles were used in the presence and absence of H2O2 and/or O3. The photodegradation of phenol was evaluated spectrophotometrically and using the total organic carbon (TOC) measurements. It was observed that the photocatalytic activity of the CNTi nanocomposites was significantly higher than that of TiO2 nanoparticles. Additionally, spectrophotometry and TOC analyses confirmed that degraded phenol was completely mineralized to CO2 and H2O with the use of CNTi nanocomposites, which was not the case for TiO2 where several intermediates were formed. Furthermore, when H2O2 and O3 were simultaneously present, the 0.1% g-C3N4/TiO2 nanocomposite showed the highest phenol degradation rate and the degradation percentage was greater than 91.4% within 30 min.

regarding g-C 3 N 4 -based photocatalysis 24 . Additionally, several researchers have reported that coupling g-C 3 N 4 with TiO 2 may improve the photocatalytic behavior of TiO 2 by lowering its band gap energy and slowing the electron-hole recombination rate 5,14,16,25,26 .
Phenol, among numerous organic pollutants found in wastewater, requires great attention due to its toxicity 27,28 . Its existence has been confirmed in several industrial wastewaters, including chemical and petrochemical industries. Phenol treatment is expensive since it is unresponsive to traditional treatment techniques 1,29 . Therefore, current investigations are directed at advanced oxidation processes (AOP) as a replacement for traditional techniques.
Although a few research studies targeted the use of g-C 3 N 4 /TiO 2 composites in photocatalysis, most investigations were directed towards the degradation of dyes in the absence of oxidizing agents and using UV irradiation, which significantly increases the cost 5, 6, 9-11, 14, 15, 17, 26, 30-36 . In this work, the focus is on investigating the extremely enhanced kinetics of the advanced oxidation process in which the photocatalytic treatment of a high concentration of phenol (20 ppm), which is more difficult to mineralize than dyes, was done using a hydrothermally-prepared nanocomposite of 0.1% g-C 3 N 4 nanosheets on top of TiO 2 nanoparticles and a low-power Xe irradiation lamp (mainly irradiates visible light) as a simulator for sunlight. Furthermore, the effect of the in-situ addition of eco oxidants e.g. hydrogen peroxide (H 2 O 2 ) and/or ozone (O 3 ) (generated using a home-made electrode) on the rate of photocatalytic degradation of phenol was also included and, to our knowledge, this work has not been previously addressed.

Experimental
Synthesis. All chemicals were of analytical grade and no extra purification treatments were carried out.
Graphitic carbon nitride (g-C 3 N 4 ) was synthesized, as reported by Wang et al. 37 . Specifically, melamine (Alfa Aesar, 99% pure) was placed in a covered crucible and heated under stationary air conditions at 550 °C with a ramping rate of 2.5 ° min for 4 h. The resulting carbon nitride powder, which was pale yellow in color, was washed with deionized water and then dried under vacuum at 60 °C.
The g-C 3 N 4 -TiO 2 nanocomposite (CNTi) photocatalyst was prepared by loading various mass ratios of CN on a commercial P25 TiO 2 (Sigma Aldrich, 21 nm particle size) support as follows. TiO 2 was added to an absolute ethanol/deionized water mixture at a 1:1 ratio and sonicated for 30 min. The same exact procedure was conducted with CN. Afterwards, the two aforementioned suspensions were mixed together, further sonicated for 30 min and nitric acid/ ammonium hydroxide solutions were used to adjust the pH to 3.5. The resultant mixture was poured into a Teflon-lined stainless steel autoclave and left overnight at 120 °C. The suspension was centrifuged, later washed with hydrochloric acid (1 M) and then with deionized water. Lastly, the product was dried overnight at 80 °C. The loading percentages of C 3 N 4 on TiO 2 were designed to be 0.1, 0.5 and 1.0% and the corresponding names are 0.1CNTi, 0.5CNTi and 1CNTi, respectively.
Characterization. X-ray diffraction (XRD) measurements were conducted using a Bruker D8 diffractometer equipped with a Lynxeye detector and a Cu K α radiation source (λ = 15.406 × 10 −2 nm). The diffractometer operated at 40 mA and 40 kV. The scanning was performed from 10-80° using a scan step of 0.015° and a step time of 0.2 s. To determine the different phases, an automatic JCPDS library search and match was used. Additionally, standard SERACH and DIFFRACT AT (Australia) computer software packages were employed.
Raman spectra were obtained using an InVia Raman microspectrometer, which worked under macro conditions (f = 3 cm) with a 785 nm excitation line and a laser control of approximately 2 mW. The measurements were conducted without prior treatment and average spectra were recorded from five registration scans. The spectra ranged from 100-3500 cm −1 with three accumulation numbers and a 10 s exposure time for each registration.
UV-Vis diffuse reflectance spectroscopy (DRS) was performed using a Cary 5000 UV-Vis-NIR (Agilent, Australia) equipped with an integrating sphere accessory. The test samples were supported by KBr, and the measured spectra were recorded at room temperature from 200 to 800 nm with a resolution of 0.05 nm.
The XPS spectra were obtained using Thermo Scientific ESCALAB-250Xi spectrometer. A monochromatic AlK α radiation source that operated at a power of 300 W (20 mA, 15 kV), was utilized. The vacuum in the analysis chamber was less than 7 × 10 −9 Torr during the measurements. Carbon pollution that was referenced to C1s at 284.8 eV with a ±0.2 eV experimental error was taken as a base for elemental binding energies.
The amount of carbon, hydrogen and nitrogen were measured using a GmBH analyzer CHNS-O EA. The BET (Brunauer-Emmett-Teller) surface areas were determined by an automatic ASAP 2010 Micromeritics sorpometer (USA) that was equipped with an outgassing platform and an online acquiring data and handling system, which functioned using numerous computer-run methods to examine the adsorption data.
Thermogravimetric analyses (TGA) were achieved utilizing a Perkin Elmer, USA/Pyris 1 TGA instrument. The air flow was heated at 5 °C min −1 to quantify the exact loading percentage of the carbonaceous material (CN) on the TiO 2 support.
The CNTi nanocomposite was characterized using a Japanese JEOL 2100 F high-resolution transmission electron microscopy (HR-TEM) operating at 200 kV and coupled with an energy dispersive x-ray unit. The sample preparation was conducted by mixing a small quantity of sample with ethanol and sonicating for 15 min. The sample was then loaded onto the TEM grid.
A SiO 2 /Ti (300 nm)/Pt (100 nm)/TiOx (100 nm)/(SnOx500 nm) electrode was used to generate ozone from a 0.5 M NaOH solution at 2 V versus the Ag/AgCl reference electrode. The electrode was manufactured at The Pennsylvania State University (USA) Materials Research Institute Nanofabrication Laboratory using a KJL CMS-18 sputtering system. The films were grown at room temperature on a quartz sample (SiO 2 ) at 5 mtorr and 200 watts. The metallic films (Ti and Pt) were grown under an argon atmosphere, whereas the oxide films were grown in a mixture of 15% O 2 in Ar. The generated ozone was transferred using Tygon tubing to the photocatalytic reactor. The electrode had an ozone production efficiency close to 20% (excluding the dissolved ozone, which was not measured in the generating solution).

Photocatalytic experiments.
For every photocatalytic run, 20 ppm of a phenol aqueous solution was prepared using deionized water. A 0.1 g catalyst was added to 100 mL of the phenol solution and stirred in the dark for 30 min until the adsorption equilibrium was achieved. The pH of the phenol solution was not adjusted. The photoreaction vessel was irradiated by a xenon lamp (150 W) without a cut-off filter at an integrated intensity of 12 mW cm −2 at 4.5 cm. The beginning of the experiment (time zero) occurred when the lamp was turned on. Samples were drawn every 5 min from the reaction vessel and filtered to dispose of the suspended catalyst particles through a nylon filter membrane with a 0.4 μm porosity. The UV-Vis spectrophotometer was used to determine the amount of phenol at a maximum absorbance of 298 nm. Experiments that involved O 3 , were electrochemically produced in an isolated cell that contained 0.5 M NaOH using a SiO 2 /Ti (300 nm)/Pt (100 nm)/ TiOx (100 nm)/SnOx (500 nm) electrode with a maximum current of 3000 mA at 8 V. The undissolved portion of the generated ozone was transferred to the photoreactor using a Tygon tube. A 4 ppm quantity of electrochemically generated ozone gas was used for each photocatalytic run. The photocatalysis system that included an O 3 generating unit (electrochemical cell plus a potentiostat), an ozone meter, a UV lamp chiller and the photocatalytic reactor is shown elsewhere 38 .
Finally, after photocatalytic experiments, the total organic carbon (TOC) for all phenol solutions were measured using a TOC-VPH analyzer, Shimadzu, Japan.

Results and Discussion
Characterization. Figure 1 displays a HR-TEM micrograph for the 1CNTi nanocomposite loaded on the TiO 2 support. It is evident that the average particle size is approximately 20 nm and C 3 N 4 was uniformly distributed on TiO 2 . It also shows the lattice spaces of TiO 2 (0.35 nm) and the lattice fringes of C 3 N 4 .
To determine the C 3 N 4 contents in the final products, TGA was employed from 20 to 700 °C in air. Pure TiO 2 shows almost no weight loss in the studied temperature range. However, pure C 3 N 4 experienced rapid weight loss from 550 to 640 °C, which indicates its decomposition. For the CNTi nanocomposite, the weight loss increased at temperatures above 500 °C. At the end of the analysis, the total combustion of CN was achieved. Therefore, the weight loss is related to the amount of CN in the composite. The results obtained from the TG studies indicated that the CN content in the respective composites corresponded to nominal values. For example, Fig. 2 shows the TGA of the 1CNTi composite. Figure 3 displays FT-IR spectra of TiO 2 , CNTi and C 3 N 4 . In Fig. 3a, a peak was detected at 1627 cm −1 , which was ascribed to deformed water molecules or a Ti-O-Ti stretching vibration 22 . Additionally, the low frequency strong and broad absorption band (below 1000) was assigned to the Ti-O-Ti vibration in TiO 2 [39][40][41][42] . The broad band in the range of 3800 to 3000 cm −1 was designated as O-H stretching vibrations of C-OH groups (from carboxylic groups) or/and intercalated water molecules. The C 3 N 4 spectrum in Fig. 3d showed several strong bands from 1130 cm −1 to 1639 cm −1 , which were associated with stretching and rotation vibrations of C-N and C = N in heterocycles 8 . The band at 805 cm −1 was related to s-triazine ring vibrations and the wide absorption peak at 2000-3500 cm −1 was attributed to the N-H bond 8,10,11,31,43 . The main characteristic peaks of C 3 N 4 were reduced and several peaks disappeared after loading on TiO 2 . These data indicated a low percentage of C 3 N 4 in the composites (Fig. 3b,c).
The XRD measurements for bare TiO 2 (Fig. 4a) show diffraction patterns of both anatase and rutile phases.  Figure 4e shows two distinctive diffraction peaks: a strong peak located 27.4° with an interlayer spacing of 3.2 Å and a broad peak located at 13.1° with an interlayer spacing of 6.8 Å, which can be indexed to the hexagonal phase of graphitic C 3 N 4 . The former peak was related to the (002) plane that was caused by the interlayer stacking of the conjugated aromatic system, whereas the latter peak was attributed to the (100) plane due to an in-plane structural packing motif. These two diffraction peaks are consistent with the results that have been reported in the literature 6,9,18,37,45,46 . The diffraction patterns of CNTi composites (Fig. 4b-d) were similar to that of bare TiO 2 without the distinctive peaks of C 3 N 4 . This is can be ascribed to either (i) the low quantity of C 3 N 4 in the composites or (ii) the peak overlap that relates to the interlayer stacking of the conjugated aromatic system with the peak at 27.5° for rutile TiO 2 7, 14, 47 . XPS measurements were performed to obtain information regarding the surface chemical composition along with their oxidation states. The survey spectra show Ti, C and O peaks in bare TiO 2 (Fig. 5Aa), C, O and N peaks in C 3 N 4 (Fig. 5Ab) and C, N, Ti and O peaks in the CNTi composite (Fig. 5Ac). Figure 5B shows the C region of bare TiO 2 , CN and CNTi. The deconvolution of the C region of bare TiO 2 (Fig. 5Bc) revealed three spectral peaks at 284.7, 286.2 and 288.8 eV, which were ascribed to adventitious carbon, -C-OH and -COOH groups, respectively 22 . In the C 3 N 4 spectrum (Fig. 5Bb), besides the spectral peak at 284.7 eV, which corresponds to C impurities, three peaks at binding energies of 286.6, 287.4 and 292.8 eV were allocated to the C-N-C, C-(N) 3 and C-NH 2 groups, respectively 6 . The CNTi composite preserved all the carbon functional groups that existed in both TiO 2 and C 3 N 4 , as indicated in the devolution of its C region (Fig. 5Bc). However, a new peak at a lower binding energy of 281.1 eV appeared, which can be assigned to the C-Ti bond 48 . The regional N 1 s spectra for C 3 N 4 and CNTi composites are presented in Fig. 5C. The spectrum of C 3 N 4 (Fig. 5Ca) was deconvoluted into four peaks, which are attributed to N = C at 397.9 eV, N-(C) 3    N functional groups existed in CNTi and a new peak at a lower binding energy of 395.4 eV, which may be ascribed to an O-Ti-N bond (Fig. 5Cb). Figure 5D displays Ti 2 P spectra of pure TiO 2 compared with that of the CNTi composite. The Ti 2p 3/2 and 2p 1/2 spin-orbit coupling peaks of bare TiO 2 appeared at 458.7 and 464.3 eV, respectively, with a spacing of 5.6 eV, which was ascribed to Ti 4+ species in the TiO 2 cluster. However, this peak showed a broadening with a 0.2 eV shift of Ti 2p 3/2 to a lower binding energy, whereas the 2p 1/2 remained un-shifted after loading C 3 N 4 on TiO 2 . The deconvolution of the Ti region of the CNTi composite (Fig. 5Cd) revealed the existence of Ti-N or/and Ti-C (at 457.8 and 463.3 eV for Ti 2p 3/2 and 2p 1/2 , respectively) with a spacing of 5.5 eV besides the TiO 2 main peak. Therefore, it can be concluded that the oxidation state of Ti (IV) did not change after the loading of C 3 N 4 on TiO 2 , and there is a chemical bond between the Ti and C and/or N of the C 3 N 4 . The negative shift in the Ti peak of CNTi due to electronic interactions between Ti and N atoms and/or C atoms caused an increase in the electron density on Ti. This was also supported by the existence of O-Ti-N in the N 1 s region (Fig. 5Cb) and the peak of C-Ti in the C 1 s region (Fig. 5Bc). The O 1 s spectrum (Fig. 5Ea) of pure TiO 2 were fit into two peaks at binding energies of 529.87 and 531.42 eV that were assigned to oxygen in TiO 2 and the surface hydroxyl, respectively 14,16,49 . In comparison with pure TiO 2 , a negative shift of the main O 1 s to 529.2 eV was observed for the CNTi composite. Note that this shift of 0.2 eV in O1s is similar to the shift found in Ti 2p 3/2 for the CNTi composite, which supports the suggestion of the existence of bonds between Ti and C and/or N and are in good agreement with previous reports 6, 14,15,33 . The XPS, XRD and FT-IR studies clearly revealed the successful preparation of nanocomposite material with chemically bound interfaces between C 3 N 4 and TiO 2 rather than a physical mixture of two separate g-C 3 N 4 and TiO 2 phases.
Argon bombardment was used to etch the CNTi nanocomposite to gain information regarding which material on top of the other. From Table 1, it is obvious that CN covered the TiO 2 surface since the quantity of TiO 2 increased while the quantity of carbon decreased with etching.
The elemental analysis of as-prepared C 3 N 4 provided a C to N ratio of 0.53, which is less than the theoretical value of ideal carbon nitride (0.75 in C 3 N 4 ). In addition to nitrogen and carbon, the elemental analysis also detected hydrogen. The presence of hydrogen in the form of N-H and NH 2 groups was established by FT-IR and XPS analyses, which describes the relatively low C to N ratio. Figure 6 shows the Raman spectra of TiO 2 , 0.1CNTi, 1CNTi and C 3 N 4 recorded with a 785 nm laser as the excitation source. The measured Raman spectrum of C 3 N 4 (Fig. 6d) is similar to the one shown by Tonda et al. 50 in which several characteristic peaks of C 3 N 4 were revealed with strong peaks at 705 and 1232 cm −1 51 . Similar to the XRD results, the Raman spectrum of C 3 N 4 loaded on TiO 2 did not show any characteristic peaks of C 3 N 4 . Only peaks corresponding to the anatase and rutile phases of TiO 2 were observed (Fig. 6a-c). This result was also observed by other researchers 51,52 .
The band gap energies of TiO 2 , 0.1CN, 0.5CN and 1CN were measured and the wavelengths at the absorption edge were calculated and recorded from the UV-Vis diffuse reflectance spectra ( Fig. 7 and Table 2). The measurements revealed that the band gap energy of C 3 N 4 equals 2.75 eV, as reported in the literature 6 . Interestingly, the band gap energy of the CNTi composite increased as the C 3 N 4 percentage in the composite increased but remained lower than that of the TiO 2 . The composite with the lowest band gap energy (2.95 eV) was 0.1CNTi compared to 3.35 eV for TiO 2 . Therefore, loading CN on TiO 2 shifted the TiO 2 absorption wavelength to the visible region and allowed the photocatalytic degradation of phenol using visible light or sunlight irradiation. Note that the BET surface area of 1CNTi composites (49.8-50.6 m 2 g −1 ) and TiO 2 (51.1 m 2 g −1 ) were similar. This proves that the surface area was not the controlling factor for this photocatalytic study.
Photocatalytic applications. Phenol degradation was studied using a xenon lamp photoexcitation source in the presence and absence of H 2 O 2 and/or O 3 . The % of phenol degradation was calculated using Equation 1, where C 0 is the adsorption equilibrium concentration of phenol before irradiation begins. C t represents the residual concentration of phenol after specific time t of irradiation. Apparent first order reaction rate law was found to fit the photocatalytic degradation of phenol with a perfect correlation constant (R 2 ) close to unity. The rate constant for the degradation, K, was obtained from the first-order plot (Equation 2).  Table 1. Argon etching of 1CNTi at different times.    Table 2. Wavelengths at the absorption edge for TiO 2 , 0.1CNTi, 0.5CNTi, 1CNTi and C 3 N 4 and their corresponding band gap energies.
Note that in absence of the photocatalysts (CNTi or TiO 2 nanocomposites), there was no degradation of phenol, even in presence of C 3 N 4 without the TiO 2 support. Furthermore, the C 3 N 4 did not peel off the TiO 2 support during and after the photocatalytic experiments, which can be attributed to the chemical bond between C 3 N 4 and TiO 2 during the synthesis process.  Table 3 summarizes the degradation rate constants in min −1 and TOC results. The following conclusions were obtained from Fig. 8 and Table 3. First, in all cases, the photodegradation percentage increased with time regardless of the type of photocatalyst. Second, all CNTi composites showed better photocatalytic activity compared with bare TiO 2 . An increase in the C 3 N 4 percentage from 0.1 to 1% in the CNTi composites decreased the photocatalytic activity, and 0.1CNTi showed the best performance in terms of the phenol degradation rate. Furthermore, the phenol degradation rate increased with the addition of H 2 O 2 regardless of the photocatalyst type. This is mostly due to the increase in hydroxyl radical formation and retardation of the electron/hole (e/h + ) pair recombination 53 due to the electron acceptor property of the C 3 N 4 where electrons move from TiO 2 to the carbonaceous materials inhibiting the recombination between the electrons and the holes.  56 . S. Obrego'n and G. Colo'n found that the modification of TiO 2 /g-C 3 N 4 composite with Pt and Mn improved the photocatalytic activity of phenol degradation under UV radiation. The best phenol rate was obtained 10.6 × 10 −8 mol/L/s over Pt-TiO 2 /g-C 3 N 4 -MnOx composite 57 . Finally, a comparison of the phenol degradation % based on UV spectrophotometer and TOC measurements indicated similar results when the CNTi nanocomposite photocatalyst was used. By contrast, when TiO 2 was used, the phenol degradation % from the UV-Vis spectrophotometer measurement was higher than that from the TOC measurement. Therefore, it can be concluded that CNTi nanocomposites undergo complete oxidation in contrast to that of the TiO 2 surface where intermediates were formed. Su    Stability evaluation. Photocatalytic measurements of the 0.1CNTi nanocomposite, which yielded the best degradation rate of phenol (in this study) in the presence of both O 3 and H 2 O 2 under Xe illumination, were repeated six times after isolating the photocatalyst. Figure 9 indicates that there was no obvious decline in the photocatalytic activity. In addition, XRD measurements both before and after the reaction was the same (the results are not shown). Therefore, the results demonstrated that the CNTi nanocomposite photocatalyst was stable under the abovementioned experimental conditions.

Conclusions
Coupling TiO 2 with C 3 N 4 (visible light-sensitized semiconductors) has proved to be beneficial for improving the photocatalytic performance due to a synergism that was ascribed to improved light harvesting, enhanced photostability and effective photoexcited charge separation. The addition of H 2 O 2 or O 3 improved the photocatalytic degradation rate of phenol. However, the rate dramatically increased when a trace amount of O 3 gas (4 ppm) was added to H 2 O 2 in the reaction mixture in the photocatalytic reactor, which was attributed to the increase in the rate of ozone degradation and, consequently, the rate of hydroxyl radical formation in the presence of H 2 O 2 .
The best C 3 N 4 ratio in the CNTi composite, which yielded the highest degradation rate of phenol, was 0.1% C 3 N 4 . This photocatalyst was stable under experimental conditions, and its performance was drastically enhanced in the presence of both O 3 and H 2 O 2 .
Finally, based on the results, the CNTi nanocomposite photocatalyst can be considered to be a promising candidate for the treatment of wastewater contaminated with phenol, especially when H 2 O 2 with traces of O 3 gas is used.