Photocatalytic decontamination of phenol and petrochemical wastewater through ZnO/TiO2 decorated on reduced graphene oxide nanocomposite: influential operating factors, mechanism, and electrical energy consumption

ZnO/TiO2 anchored on a reduced graphene oxide (rGO) ternary nanocomposite heterojunction was synthesized via the multi-step method including hydrothermal, solvothermal and sol–gel methods. XRD, Raman, FESEM, EDX, Dot Mapping EDS, BET, FTIR, UV-VIS, TGA, and EIS techniques were utilized for characterizing as-synthesized catalysts. The XRD and Raman data proved the formation of anatase phase TiO2 and wurtzite phase ZnO in the prepared samples. Further, the UV-Vis spectrum confirmed that the band gap value of ZnO/TiO2 diminished on introduction of graphene oxide. Photocatalytic performance of the fabricated catalysts was investigated by decontamination of phenol in aqueous solutions. The effect of different operational factors such as pH, catalyst dosage, phenol concentration, and light illumination was investigated to find the optimum decontamination conditions. According to the results, complete degradation of phenol was achieved at pH = 4, catalyst dosage of 0.6 g L−1, light intensity of 150 W, and phenol initial concentration of 60 ppm at 160 min under visible light illumination. With the addition of graphene oxide to the composite, a significant increase was detected in the photocatalytic performance due to the higher available surface area and lower electron/hole recombination rate. In addition, the scavenging experiments revealed that the ·OH is responsible for the degradation of phenol during the reaction. The degradation mechanism, economic performance, mineralization, and recyclability were also investigated. Kinetic studies confirmed that photocatalytic degradation process followed the pseudo-first-order kinetic model. A case of real wastewater treatment was used to examine the performance of the catalyst for real case studies.


Introduction
Nowadays, providing clean and healthy water resources has become one of the most important concerns of governments due to the population growth, considerable increase in water withdrawal, and production of large amounts of wastewater by different industries. Many industries produce water effluent containing organic compounds which are considered harmful environmental contaminants. The discharge of these contaminants makes changes to the ecosystem of aqueous systems by reducing the concentration of dissolved oxygen in natural channels. Petrochemical plants, petroleum reneries, and textile industries are the most important sources of phenol and heavy aromatic compounds. Phenol and phenolic compounds are bio-recalcitrant toxic organic compounds which are very harmful to the environment. Based on the European Union regulation no. 80/778/EC, the maximum allowable concentration of phenolic compounds is 0.5 mg L À1 in drinking water. Accordingly, wastewater containing a high concentration of aromatic and toxic contaminants should be effectively treated before being discharged into natural channels. [1][2][3][4][5][6] In particular, due to non-biodegradability of phenolic compounds in aqueous solutions, conventional biological methods are inefficient for effective treatment of these compounds. In recent years, advanced oxidation processes (AOPs) have changed into an alternative solution in the wastewater treatment plants for degradation of recalcitrant organic compounds thanks to their high degradation efficiency, low-cost, fast reaction rate, and non-selective degradation. AOPs are based on production of very reactive radicals involving ozonation, H 2 O 2 /O 3 combined with UV, semiconductor photocatalysis, Fenton, photo-Fenton, ultrasound irradiation and wet air oxidation. 3,[7][8][9] Amongst AOPs, heterogeneous semiconductor photocatalysis has been proven as one of the most appropriate and effective methods for degradation of recalcitrant organic contaminants in water thanks to its high performance, low material consumption, and cost-effectiveness. Recently, many semiconductors have been under usage as photocatalyst such as TiO 2 , ZnO, WO 3 , MgO, SiO 2 , Fe 2 O 3 , and CdS. [10][11][12][13][14][15][16] Semiconductor photocatalysts have recently demonstrated high photocatalytic activity for decontamination of organic compounds in aqueous solutions. Typically, titania has been revealed as a promising catalyst and has been widely used in photocatalytic degradation of organic contaminants due to its electrochemical properties, low cost, availability, high stability and non-toxicity. However, pristine TiO 2 has a fast electron/hole (e/h) recombination rate which restricts the photo-generated electron-holes lifetime. Further, these semiconductors have a large band gap and low efficiency in degradation of contaminants under visible light regions. Alternatively, several techniques have been chosen to improve the efficiency including metal/nonmetal doping, hetero-structures, multi-component composites using other semiconductors, architectural control, etc. Combination of wide bandgap semiconductors with narrow bandgap semiconductors and formation of type I heterojunctions can enhance the photocatalytic activity of titania. [17][18][19][20][21][22][23][24][25] The composite of TiO 2 and ZnO has shown to have a higher photocatalytic activity than pristine TiO 2 and/or ZnO because of type II hetero-junction formation which inhibits the e/h recombination rate. Furthermore, the heterogeneous structure formation by two semiconductors extends the photo response of composites to visible light area. [26][27][28][29][30] Co-adsorbents such as zeolites, carbonaceous materials, and nano-clays are another way for further enhancing the photocatalytic activity and adsorption capacitance of TiO 2 . Recently, graphene nanosheets have been used as an effective ESI † due to its favorable properties like solar radiation absorption, separation of efficient charge carriers, exposed reactive sites, and high surface area. Graphene has superior properties such as optical transparency, high surface area, and high electrical conductivity along with other physiochemical properties. It was shown that this support can enhance the degradation efficiency through making a hybrid by semiconductor metal oxides. Many heterogeneous nanostructure photocatalysts coupled with graphene nanosheets have been synthesized and developed in recent years. 17,19,26,[31][32][33][34] Photocatalytic process efficiency is signicantly affected by electron-hole recombination rate which could be controlled by various parameters such as interfacial properties, electronic properties, chemical composition, and physical dimension. 23 Various statements have been suggested about the characteristics of graphene in hetero-structure nanocomposites which included graphene nanosheets in their frameworks: (I) formation of Zn-O-C, Zn-C, and Ti-O-C bonds by graphene defect states improves the solar or visible light adsorption capacity of catalysts; (II) graphene can accept or donate electrons due to its high electron mobility in order to reduce the e/h recombination rate of ZnO/ TiO 2 nanocomposite, thus prolonging the life-time of excited electrons during the reaction; (III) graphene enhances the adsorption capacity of nanocomposites and creates more reactive sites for sorption of target contaminants because of its high BET specic surface area. 23,35,36 However, graphene nanosheets are hydrophobic, making it challenging to disperse graphene in aqueous solutions for preparing homogenous solutions. Homogeneous dispersion of graphene oxide (GO) in aqueous mediums is easier than that of graphene due to its functional groups such as -OH and -COOH. Therefore, GO nanosheets are commonly used in the synthesis of photocatalytic nanocomposites. Thus, we expected that the introduction of rGO in the TiO 2 /ZnO framework would resolve the TiO 2 /ZnO restrictions and considerably enhance the photocatalytic activity of TiO 2 /ZnO under visible light illumination. To the best of our knowledge, many studies have explored enhanced photocatalytic activity of titania via ZnO and/or graphene. Shahbazi et al. used TiO 2 nanoparticles over graphene nanosheets synthesized in hydrothermal process as a photocatalyst for degradation of phenol in aqueous media under UV irradiation. 3 17 Malekshoar et al. used TiO 2 /graphene, ZnO/graphene and their physical mixture for degradation of phenol with sun light irradiation. 26 Liu et al. synthesized ZnO/TiO 2 /rGo nanocomposite via microwave-assisted method and evaluated its photocatalytic performance by reduction of Cr(VI). 27 Due to low quantum yield of bare TiO 2 and its wide band gap value, ZnO as a coupling semiconductor material is able to suppress the electron/hole recombination rate of TiO 2 and improve its photocatalytic activity. Besides, reduced graphene oxide was used for surface modication of titania and improving its adsorption as well as charge conductivity performance. To the best of our knowledge, there are limited studies which used couple of modications for degradation of phenol and real petrochemical wastewater under visible light irradiation. In this study, catalysts were synthesized via multi-step method involving hydrothermal, solvothermal and sol-gel methods. The prepared composite was characterized and studied for phenol photocatalytic degradation under visible light illumination in aqueous solutions. The effect of different operational parameters including pH, phenol concentration, catalyst dosage, and light illumination were examined. In addition, stability of fabricated nanocomposite during reactions, mineralization rate, economic estimation of the process, possible degradation mechanism, and trapping studies were also investigated. In the end, a petrochemical wastewater was treated via the fabricated catalyst.
as the precursor of TiO 2 . Also, zinc nitrate hexahydrate (ZN) was employed as the precursor of ZnO. Other chemicals were purchased from Merck (Germany). All materials were utilized without further purication because of their analytical grade.

Real wastewater sampling
The real petrochemical wastewater was collected from wastewater effluent of Maroun Petrochemical Company, Mahshahr, Iran. The collected samples were stored in 10 L containers at 4 C. The physiochemical features of the collected wastewater such as COD, TDS, BOD, TSS, and pH were determined based on specic standard techniques (ISO 5667-10:1992).

Preparation of TiO 2
For this purpose, 10 mL of TB was added to 40 mL of absolute ethanol (AE) and sonicated for 30 min (mixture 1). Another solution containing 10 mL ethanol, 80 mL deionized water (DI) and 3 mL HNO 3 was added to mixture 1 dropwise under constant stirring. Aer 1 h of stirring, the homogeneous solution was heated at 80 C in a hot water bath for 30 min. Then, the hot solution was allowed to cool and form the gel. Next, the gel remained still for 24 h, which was then dried in an oven at 100 C for 8 h. At the end, the ne powder of dried gel was annealed at 400 C for 4 h to achieve the anatase phase of TiO 2 (Ti). 33,37

Preparation of ZnO
To this aim, 10 gr ZN was dissolved in 250 mL DI water and sonicated for 2 h. The prepared solution was heated at 90 C under constant stirring for another 5 h. Next, the obtained mixture was dried at 100 C overnight. Finally, the obtained ne powder was annealed at 600 C for 2 h to achieve the wurtzite phase of ZnO. 38 2.5 Preparation of rGO/ZnO/TiO 2 rGO/ZnO/TiO 2 (GZnTi) was synthesized through facile hydrothermal method. A suitable amount of GO (0.05, 0.075 and 0.1 gr GO for 5, 7.5 and 10 wt%, respectively) was dissolved in a mixture containing 20 mL DI water and 40 mL AE and sonicated for 1 h. Then, 0.1 g of ZnO and 0.9 g TiO 2 (10% ZnO/90% TiO 2 ) were added to the prepared mixture and sonicated for another 2 h to attain a homogeneous suspension. The attained suspension was transferred to 100 mL Teon sealed hydrothermal autoclave and maintained at 120 C for 3 h to develop the nanocomposite of rGO, ZnO and TiO 2 . Finally, the achieved composite was ltered and washed with ethanol several times and dried at room temperature. ZnO/TiO 2 (ZnTi) nanocomposite was synthesized by the mentioned hydrothermal method without adding GO. 39
The X-Ray Diffraction (XRD) prole of the synthesized samples was recorded by X-Ray diffractometer (Quantachrome, NOVA 2000) through graphite monochromatic CuK a irradiation (0.15406 nm wave length) within the range of 10-80 with 0.05 step size. The eld emission scanning electron microscope (FESEM), Energy Dispersive X-ray (EDX) and Dot-Mapping of samples were recorded using Mira-3-Taksan operating at 15.0 kV. Brunauer-Emmett-Teller (BET) specic surface areas of the samples were determined by N 2 -adsorption-desorption isotherms at 77 K using Beckman-Coulter 3100 instrument. Barrett-Joyner-Halenda (BJH) model was used for determining the pore mean diameter using N 2 -adsorption measurement with 0.99 bar relative pressure. UV-Vis Diffuse reectance spectroscopy (UV-VIS DRS) of specimens was examined using a SHI-MADZU spectrometer model UV-1240 (ranging from 200-900 nm with BaSO 4 as reference). The electrochemical impedance spectroscopy (EIS) of the synthesized catalysts was investigated via potential-galvanostat PGSTAT302N instrument in a three-electrode-cell using saturated calomel electrode and platinum foil as the reference electrode and counter electrode, respectively. AC voltage amplitude, frequency range, and electrolyte were 10 mV, 10-100 kHz, and NaCl 3.5 wt%, respectively. The total organic carbon (TOC) of the degraded solution was studied via the DC-190 of Rosemount-Dohrmann analyzer. Thermogravimetric analysis (TGA) of the as-synthesized nanostructures was performed by a Q-600 TA-apparatus under air ow at a warming rating of 5 C min À1 ranging from ambient temperature to 800 C.

Photocatalytic activity
To prevent any ion interference during the reaction, DI water was used for preparing all solutions. In order to evaluate the photocatalytic activity of as-synthesized photocatalysts, phenol degradation (Ph-degradation) in aqueous medium was used as the objective contaminant. In the reaction process, three visible 270.7 Cd lamps equipped with l > 420 nm lter was utilized as the visible light source, and the experiments were performed at the ambient temperature. First, a 250 mL solution was prepared with a certain amount of phenol. Then, a specic amount of assynthesized catalyst was added to the phenol solution. With addition of NaOH 0.5 M or HCl 0.5 M L À1 , the pH of the solution was adjusted. The pH of the media was determined using S210-Std-Kit Mettler-Toledo pH-meter under 250 rpm stirring. To achieve adsorption-desorption equilibrium, the mixture was loaded on to a 250 mL reactor and stirred (at 250 rpm and 27 AE 1 C) for 60 min in the dark. The photocatalytic reaction was initiated by exposing the prepared suspension to visible light irradiation.
At specic time periods, 5 mL of the solution was sampled to determine the phenol concentration. PIT320 Universal centrifuge apparatus was used to remove the photocatalysts from the aqueous solution. The phenol concentration of the sample was calculated by UV2100 Unico UV-visible spectrophotometer to record the specic absorption at l max ¼ 270 nm (maximum absorption wavelength). 3 The removal and mineralization efficiencies of the system were calculated using the following equations: where, C 0 and C t represent the concentration of phenol at the initial time and sampling time. T 0 and T t denote TOC of the solution at the initial time and sampling time, respectively. The adsorption capacity of the system was obtained through the following equation: where C t0 was initial phenol concentration (mg L À1 ), C e is the concentration of phenol at the adsorption equilibrium state (mg L À1 ), V is volume of the mixture (L) and M c is the weight of GZnTi (g).

Results and discussion
3.1 Characterization of photocatalysts 3.1.1 BET. Fig. 1A and B illustrate nitrogen adsorptiondesorption isotherms and BJH patterns of Ti, ZnTi and GZnTi samples. According to IUPAC classication, N 2 adsorption/ desorption isotherms belong to type IV with H 4 type hysteresis loops at P/P 0 ranging around 0.03-0.99, suggesting the presence of abundant mesoporous structures in all samples. This isotherm is used for porous materials and oen seen for industrial catalysts. BET surface areas of Ti, ZnTi and GZnTi were found to be 76.805, 66.112 and 149.95 m 2 g À1 , respectively. The obtained results indicated elevated BET surface area of GZnTi, compared with ZnTi, which is an effect of graphene addition. This phenomenon revealed that graphene is an efficient support for ZnTi photocatalyst thanks to enhancing the adsorption capacity of the catalyst. From BJH patterns, it is clear that the pore diameter distribution of catalysts was around 5 nm, representing the mesoporous structure of the samples. 1 3.1.2 XRD. XRD is a well-known and powerful method for structural analysis of nanocomposites, and has been used by researchers since 1950s. Fig. 2A displays the XRD patterns of Ti, ZnTi and GZnTi. As can be seen in Fig (1 1 2) respectively, according to JCPDS . Further, the ZnTi pattern reveals that the addition of ZnO does not make any changes in the TiO 2 crystallite structure, and its remain in anatase phase. The major phase of anatase and wurtzite can be seen from the GZnTi pattern, revealing the presence of TiO 2 and ZnO in the composite, respectively. This nding suggests that GO preserved the crystalline structure of the composite. The characteristic peak observed at 2q ¼ 11.04 of GO pattern matched (0 0 1) facet of GO. This peak could not be seen in the GZnTi pattern because GO was reduced during the thermal process. 40,41 However, it is difficult to recognize the characteristic peak of graphene in the GZnTi prole. The reason can be related to the fact that the characteristic peak of graphene located at 2q ¼ 25.1 overlapped with the characteristic peak of TiO 2 at 2q ¼ 25.35 (see Fig. 2A). Furthermore, the average grain sizes of as-prepared samples were estimated by the Debye-Scherrer's equation. The crystalline size of TiO 2 for Ti, ZnTi and GZnTi samples was found to be 8.59, 6.68, and 6.31. The obtained results reveal that the addition of GO into the ZnTi composite considerably hindered crystalline growth.
3.1.3 Raman. Fig. 2B  , and 2LO wurtzite active-modes of ZnO, respectively. The obtained results are in good agreement with XRD reports. Furthermore, compared with the Raman prole of Ti, ZnTi and GZnTi patterns shied to higher wavenumbers. This could be ascribed to distortion of titania due to incorporation of ZnO or GO. Raman shis of GO and GZnTi indicate new peaks at 1354 and 1576, matching D-band and G-band of carbon materials contained in the mentioned samples. Meanwhile, I D-band /I G-band ratio is an efficient tool for determining the functionalization degree of samples and defects in carbonaceous materials. The I D-band /I G-band ratio of GZnTi increased to 1.175 compared to that of GO which was found to be 0.858. This suggests successful reduction of GO to rGO during the thermal process, which is in excellent accordance with previous reports.    Fig. 4 represents the UV-visible adsorption spectra of Ti, ZnTi, and GZnTi within the range of 200-900 nm. As can be observed, all of the samples have had intense adsorption in UV-light area (200-400 nm), suggesting their activation within this range. Notably, the adsorption intensity of GZnTi in visible-light region (400-700 nm) has been considerably higher than that of Ti and ZnTi. The inset of Fig. 4 shows the colors of photocatalysts. According to this gure, incorporation of GO into the ZnTi host changes the color of the catalyst from light (white for Ti and ZnTi) to dark (gray for GZnTi). Thus, darker catalyst color represents more intense adsorption. A slight positive shi of adsorption edge was caused by the interaction of ZnO with the titania framework. Furthermore, the adsorption edge of GZnTi nanocomposite had a red shi toward higher wavenumbers for 27 nm than that of ZnTi sample, since GO layers have successfully interacted with ZnTi nanoparticles. Hence, GO and ZnO are able to enhance the light adsorption intensity and optical response of titania across visible regions, which are favorable for the photocatalytic efficiency of systems on industrial scales. The band gap values of specimens were determined using Planck equation as follows: where, l is the cut-off wavelength, c represents the speed of light (2.99 Â 10 8 m s À1 ), and h denotes the Planck constant (4.13566 Â 10 À15 ). The band gap calculation details of the samples are listed in Table 1. As can be observed, the band gap value of GZnTi was lower than that of Ti and ZnTi. The obtained results can be attributed to the characteristic light adsorption potential of GO contained in the nanocomposite and to high rate of charge migrations between the semiconductors (titania and ZnO) and rGO layers. 43 3.1.6 FESEM. Fig. 5(a-d) illustrates the FESEM micrographs of Ti, ZnTi, GZnTi and GO with corresponding EDX tests. As can be seen in Fig. 5a and b, neat TiO 2 nanocomposite appeared with uniform sphere-shaped morphology and a trace content of rod-shaped ZnO was detected in ZnTi sample. In this regard, similar morphologies has been obtained in the literature. This indicates that introduction of ZnO preserved the initial morphology of titania. Fig. 5c and d demonstrate the FESEM micrographs of GZnTi and GO, respectively. The at plane-like morphology of GO has been presented in Fig. 5c and d. Also, the interaction of titania and ZnO with GO can clearly be observed in Fig. 5c. Furthermore, the EDX proles of specimens revealed major peaks of Ti, O, Zn, and C, conrming their high purity. According to the obtained results listed in Table 2, the theoretical weight composition of elements are a little different with the EDX results for each synthesized sample. The reason is     ascribed to evaporation of moisture adsorbed on the surface of samples. The order of the primary decomposition ratio of the samples was found to be GZnTi > Ti > ZnTi, which is in excellent accordance with BET results, where S BET of GZnTi was greater than that of Ti and ZnTi. Thus, a higher moisture content could be adsorbed on the surface of catalysts. 1,3 Notably, in the GZnTi prole, a secondary decomposition step was initiated from 200 C to 800 C, associated with combustion or decomposition of rGO contained in the GZnTi catalyst. The results indicate that the real weight ratio of rGO in the GZnTi nanocomposite is almost the same as the theoretical weight ratio of rGO before synthesis, conrming the accuracy of the synthesis process.

EIS.
The electrochemical properties of the samples were explored by powerful EIS spectra. Fig. 7B exhibits the EIS Nyquist diagrams of Ti, ZnTi, and GZnTi. EIS Nyquist represents three types of resistance; (i) contact resistance, (ii) intrinsic material resistance, and (iii) electrolyte resistance. The radius of the semicircle of Nyquist plots is related to the interfacial charge migration, which depends on the charge separation efficiencies of samples. As can be seen from Fig. 7B, the order of radii of semi-circles was found to be Ti < ZnTi < GZnTi, indicating that introduction of ZnO and (or) GO enhanced the electron migration rate of the sample, thereby improving the electron/hole separation rate of the GZnTi photocatalyst. 1 Accordingly, the obtained EIS results suggest that the GZnTi photocatalyst is desirable for photocatalytic degradation reactions.

Photocatalyst performance
3.2.1 Effect of GO content. The photocatalytic performance of GZnTi, ZnTi, and Ti composites was investigated to nd the optimum concentration of GO in catalysts for the maximum Phdegradation efficiency in the presence of visible-light irradiation at pH of 5, catalyst dosage of 0.8 g L À1 , phenol concentration of 60 ppm. Ph-degradation rate of Ti was 40%, indicating low efficiency of pristine titania for decontamination of phenol under visible irradiation due to its high e/h recombination rate and wide band gap value. Also, with introduction of ZnO into TiO 2 , the Ph-degradation efficiency increased from 40% to 56%. This increase could be ascribed to the positive effect of ZnO on photo-excited electron trapping. In contrast to ZnTi, the Phdegradation rate of all GZnTi nanocomposites grew, indicating that addition of GO considerably enhanced the photocatalytic performance of catalysts under visible light irradiation. Apparently, for the GZnTi nanocomposites, by increasing the weight percentage of GO from 5 to 7.5 wt%, the photocatalytic activity rose from 40% to 86.4%. However, by raising the GO concentration further to more than 7.5 wt%, the removal rate diminished to 73.5%. Mechanistically, GO can improve the adsorption capacity and reduce electron-hole-recombinationratio, thereby enhancing the degradation efficiency. 23,45 This effect was clearly observed in the case of GZnTi 5 wt% and 7.5 wt%, where the Ph-degradation efficiency increased considerably from 40% to 86.4%. Nevertheless, higher GO concentrations can scatter the light irradiations and reduce the available light adsorption surface of TiO 2 and ZnO NPs, and compromise the photocatalytic performance. 45 Accordingly, at high weight percentages of GO, TiO 2 , and ZnO, NPs aggregate on the surface of GO, thereby lessening the adsorption capacity of GO sheets and light adsorption of TiO 2 and ZnO NPs. This, in turn, leads to lower photocatalytic performances. Accordingly, the GO concentration 7.5 wt% is an optimum concentration of GO in the GZnTi composite.

Effect of pH.
Based on studies, the initial pH value of a solution is one of the important operational parameters which can inuence the degradation rate of organic compounds during heterogeneous photocatalyst degradation. pH has a signicant inuence on the contaminant hydrolysis, catalysts' surface properties, oxidant, and contaminant ionization degree. The experiments were conducted at pH of 4-10, catalyst dosage of 0.8 g L À1 , irradiation time of 160 min and phenol concentration of 60 ppm. As can be seen in Fig. 8B, the Ph-degradation in the acidic solution is higher than in neutral solution and alkaline solution, which dropped from 91% to 51.9% with pH rise from 4 to 10. These differences between degradation efficiencies at various pH ranges could be described by two reasons: (I) the oxidation ability of hydroxyl radicals is greater under acidic conditions; however, presence of HCO 3 À and CO 3 2À under alkaline conditions interferes in the reaction of hydroxyl radicals, thus reducing its oxidation potential. 46 This validates that the Ph-degradation by photoinduced radicals is higher under acidic solutions than in alkaline solutions. (II) The low adsorption of negatively charged system components and interference of HCO 3 À and CO 3 2À (i.e. at pH value of 10) are the results of very low Phdegradation rates. Therefore, effective Ph-degradation can occur within the range of 4-6.5 (lower than pH ZPC in the acidic area). The results of degradation experiments revealed that the maximum Ph-degradation was 91% at pH ¼ 4.

3.2.3
Effect of GZnTi dosage. The amount of photocatalyst has a signicant effect on the Ph-degradation efficiency. The effect of GZnTi photocatalyst was studied by varying the catalyst dosage. The experiments were performed at initial pH of 4, catalyst dosages of 0.2-1 g L À1 , irradiation time of 160 min and phenol concentration of 60 ppm. The degradation rate of phenol for different catalyst dosages is shown in Fig. 8C. By raising the photocatalyst dosage from 0.2 g L À1 to 0.6 g L À1 , more active sites were available and the specic surface area increased, resulting in enhanced Ph-degradation rate from 58.2% to 100%. Also, the superoxide and hydroxyl radical generation rate grew within this range. However, with addition of extra contents of the catalyst over 0.6 g L À1 , the Phdegradation rate declined from 100% to 72.4%. This could be resulted by: (I) lower solution transparency, light scattering and interception results due to excessive dosage of photocatalyst and prevention from light induction of some particles; (II) diminished pore volume and available surface area of the catalyst; (III) prevention of collisions between phenol and hydroxyl radicals; and (IV) excessive hydroxyl generation with higher doses of the catalyst. This recombines formation of H 2 O 2 or reaction with O 3 for production of HO 2 radicals, which are less active than hydroxyl radicals. 7,19,47 3.2.4 Effect of phenol concentration. The effect of phenol concentration was studied at different phenol concentrations of 60-90 ppm, initial pH of 4, catalyst dosage of 0.6 g L À1 , irradiation time of 160 min. Fig. 8D displays the obtained results for different phenol concentrations. As can be seen, the Phdegradation efficiency is 76% at the concentration of 90 ppm, which signicantly increases as the phenol concentration drops from 90 to 60 ppm. This can be ascribed to the fact that elevation of phenol concentration and its intermediate by-products deactivates more active sites and reduces the light penetration to active sites situated on the surfaces of catalysts, which are responsible for producing oxidative radicals. Also, at high phenol concentrations, a major portion of light illumination Herein, to a better assess, the catalytic activity of the synthesized GZnTi was compared with the detoxication of phenol reported in previous studies (see Table 3). GZnTi showed more photocatalytic performance because it indicated higher amounts of decontamination in less time under visible light than other catalysts. Therefore, it was concluded that the GZnTi nanocomposite was an efficient catalyst for decontamination under selected experimental parameters. It may be benecial to use this nanocomposite for commercial applications due to its slight cost and great performance.

Effect of light distance.
Position of light from the mixture plays a critical role to enhance the incident illumination intensity on the surface of the photoreactor. The effect of light distance on Ph-degradation was investigated within the range of 5-25 cm at pH of 4, catalyst dose of 0.6 g L À1 , irradiation time of 160 min, and phenol concentration of 60 ppm. The results presented in Fig. 8E reveal that by increasing the light intensity from 5 to 15 cm, the Ph-degradation did not considerably change. This nding suggests that in the light distance range of 5-15 cm, sufficient content of photons delivered from the lamp to the photocatalyst surfaces for effective decontamination of organics in the process. However, at distances higher than 15 cm, there was a decrease in Phdegradation rate which indicates the limitation in the light distance from the mixture in photocatalytic systems that in accordance with the ref. 58.
3.2.6 Recyclability. In real application of heterogeneous photocatalysts in wastewater treatment plants, stability and recyclability of catalyst are absolutely important. The stability Fig. 9 Recyclability of GZnTi in Ph-degradation process in the presence of visible light illumination through 6 cycles at pH: 4, catalyst dosage: 0.6 g L À1 , and phenol concentration: 60 ppm (at the optimum Ph-degradation conditions). and recyclability of GZnTi were examined in a batch reactor under the optimum Ph-degradation condition (pH of 4, catalyst dosage of 0.6 g L À1 , irradiation time of 160 min, and phenol concentration of 60 ppm). Aer each experiment, the catalyst was withdrawn from the solution and washed with water and ethanol mixture (2 : 1 volume ratio) to remove the phenol and its intermediate products on their pores and surfaces. Then, the mixture was dried at 80 C for 2 h. This sequence was repeated for six times to nd the Ph-degradation efficiency of each cycle. The results demonstrated in Fig. 9 indicate that the Phdegradation efficiency was maintained within the range of 90-100% aer six cycles, and showed good stability and recyclability for GZnTi. The photocatalytic activity reduction of GZnTi from 100% to 90.2% aer six cycles could be explained by loss of catalyst during the washing process, shrinkage of the pore volume, and reduction of the BET specic area. 19 3.2.7 Trapping/radical scavengers. It has been proven that different reactive species including hydroxyl radical, hole, and superoxide radical are responsible for the oxidation of organic contaminants in aqueous solutions in photocatalytic degradation systems. The scavenger effect was studied under visible light irradiation, at pH of 4, catalyst dosage of 0.6 g L À1 , irradiation time of 160 min, and phenol concentration of 60 ppm. For this aim, 10 mL of IPA (isopropyl-alcohol), BQ (p-benzoquinone), and EDTA 0.01 M (disodium ethylene diamine tetra acetic acid) was separately added to the photo-reactor as scavengers of hydroxyl radicals ($OH), superoxide radicals ($O 2 À ), and holes (h + ), respectively. 2,59 The Ph-degradation efficiency for different radical scavengers is exhibited in Fig. 10A. As can be seen, the Ph-degradation rates in the presence of $OH, $O 2 À and h + scavengers were found to be 24.75, 53.82, and 73.65, respectively. Accordingly, the dominancy order of species was: $OH>$O 2 À > h + , indicating that Ph-degradation process was mainly controlled by hydroxyl radicals.

Mineralization.
The mineralization rate of phenol was investigated through comparing TOC and Ph-degradation efficiency for different times under optimal conditions including pH of 4, catalyst dosage of 0.6 g L À1 , and phenol concentration of 60 ppm.
The experimental results reveal that TOC removal rate rises to 100% aer 360 min, while the Ph-degradation rate reaches 100% aer 160 min (Fig. 10B). The difference between TOC and Ph-degradation can be ascribed to decomposition of phenol molecules to the intermediate by-products.
3.2.9 Degradation kinetic. The pseudo-rst order kinetic parameters of Ph-degradation kinetics were calculated by plotting ln(C 0 /C) vs. time for phenol concentrations of 60, 75 and 90 ppm under pH of 4, catalyst dosage of 0.6 g L À1 , and irradiation time of 100 min. The rate constant and correlation coefficient were determined for different initial phenol concentrations and listed in Table 4. Based on the results, it can be established that the Ph-degradation kinetics at 60 ppm was best tted by the pseudo-rst-order reaction with 0.9859 regression coefficient and 0.0124 L min À1 rate constant (Fig. 11). 4 3.2.10 Estimating electrical energy consumption. According to the kinetic results, decontamination rate of phenol versus time at different phenol concentrations matched pseudo-rstorder kinetic model. On the other hand, heterogeneous photocatalysis technique is electricity intensive, so electric energy cost is one of the main portions of total costs. Thus, determination of simple gures of merit, which is based on electrical energy consumption, could be helpful and benecial. The photochemistry commission of IUPAC for low-content contaminants and rst-order reaction kinetic models suggested gures of merit by an equation for AOPs. According to this statement, the suitable gure of merit is the electrical energy per order (EEO), and introduced as the number of kW h of electrical energy necessary for reducing a contaminant content by one-order of magnitude (almost 90%) in 1000 L of contaminated water. The EEO kW h m 3 Â order of the system used in this investigation could be estimated from the following equation: 60 where, P is the rated power (kW) of the photocatalysis process, V represents the volume (L) of the solution contained in the photo-reactor, and k denotes pseudo-rst order rate constant 1 min of Ph-degradation kinetic model. EEOs of Based on previous reports, the possible mechanism for photocatalytic Ph-degradation via GZnTi nanocomposite could be suggested as Fig. 12. As the nanocomposite is exposed to the visible light illumination, electrons in the valence band of titania migrate to the conduction band of ZnO. This process improves when rGO is used as a ESI. † Indeed, graphene acts as an excellent adsorbent for organic contaminants, sink for photo-excited electrons, and electron transporter to the composite surface. In this way, the effective roles of rGO nanosheets result in reduction of e/h recombination rate, thus enhancing photocatalytic performance of the nanocomposite. 38 The created electrons and holes are trapped by O 2 and H 2 O molecules, generating highly active peroxide and hydroxyl radicals, respectively. Therefore, the generated radicals attack the organic pollutants such as phenol molecules and oxidize them to intermediate by-products, which are nally mineralized to CO 2 and H 2 O. 38 3.2.12 Real wastewater treatment. In order to evaluate the effectiveness of GZnTi nanocomposite in practical industrial uses, real petrochemical wastewater was utilized for treatment via GZnTi at pH ¼ 4 and catalyst dosage of 0.6 g L À1 under visible light irradiation. The properties of the collected wastewater are listed in Table 5. The TOC and COD removal ratio of the wastewater as a function of time are presented in Fig. 13A and B. As can be seen, the TOC and COD removal ratio grew by prolonging the reaction time, which were found to be 85.9% and 92.1%, respectively, aer 440 min. As a possible explanation, with prolongation of reaction time, more photons are adsorbed by the nanocomposite, leading to production of more oxidative radicals which are responsible for reducing TOC and COD of the wastewater during the process. To the best of our knowledge, some studies have reported similar ndings for treating industrial wastewaters using AOPs. 4,61-63 3.2.13 Adsorption studies. Adsorption has a key role in removal of organics in the photocatalytic systems because it determines the sorption mechanisms and also, reveals the different parameters impacting on the degradation rate. Interestingly, batch adsorption kinetics reports are useful for designing the industrial columns and scale-up applications. Noting that the experiments were performed at dark conditions, 25.6 C, pH of 4, catalyst loading of 0.6 g L À1 and phenol concentration of 30-50 ppm. As could be seen in Fig. S1, † the adsorption rate of phenol (30 ppm) on GZnTi was very quick in the rst stage of experiment (t < 30 min) which indicated the happening of a physical adsorption process. Finally, the adsorption process reached to the equilibrium adsorption state without any further phenol adsorption aer 60 min which revealed the presence of a large number of reactive sites on the surface of the adsorbent saturated by phenol molecules during the process (Fig. S1 †). In order to investigate the adsorption character and its performance, several experiments were carried out for kinetics and isotherms of phenol on the GZnTi photocatalyst. Three well-known kinetic models including intraparticle diffusion, pseudo rst-order (PFO) and pseudo second-order (PSO) models and two isotherm models including Langmuir and Freundlich were considered for evaluation of experimental results. ESI † and constants of the aforementioned kinetics and isotherms as well as the corresponding regression coefficients (R 2 ) are represented in Table 6. The correlation coefficient of intra-particle diffusion and PFO models were 0.9454 and 0.9121, respectively; whereas, the correlation coef-cient of PSO model was expressed 0.9811. In other words, the obtained results from the kinetic studies revealed that the PSO model better describes the behavior of phenol adsorption on the GZnTi. It is noteworthy that PSO model demonstrates the fact that the chemisorption process is the dominant step in the adsorption experiment which occurred by donating or exchanging of electrons between phenol molecules and GZnTi binding sites. Furthermore, as given in Table 6, the R 2 value for Langmuir model was greater than that of Freundlich model which indicates that the phenol adsorption followed Langmuir isotherm model. The maximum mono-layer adsorption capacity

Conclusion
In this investigation, visible light active GZnTi ternary nanocomposites were successfully prepared with different weight ratios of GO. They were conrmed via XRD, Raman, FT-IR, BET, UV-Vis DRS, EIS, TGA, FESEM, EDX, and EDS analysis. According to the obtained ndings, GO in the GZnTi nanocomposite was effectively reduced through the synthesis process. Furthermore, presence of ZnO nanorods was detected via FESEM micrographs. Compared to pure titania, within corporation of GO and ZnO, the band gap value diminished, BET surface area increased, and the recombination rate of charges dropped. The catalytic performance of as-prepared ternary nanocomposites was evaluated by Ph-degradation under visible light irradiation. According to the experiments,  the GZnTi nanocomposite demonstrated complete degradation of phenol when the GO content was 7.5 wt%, pH value was 4, photocatalyst dosage was 0.6 g L À1 , phenol concentration was 60 ppm, light intensity was 150 W, and reaction time was 160 min. Furthermore, based on the electrical energy consumption study, 3.81832, 5.9184, and 10.0739 euro/m 3 were required for Ph-degradation with a concentration of 60, 75, and 90 ppm, respectively. In addition, quenching experiments revealed that the photocatalytic process via GZnTi catalyst was controlled by highly reactive hydroxyl radicals ($OH). In the end, real petrochemical effluent was effectively treated under visible light irradiation and optimal operational conditions. The TOC t / TOC 0 and COD t /COD 0 removal ratio was reduced to 14% and 8% aer 440 min. Therefore, the obtained results suggest that GZnTi has excellent photocatalytic abilities under visible light for decontamination of non-biodegradable compounds and real effluents.

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