Photocatalytic degradation of methylene blue dye by fluorite type Fe2Zr2-XWXO7 system under visible light irradiation
Abstract
Nano-sized Fe2Zr2-XWXO7 system was prepared using the Pacini method where x = 0, 0.05, 0.1 and 0.15. All the samples were characterized using chemical analysis, X-ray diffraction (XRD), Fourier-transform infrared (FT-IR), transmission electron microscopy (TEM), UV–vis diffuse reflectance measurements (DRS) and surface area measurements. The undoped Fe2Zr2O7 was crystallised in the cubic fluorite phase as a major phase in addition to rhombohedral phase of Fe2O3 and monoclinic phase of ZrO2 as the minor phases. Meanwhile, single cubic fluorite phase was defined for Fe2Zr0.85W0.15O7 sample. The last has the lowest band gap (1.69 eV) and the highest surface area (106 m2/g). From TEM, the average particle size of the prepared samples was in the range of (3–7 nm). The photocatalytic efficiency of the prepared Fe2Zr2-XWXO7 system was manifested by the degradation of methylene blue and real textile wastewater of blue colour. Ascending degradation efficiency was exhibited with increasing tungsten concentration which is in accordance with their band gap as well as their surface area values. The degradation rate using Fe2Zr0.85W0.15O7 sample obeys the pseudo-first order kinetic at the optimum degradation conditions (1.5 g/L catalyst and pH11). Fe2Zr0.85W0.15O7 showed promising photocatalytic activity for real textile wastewater where the 69% COD removal was obtained under the same conditions used for methylene blue degradation.
1. Introduction
The value of water resources is universally recognized and the water shortage problem is increasing in many countries. The need to preserve water resources is the driving force behind the identification and ex- ploitation of non-conventional water sources. For industry, in particular textile production, wastewater reclamation appears to be a technically feasible solution especially for water reuse (Haddad, 2011).
The textile industry consumes a huge amount of water in its pro- cessing stages. The matriX of industrial textile wastewater includes various pollutants such as inorganic compounds, dyes, colour residues and cleaning solvents which lead to many environmental issues. All over the world, an output of more than one million tons of dyes are fabricated per year resulting 0.28 million tons wastewater (Anjum et al., 2017).
Methylene blue (MB) is one of the dyes that is extensively used for dying silk, cotton and wood. The existence of this dye in wastewater has several health problems such as nausea, vomiting, diarrhea and eye burns (Salehi et al., 2012). Thus, it is obligatory to be decomposed before dumping in the rivers or lakes.
Textile wastewater treatment includes physical treatment (such as Flotation and Adsorption) chemical treatment (such as H2O2 and ozo- nation) as well as biological treatment. Each process has its merits and infirmities. Thus, the choice of a suitable process that gives the op- timum outcome with less toXic by-products and low cost is a hard as- signment.
Recently, the Advanced OXidation Processes (AOP) gained special consideration because of its ability to get rid of the textile effluents wastes. In AOP, the catalyst leads to the production of free radicals which are powerful oXidizing agent of the organic materials that can be mineralized to carbon dioXide, water and inorganic ions (Wang and Xu, 2012). Heterogeneous AOP for wastewater treatment using nanoma- terials gained considerable attention. As the photocatalysts particle size decreases, the surface area increases leading an increment of light absorption efficiency and the probability of holes and electrons move- ment to the surface. Therefore, photocatalyst with nano-sized particles is better than the micrometer sized ones in promoting the photo- catalytic reaction. Several nanomaterials are used for wastewater treatment such as semiconductors, nano-clay, nanocatalyst, na- noclusters, nanorods, nanocomposites. Some of the nanomaterials re- ported for treatment of wastewater are Pd/Fe3O4 (Hildebrand et al., 2009), TiO2 (Tangestaninejad et al., 2008) and magnetic chitosan (Geng et al., 2009).
A complex oXide with A2B2O7 composition has either a pyrochlore- type structure or a defect fluorite-type structure. Simply, in a pyro- chlore-type structure, the A-site is occupied by larger, trivalent and 8- coordinated cations, typically rare earth elements. The B-site is occu- pied by smaller, tetravalent and 6-coordinated cations, usually transi- tion metals, such as Zr, Ti and Hf (Moriga et al., 1990; Subramanian et al., 1983). For defected fluorites or disordered pyrochlores there is a lack of 1/8 oXygen ions in relation to MO2 fluorite. The crystallization of A2B2O7 pyrochlore or fluorite phases structure depends on the radius ratio of A and B cations (rA/rB) as well as the conditions of samples processing (Catchen and Rearick, 1995; Mandal et al., 2007, 2006). A2B2O7 crystallizes in the stable pyrochlore structure when the (rA/rB) is in the range of 1.46–1.78 (Whittle et al., 2009). The defect fluorite structure is detected for the lower or upper limits of the above-men- tioned range of (rA/rB).
The complex oXide with A2B2O7 composition has appealing properties, such as high melting point, high thermal expansion coefficient, low thermal conductivity, high thermal stability, high stability of ra- diation and high electrical conductivity (Liu et al., 2010). Pyrochlore oXides are used as ionic and electronic conductors in the Solid OXide Fuel Cells (Martínez-Coronado et al., 2011; Wuensch et al., 2000) as well as the host matriXes for the immobilization of the nuclear waste rich in the actinides (Ewing et al., 2004; Weber and Ewing, 2000). Pyrochlore based Lanthanide’s zirconate could be used as the inert matriX fuel materials (Moudir et al., 2013; Zhang et al., 2008).
Recently, Nd2Zr2O7–Nd2O3 nanocomposites were prepared using the Pechini method. The prepared nanocomposite was found to be more effective in the degradation of erythrosine dye than that of Nd2O3 under UV light (Zinatloo-Ajabshir et al., 2016). Nanocrystalline neodymium zirconate (Nd2Zr2O7) was prepared and its photocatalytic activity for degradation of eosin Y was investigated (Zinatloo-Ajabshir and Salavati-Niasari, 2017).
Most of the literature focuses on the preparation and the char- acterization of the A2B2O7 composition with the pyrochlore-type structure to be used in different applications. There is a lack in the literature concerning the preparation and characterization of the A2B2O7 composition with fluorite-type structure For the best of our knowledge, the nano-sized undoped or W-doped Fe2Zr2O7 (with fluorite-type structure) never been prepared or characterized in the previous studies. Accordingly, the photocatalytic activity of nano-sized undoped or W-doped Fe2Zr2O7 is not reported yet. In this frame, the aim of the present work is the preparation and characterization of na- nosized undoped and W doped Fe2Zr2O7 using Pechini method which is considered reliable, cost-effective, eco-friendly and easy method to optimize the shape and the grain size of the nano-sized metal oXides. Moreover, studying systematically the impact of W doping on both the
structural properties and the photocatalytic activity for the degradation of methylene blue.
2. Experimental
2.1. Preparation and characterization of the prepared materials
Fe2Zr2-XWXO7 systems were prepared; where x = 0, 0.05, 0.1 and 0.15 using the citrate technique (Pechini method) which is a wet-che- mical method based on polymeric precursor (Pechini, 1967) that was 2016; Ribeiro et al., 2013; Sunde et al., 2018; Y. Zhang et al., 2014).In this method, α-hydroXyacid (citric acid) is used to chelate the cations forming a polybasic acid. PolyhydroXy alcohol (ethylene glycol) reacts with these chelates forming ester and water. Heating the miXture leads to polyesterfiction and after the evolution of nitrous oXide and water, a gel is obtained. The thermal decomposition of this gel results in a chemically homogeneous powder containing the desired stoichio- metry (Carrillo et al., 2016; Ribeiro et al., 2013).
Zirconium (IV) oXynitrate hydrate (Sigma Aldrich), Tungsten (VI) chloride (Sigma Aldrich), Iron (III) nitrate monohydrate (Guangdong Guanghua Sci-Teach), Ethylene Glycol (ROAD, Sandy Croft, Deeside, CLWYD) and Citric Acid Anhydrous EXtra pure (Loba Chemie) are used as starting materials. All chemicals were reagent grade and used as received without any modification.
Fe2Zr2O7 was prepared using the Pechini method as follows: aqu- eous zirconium oXynitrate and iron nitrate solutions were miXed (so- lution A), considering the desired stoichiometry of the metal oXides in the final ceramic powder. Citric acid (CA) was then added to the so- lution (A) to chelate metal cations at the CA: Me molar ratio of 4:1. Me denotes Fe3+, Zr4+ in the final powder. After dissolving of CA, ethylene glycol (EG) was added into the solution at a CA: EG molar ratio of 1:1.5. The solution was then heated at 140 °C and kept under stirring to promote the esterification and polymerization reactions. After the elimination of nitrous oXides and water, a gel was obtained. The gel was charred gradually up to 300 °C then heated in the muffle furnace at 300 °C for 2hrs. The charred gel thus produced was grounded and calcined for 2 h at 500 °C and then grounded and calcined for 2 h at 600 °C. Fe2Zr2-XWXO7 system where x = 0.05, 0.1, 0.15 was prepared using the same sequence. For the preparation of Fe2Zr2-XWXO7 samples, tungsten chloride was dissolved in ethanol and then added to solution (A). Flowchart of the preparation of Fe2Zr2O7 powder was presented in the supplementary data (Fig. S1).
The chemical composition of the catalysts was determined using inductively coupled plasma–optical emission spectroscopy (ICP–OES) 5100 Agilent after samples were digested in a miXture of inorganic sulfuric and nitric acid. X-ray diffraction (XRD) is the standard tech- nique for the determination of the crystal structure of a solid. XRD is used to identify the crystal structure, to determine the lattice para- meters and unit cell volume. The XRD measurements were carried out using 7000 Shimudzo 2 kW model X-ray spectrophotometer with a nickel filtered Cu radiation (CuKα) with λ = 1.54056 Å. The scanning 2 θ range was 5–80 with a step size of 0.2. The lattice parameters were determined using a program called UnitCellWin (Holland and Redfern, 1997). FTIR spectra were recorded in the frequency range 400–4000 cm−1 with a resolution of 4 cm−1 using FTIR 6100 Jasco. Japan Spectrum equipment. The microstructures were studied by transmission electron microscope (TEM, JEOl JEM2100). Diffuse re- flectance measurements were performed to study the optical properties of the prepared samples using Shimaduz UV-3600. The specific surface area of the prepared samples was determined by NOVA Surface Area Analyzer from Thermo Pascal 140 mercury porosimetry under a pres- sure range of 0.1–200 MPa. Mercury surface tension of 480 Dyne/cm and contact angle of 141.3° were used.
2.2. Design of experiment
The experimental design was conducted using Minitab 18 software, the low and the high levels of each factor in addition to the matriX of the full factorial design with removal efficiency results and fits were represented. All the statistical data of the factorial design analysis were analysed using ANOVA software.
2.2.1. Experimental setup and MB determination
A slurry mode batch reactor equipped with a cooling jacket was exercised for photocatalytic assessment experiments. All reaction
used to prepare several metal oXides (Carrillo et al., 2016; Danks et al., commercial visible metal halide lamp (HQI-T250/Daylight, OSRAM GmbH, Germany) with a luminous efficacy of 82 lm/W and luminous fluX of irradiation 20,000 lm (wavelengths 380–780 nm) was used. A schematic of diagram of the photocatalytic degradation experimental set-up is presented in Fig. S2 in the supplementary data. Five millilitres were withdrawn from the reaction vessel at fiXed time intervals and then centrifuged to be separated from the photocatalysts. Different parameters were optimized during the experimental setup such as ir- radiation time (0–120 min), pH (3–12), amount of catalysts (0.50–2.0 g/L), the concentration of MB (20–100 mg/L). The con- centration of MB colour was assessed by the direct fraction of the ab- sorbance concentration curve using Carry 100 UV/Vis spectro- photometer. TOC was analysed by Tekmar – Teledyne TOC analyzer. TOC, pH, TSS and COD (closed refluX colorimetric method) were ana- lysed according to APHA, 2017 (APHA, 2017).
2.2.2. Experimental setup textile wastewater treatment
An experiment was conducted on real blue textile wastewater ac- companied with high organic load under the same optimum experi- mental conditions determined for synthetic MB dye (pH = 11, catalyst dosage = 1.5 g L−1, reaction time = 45 min). Homogeneous textile wastewater sample was collected from the sink of three Dimension Company for Readymade Clothes – Al Dulayl industrial complex – Zarqa city – Jordan. The sample was collected for 24 h to be representative of the daily industrial wastewater production in the factory. The samples were collected by a continuous dosing pump of 250 mL/h making 6 L sample during the day. The sample then preserved in ice boX at 4 °C until it reached the lab for characterization and performing degradation experiments.
The wastewater was characterized before and after treatment by measuring pH, Total suspended solids (TSS) and chemical oXygen de- mand (COD) as an indicator of the organic load of textile wastewater. A slurry mode batch reactor equipped with a cooling jacket was exercised for photocatalytic assessment experiments with vigorous stirring. A commercial visible metal halide lamp with the pre-men- tioned specifications was used in the supplementary data (Fig. S2). Five millilitres were withdrawn from the reaction vessel at fiXed time in- tervals and then centrifuged to be separated from the photocatalysts then COD was measured as an indication for organic loads removal.
2.2.3. Scavenger experiment
To investigate the major active species responsible for MB photo- degradation, different scavenger experiments were conducted. 10 mmol of isopropyl alcohol (IPA), EDTA and benzoquinone (BQ) were added separately to study the influence of HO • radical, photogenerated holes (h+) and O2 radical, respectively.
2.2.4. Toxicity
The treated water was tested via a MicrotoX Model 500 analyzer and the acute toXicity results were given as EC50 (mg/L) according to ISO 11348-3, 2007.The MicrotoX acute toXicity test is based on the luminescence in- hibition of the marine gram negative bacteria. Lyophilized Vibrio fi- scheri (NRRLB-11177) was used as the test organism. The bacterial suspension was added to the sample osmotically arranged with 2% NaCl and to the sample dilutions. After the bacteria were exposed to the sample, photometry was carried out periodically.
3. Results and discussion
3.1. Characterization of the prepared materials
In order to confirm the chemical composition of the prepared cat- alysts, ICP analysis was conducted by determination of the weight percent of Zr, Fe and W for the prepared samples. The experimental weight percent values of Zr, Fe and W in all the prepared samples were accorded with the expected weight percent (Table S1 in the supple- mentary data).
The stability of A2B2O7 whether pyrochlore or defect-fluorite structure depends on the ionic radius ratio of the A and B cations (rA/rB) (Catchen and Rearick, 1995; Mandal et al., 2007, 2006). When rA/ rB = 1.46–1.78 (Whittle et al., 2009), pyrochlore structure is formed while the defect-fluorite structure is formed for the lower or upper limit of the mentioned values for rA/rB for the pyrochlore structure.
In the present study, for Fe2Zr2O7 (ZF sample), the (rA/rB) = 0.76 where A = Fe3+ (ionic radius ls = 55 p.m. (Shannon, 1976)) and B]Zr4+ (ionic radius = 72 p.m. (Shannon, 1976)). From the XRD pattern of ZF sample (Fig. 1), ZF sample is crystallised in cubic fluorite phase of Fe2Zr2O7 as major phase (as expected from (rA/rB) ratio) in addition to rhombohedral phase of Fe2O3 (PDF 24–0072) and monoclinic phase of ZrO2 (PDF 37–1484) as minor phases. The cubic fluorite phase of Fe2Zr2O7 is best matched to the PDF 78–1299 for Er0.5Zr0.5O1.75 where there is no card for the novel Fe2Zr2O7 material. The shift to lower theta value of the cubic fluorite phase of Fe2Zr2O7 as compared to that of Er0.5Zr0.5O1.75 is due to the difference in the ionic radii between Er3+ ion (ionic radius = 89 p.m.) and Fe3+ ion (ionic radius ls = 55 p.m. (Shannon, 1976)). The peaks at 2 θ about 14°, 28°, 37°, 45° corresponding to (1 1 1), (3 1 1), (3 3 1) and (5 1 1) planes respectively (L. Zhang et al., 2014) characteristic for the pyrochlore structure aren’t detected. Fig. 1 shows the X-ray diffraction pattern for Fe2Zr2-XWXO7 system, it was found that as the concentration of W increases, the peak intensity of the peaks characteristic of Fe2O3 decreases until it dis- appeared for ZFW3 sample where single cubic fluorite phase was de- fined. This might be due to W6+ ion (ionic radius 60 p.m. (Shannon, 1976) which is smaller than that of Zr4+ ion (ionic radius 72 p.m. (Shannon, 1976)) and larger than Fe3+ ion (ionic radius ls = 55 p.m. (Shannon, 1976)) leads to improving the stability of the cubic fluorite phase structure of Fe2Zr2O7.
The calculated cubic lattice parameter and unit cell volume, as well as the samples names and composition, are presented in Table 1. For Fe2Zr2-XWXO7 system, the cubic lattice parameter and unit cell volume for W doped Fe2Zr2O7 samples is smaller than that of the undoped Fe2Zr2O7. This might be attributed to W6+ ions (ionic radius 60 p.m. (Shannon, 1976)) substitutes Zr4+ ions (ionic radius 72 p.m. (Shannon, 1976)).
Fig. S3 in the supplementary data showed the FTIR spectra for the prepared samples. The broad absorption peak at about 3429 cm−1 was detected for the undoped ZF sample which is assigned to the stretching vibration of OH group in water molecule. Besides, the absorption peak characteristic to the bending vibration of water molecule was detected at 1640 cm−1 (Yakout and Hassan, 2014). The band detected at 454 cm−1 is assigned to Zr–O vibration (Gurushantha et al., 2015). The peaks at 1040 cm−1 might be associated with stretching vibrations of Zr–O terminals (Badenes et al., 2002). The peak at about 873 cm−1 corresponds to the bending vibration of hydroXyl groups bounds to zirconium oXide (Yakout and Hassan, 2014). The bands at about 540 and 455 cm−1 are assigned to Fe–O stretching vibration (Zolghadr et al., 2017).
Interestingly, a little shift in the peak positions and peak intensities with increasing the W concentration was observed for the doped sam- ples (ZFW1, ZFW2 and ZFW3 samples). The region (600–900 cm−1) might correspond to O–W–O stretching modes (Rougier et al., 1999).
Diffuse reflectance spectroscopy was used to study the optical properties of the prepared samples. Fig. S4 in the supplementary data shows the diffuse reflectance spectra of ZF, ZFW1, ZFW2 and ZFW3 samples. It is clear that the absorption edge for Fe2Zr2-XWXO7 system lies in the visible light range. The band gap value is obtained by plotting of F (R∞) E)2 versus photo energy according to Kubelka-Munk (Fig. S5) in the supplementary data. The linear fit is performed to determine the intersection with the x-axis which represents the band gap value (Fig. S6 in the supplementary data). The obtained band gap values were presented in Table 1. The band gap of ZF sample is 1.96 eV. As W concentration increases the band gap decreases. The lowest band gap value was obtained for ZFW3 sample (1.69 eV) which enables the prepared Fe2Zr2-XWXO7 system to be used as photocatalyst in the visible light range.
ZrO2 has a wide band gap (5 eV) which renders its absorption of visible light (Sayama and Arakawa, 1996, 1994). In the present work, it is clear that the undoped, as well as W doped Fe2Zr2O7, has a smaller band gap than that of ZrO2 and some pyrochlore structured oXides such as Sm2Zr2O7 (2.86 eV) (Uno et al., 2006) and Nd2Zr2O7 (2.67 eV) (Uno et al., 2006).
For Sm2Zr2O7 and Nd2Zr2O7, the conduction band composed of Zr 4d orbitals whereas the valence band consists of O 2p bands. The Sm or Nd 4f orbitals lie on the top of O 2p bands thus raising the band gap level which in turn results in decreasing the band gap (Uno et al., 2006). In Fe2Zr2O7, Fe might form a new energy level on the top of O2p bands resulting in decreasing the band gap (Fig. S7) in the supple- mentary data. This is in agreement with that for SrTi0.7Fe0.3O3 (Abbas and Jamil, 2016), Ga doped and Sc doped SrTi0.7Fe0.3O3 (Abbas et al., 2016; Jamil et al., 2017).
For Fe2Zr2-XWXO7 system, tungsten as a dopant might form a level in the bottom of the conduction band resulting in decreasing of the band gap (Fig. S7) in the supplementary data. This is in agreement with (Abbas et al., 2014) for Zr doped indium tin oXide (ITO).
Fig. S8 (a-d) in the supplementary information represents the mi- crographs of all the prepared samples. Small quasi-spherical particles, which agglomerate into denser aggregates were observed. The particle size range determined by TEM for all samples is presented in Table 1. All the samples are in the nano-size range (3–7 nm). Fig. S9 in the supplementary data for ZFW3 sample (the highest surface area, lowest band gap and the best photocatalytic activity sample) shows lattice fringes spacing with interplanar distances of 0.2681 nm which could be
identified with (104) plane for rhombohedral phase of Fe2O3 structure. Interestingly, the sum of the 10 interplanar distances (2.681 nm) is close to the d-value (2.70620) of (104) plane of Fe2O3.
The surface area is one of the important parameters used to de- termine the photocatalytic activity of the catalyst. More dye molecules are adsorbed on the high surface area catalyst (Abhilash et al., 2019). BJH method was implemented to define the specific surface areas for the prepared materials which were significantly increased with the increment of tungsten concentration (Table 1). A type-IV isotherm with perfect mesopores firm materials were manifested for the intended materials as shown in Fig. S10 in the supplementary data which is declared by the hysteresis loop that is related to the aggregates pre- sented in the mesopores and the little marked uptake over the high of P/Po (Sharaf El-Deen et al., 2016). There is significant variation in the calculated surface area since; the surface area of ZFW3 was increased by 3.6 times than that of the parent ZF. The high surface area of the intended materials presumed that surface area may have a significant impact on the photocatalytic mechanism.
3.2. Photo assisted degradation activity of methylene blue dye
3.2.1. Degradation time effect
Fig. 2a demonstrates the impact of degradation time on photo- degradation of methylene blue (MB) by the visible/Fe2Zr2-XWXO7 system. MB degradation increases with extending the reaction time up to 45 min. Moreover, no noteworthy change in the photodegradation reaction was detected. Superior photodegradation rate at the first 20 min was noticed. This can be reasoned to the direct reaction between the catalyst and MB. Since, in the early reaction stages, there is no re- action by – products that might compete with MB itself in the reaction of the catalyst that can be adsorbed on the surface of the catalyst and hinder the photodegradation rate (Yang et al., 2017). The reaction rate was almost steady after 45 min until the end of the second hour in- dicating that the optimum reaction time is 45 min.
Under visible light, the energy gained by the electrons shifts them from the valence band to higher conduction band generating positive holes and negative electrons on the Fe2Zr2-XWXO7 surface. The hole interacts with water or OH− creating OH•. Electron in the conduction band ( CB) reacts with the atomic oXygen forming superoXide anion. HydroXyl (HO· ) and superoXide (O· 2 ) radicals are viewed as the re- ceptive species that will be adsorbed on the surface of Fe2Zr2-XWXO7 (Liu and Chen, 2009; Lu et al., 2011). These created radicals interact with MB to accomplish decomposition (equations (1)–(5)).
Fig. 2. Photocatalytic degradation for the prepared samples as a function of a) time, b) catalyst dose and c) pH under visible light. Fig. 2d Pseudo first order kinetic for different MB concentrations with Fe2Zr1.85W0.15O7 (ZFW3 sample).
3.2.2. The impact of photocatalytic loads
Variable loads of Fe2Zr2-XWXO7 systems (from 0.5 to 1.5 g/L) were used in this part with an initial dye concentration of 20 mg/L at pH7 and 45 min. Fig. 2b demonstrates that as the catalyst dose increases up to 1.5 g/L, the MB photocatalytic degradation increases. This might be due to increasing the adsorption of MB molecules on the catalyst sur- face and also increasing surface-active sites as well as increasing hy- droXyl radicals creation (Zhao et al., 2011). At higher loads (2.0 g/L), the turbidity of the solution increases causing scattering of the visible light and hindering its effort to penetrate the solution. Thus, the high loads of catalysts could not be illuminated by visible light; which in turn decreases photonic productivity (Reza et al., 2017). Accordingly; 1.5 g/L is the optimum catalyst load that can be used.
3.2.3. The impact of pH value
The experiment was accomplished at the optimum contact time and catalyst load using 20 mg/L MB concentration with variable pH values (3, 5, 7, 9, and 11). Fig. 2c displayed the influence of pH value on photo-degradation by visible/Fe2Zr2-XWXO7 system. Certainly, it was
found that the decolourization efficiency increased as the reaction pH increased from 3 to 11, due to a higher rate of generation of OH.in alkaline solution by oXidizing a greater amount of OH− available on the surface of the catalyst. Thus, MB was enhanced the decolourization in the alkaline range. The lowest determined colour was noticed at pH 11 indicating that pH 11 is the optimum pH for that reaction. After pH 11, full saturation of the surface of the catalyst by MB occurred hindering the availability for adsorption and reducing photocatalytic activity (Sahoo and Gupta, 2015). In the acidic pH range MB, being an anionic dye undergo repulsion with the catalyst surface preventing the arrival of dye to its surface. Thus, the decolourization efficiency was lower in the acidic pH range (Jing et al., 2014).
The effect of W doping was deeply investigated through this study including XRD, IR, DRS, BET as well as the photocatalytic activity. The doping has a positive role in improving the photocatalytic efficiency of the catalyst. This might be attributed to i) new energy level formation between the valence and the conduction bands besides, (ii) enhancing the adsorption of the pollutant molecules on catalyst surface by varying the acid-base properties of catalyst surface. For the prepared samples, the W doped samples have higher photocatalytic activity than that of the undoped sample. As the W concentration increases, the photo- catalytic degradation of MB increases as shown in Fig. 2a, b, c which is in accordance with the band gap and surface area. Increasing the photocatalytic activity for ZFW3 sample as compared to that of the undoped ZF sample might be attributed to the substitution of Zr4+ by W6+ in the Fe2Zr2O7 lattice which is reflected in the values of both the cubic lattice parameter and unit cell volume and introduction of new levels between the valence band and the conductance band reducing the band gap values. In the visible light irradiation, an electron is promoted from the valence band to the conduction band through the created new level leaving a hole in the valence band. The promoted electron and the hole react with the adsorbed oXygen and the OH− respectively forming O2 • and HO• on the catalyst surface which are responsible for MB degradation (Abbas et al., 2016).
3.2.4. The impact of MB dye concentration
The order of the degradation rate was manifested under the opti- mized conditions of pH and catalyst dose (1.5 g/l catalyst dosage and pH11) using different MB concentrations (20, 30, 40, 50 and 100 mg/L) then withdrawing samples at variable time intervals. Fig. 2d described that MB degradation decreases with increasing the initial dye con- centrations from 20 to 100 mg/L. This can be attributed to the decrease in the facility of the visible light penetration of the solution decreases with the increment in the MB concentration, subsequently, the photonic efficiency increased with lowering the initial dye concentration. The decrease of photonic efficiency with increasing initial dye concentra- tion is credited to the increment of collision frequency between dye and photons collision frequency which is in alliance with (Dai et al., 2013). At high concentrations, the interaction between free radical and elec- tron-holes can be hindered due to the extensive dye adsorbed on the surface of photocatalyst causing blocking of the surface of the catalyst (Rauf et al., 2011).
The photocatalytic degradation rate of MB dye in heterogeneous photocatalytic systems under visible-light illumination followed the Langmuir–Hinshelwood equation (7) (Dung et al., 2005; Mills and Le Hunte, 1997; Muhd Julkapli et al., 2014; Vallejo et al., 2017) that can be represented as ln C0/C = kKt = kapp_t (7) Ln C0/C was plotted versus t giving the apparent rate constant for photocatalysis of MB from the slope of curve fitting line and the in- tercept is equal to zero. In the meantime, the direct connection between ln C0/C and t demonstrate that the photocatalytic degradation reaction also follows the pseudo first-order reaction, the apparent rate constants were figured to be 0.135, 0.175, 0.205, 0.225 and 0.289 min−1 for Fe2Zr2-XWXO7 system where x = 0, 0.05, 0.1, 0.15 respectively.
In the other another hand, the mechanism of MB removal using the Fe2Zr1.85W0.15O7 nanoparticle as photocatalyst was studied by pseudo- second-order kinetic models. The linear plot of t/qt vs. t according to equation (8) is shown in Fig. S11 in the supplementary data. The value of the correlation coefficient R was lower than that of the pseudo-first- order model (> 0.99) indicating that the photocatalytic degradation of MB dye in aqueous solutions using Fe2Zr1.85W0.15O7 follows the pseudo-first-order kinetic model.
3.2.6. The reusability for Fe2Zr1.85W0.15O7 sample
One of the vital obstacles of the photocatalytic implementation in industrial wastewater remediation is the reusability of the catalyst. So as to underline the number of cycles reuse of Fe2Zr1.85W0.15O7 sample in decolourization of MB. Ten cycles for MB decolourization over Fe2Zr1.85W0.15O7 sample were operated under visible light at the opti- mized reaction conditions. The catalytic solution was settled for enough time then the catalyst precipitated at the end of the reactor, after se- paration from the supernatant, the catalyst was introduced for to the decolourization of another sample of MB under the same conditions. The variation in % MB removal with different cycles is shown in Fig. 4. The slight reduction in photocatalytic efficiency of Fe2Zr1.85W0.15O7 sample (100% in the 1st cycle to 95% in the 10th cycle) pointed to acceptable results for MB decolourization using Fe2Zr1.85W0.15O7 sample under visible light irradiation confirming that the pre-men- tioned sample can be reused without losing the profitable synergy action.
3.2.7. Photo-degradation of a real textile wastewater
In order to examine the competence of ZFW3 as a photocatalyst under visible irradiation. An experiment was conducted on real blue textile wastewater accompanied with high organic load under the previously optimized experimental conditions. The COD concentration of raw textile wastewater was 749 mg L−1 (Table 2). The residual COD concentration of raw textile wastewater was 231 mg L−1 after 45 min, corresponding to COD removal of 69.2% which is lower than % COD removal for synthetic MB dye (97%) as shown in Fig. 5, TOC residual was < 0.1 mg C/L and the initial TOC of 20 mg/L MB was 11.95 mg C/ L corresponds to TOC removal around 99.9% for synthetic MB dye. However, in the case of real textile wastewater, the residual TOC was 650 mg C/L corresponds to TOC removal 73.7% and residual MB was The results showed that W concentration up to 15 mol % had a remarkable effect in the colour removal of MB because of the gradual increase in the surface area shifting the photo excitation reaction of Fe2Zr2O7 (ZF) to the visible region and decreasing the band gap energy. Moreover, the parent ZF had Fe2O3 phase which is the reason for lowering the surface area which has a vital role in the low photo- catalytic activity of Fe2Zr2O7. Fe2Zr1.85W0.15O7 nanoparticles have been elected as a promising new photocatalysts due to its superior activity compared with other photocatalysts (Table S2). 3.2.5. The photocatalytic degradation pathway of MB GC-MS study carried out at the end of the photocatalytic degrada- tion of MB using Fe2Zr1.85W0.15O7 under the optimum operating con- ditions was presented at Fig. S12 in the supplementary data. MB de- gradation was started by hydroXyl radical attack on the central imino- group (N = C) underwent double bond cleavage giving C16H21N3SO with m/z = 303. Followed by the cleavage of the double bond of the –S = group in the Para position in the central aromatic ring yielding a substituted aniline as seen for products at m/z = 189.2, 201.2 and 221 in Fig. 3. Consequently, the amino group can be substituted by hydroXyl radical forming the corresponding phenol derivative with m/ z = 190.18, 110 and 94. In an advanced stage of degradation, the destruction of benzene rings occurred producing short chain acids such as acetic acid and oXalic acid (Houas et al., 2001; Jing et al., 2014). Finally, short chain acids mineralized to inorganic substances (carbon dioXide, water, nitrate and sulfate). Accordingly, the concentration of nitrate and sulfate was not in the real wastewater sample over the synthetic sample as well as the higher suspended solids content that increase the turbidity of the so- lution. The high turbidity in case of real waste hinders the penetration of irradiation light resulting in lower color removal (Khan et al., 2016). As well, the degradation intermediates produced in case of treated real textile wastewater need more OH• radicals than that for the synthetic dye, as a result, further oXidation of intermediates need relatively longer reaction times (Jorfi et al., 2018). 3.2.8. Scavenger experiment To evaluate the radical responsible for photocatalytic degradation, scavenger experiment was performed for ZFW3 under the optimum conditions (MB concentration 20 mg/L at 1.5 g/L catalyst dose and pH 11 for 45 min) using 10 mmol L−1 of different scavengers (isopropyl alcohol (IPA), EDTA and benzoquinone (BQ)) were conducted. When 10 mmol L−1 isopropyl alcohol (IPA) was incorporated as HO• sca- venger, the degradation of MB was lightly influenced (MB % removal achieved 89%), which demonstrated that the HO• was not the major interactive species. However, the degradation process could be pre- vented proficiently when 10 mmol L−1 Benzo Quinone (BQ) was used since; MB removal was diminished to 56%. It indicated that the O2 assumed a demonstrating role in the catalysis process. At the point when 10 mmol L−1 EDTA was included, the catalytic degradation of MB could be strongly inhibited (MB removal achieved 25%), demon- strating that the h+ also showed a significant role in the catalytic process. Along these lines, h+ and O2 had the major contribution to photocatalytic degradation of MB while, HO• had a minor contribution (Chiu et al., 2019). Fig. 4. Number of cycles for MB over Fe2Zr1.85W0.15O7 sample under visible light at optimum conditions. 3.2.9. Toxicity A solution of 107 cell/mL Vibrio Fischeri organism at pH7 was used to assess the toXicity of the treated water using ZFW3 by incubation for 10 min and subjected to heavy light. At the same time, a control sample was typically treated. No significant difference was noticed between the control sample and the Vibrio Fischeri sample. The treated water by ZFW3 was proved by > 100 mg/L MicrotoX EC50 reading according to (Jamil et al., 2018).
3.2.10. Experimental design (factorial design)
Applying statistical data analysis of the factorial design results were estimated in order to determine the significance between and within the low and the high levels of each factor for Photocatalytic activity of ZFW3 (Table S3 in the supplementary data) and Pareto chart, was plotted and interpreted in Fig. S13 in the supplementary data.
3.2.10.1. Pareto chart. Values of the main and the interaction effects were expressed in the Pareto chart plot (Fig. S13 in the supplementary data) and represented by the horizontal columns. The t-value revealed by the vertical line “reference line” existed on the graph indicated the determine the relative impact of each factor because the coefficients are scaled to accommodate the units of each factor and the intercept is not at the center of the design space.
Transformation was performed as shown in BoX CoX transformation plot Fig. S 15 (in the supplementary information), the response range from 0 to 100 and Range ratio max to min is 100. For ratio > 10 transformation is required and the square root equation is represented as shown in equation (9).
Fig. 5. Photocatalytic degradation of MB dye and real textile wastewater using Fe2Zr1.85W0.15O7 sample under visible light at optimum conditions.
3.2.10.2. ANOVA analysis for selected factorial model. Analysis of variance (ANOVA) was estimated in order to determine the significance between and within the low and the high levels of each factor (Tables S4–S7 in the supplementary information).
The Model F-value of 60.39 implies significant of the model. There is only a 1.64% chance that F-value this large could occur due to the noise. Meanwhile, P-values < 0.0500 indicating that model terms are significant. In this case irradiation time, catalyst dose, interaction be- tween irradiation time and catalyst dose are significant model terms (Fig. S 14 in the supplementary data). Values > 0.1000 indicate that the model terms are not significant. If there are many insignificant model terms (not counting those required to support hierarchy).
The Predicted R2 of 0.8947 is in reasonable agreement with the Adjusted R2 of 0.9770; i.e. the difference is less than 0.2. Precision, measures the signal: noise ratio, is 19.111 with adequate signal which is accepted since it is > 4. This model can be used to navigate the design space.
The coefficient estimation represents the expected change in the response per unit change in factor value when all remaining factors are held constant. The intercept in an orthogonal design is the overall average response of all the runs. The coefficients are adjustments around that average based on the factor settings. When the factors are orthogonal the VIFs are 1; VIFs > 1 indicate multi-collinearity, the higher the VIF the more severe the correlation of factors. As a rough rule, VIFs less than 10 are tolerable.
4. Conclusion
Nano-sized Fe2Zr1-XWXO7 system was prepared and characterized by XRD, FTIR, TEM, DRS and BET. The single cubic fluorite phase was detected for Fe2Zr1.85W0.15O7 sample. Such sample is characterized by the lowest band gap and the highest surface area. Fe2Zr1.85W0.15O7 showed a promising photocatalytic degradation of MB where complete decolourization was noticed at the optimum conditions (1.5 g/L cata- lyst dose, pH 11 after 45 min). Degradation of MB following the pseudo first order kinetic. The h+ and O2 are the major contributors to pho- tocatalytic degradation of MB while, HO• had a minor contribution. Degradation by-products had been studied by GC/MS at the end of the reaction. NontoXic prepared materials as well as the treated water using ZFW3 has been proved by EC50 > 100 mg/L MicrotoX. Fe2Zr1.85W0.15O7 sample was applied to photocatalytic degradation of real textile wastewater obtaining 69.2% COD removal under the same operating condition. Accordingly, Fe2Zr1.85W0.15O7 is proved to be used as heterogeneous photocatalyst for removal of dyes in textile industrial wastewater especially with its noticeable reusability for 10 times with only 5% decrease in its efficiency.