Recently, in most studies, TiO2 nanoparticles have been used for photocatalytic removal of pollutants from water because of their special photocatalytic properties. But the major disadvantages of the application of this nanoparticle in water treatment are the high cost of manufacturing nanoparticles and the probability of their occurrence in the environment. To overcome these problems, one of the suggested options is fixing materials on a static surface. However, this action reduces the number of active catalyst sites and consequently reduces the photocatalytic ability of these materials compared to non-supported nanocatalysts [1, 2].
Another promising solution for solving the problem of recycling nanoparticles in slurry systems is using magnetic nano-photocatalysts in the water treatment processes. When magnetic nanocomposites are used, they can be easily separated from the water by an external magnetic field . Nanocomposites recycling will make the water purification processes with nanoparticles will be environment-friendly, more efficient and cost-effective .
So far, several magnetic nanocomposites have been used as photocatalyst for water purification, which in most cases follow a core-shell structure, consist of a formulated composite with photocatalytic and magnetic nanomaterials. Usually, the internal core consists of magnetic elements such as iron, nickel, cobalt or their oxides, such as magnetite (Fe3O4), megamet (c-Fe2O3), cobalt ferrite (COFe2O4) . In addition, the shell is usually made with photocatalytic materials such as TiO2, ZnO, BiOCl and AgBr . Recently, extensive studies have been carried out on the use of TiO2 as coatings and SnO2 and Fe3O4 oxides as a magnetic core . Composition of TiO2 photocatalytic nanoparticles with Fe3O4 superparamagnetic nanoparticles is not perfect because composition of these two nanoparticles together prevents the absorption of high-frequency light by TiO2 nanoparticles. Also, the direct contact of the magnetic material and the TiO2 nano photocatalyst causes the electron incompatibility issue [8, 9]. As a result, this phenomenon will cause a reduction in photocatalytic activity . One solution to increase the photocatalytic activity of magnetic nanocomposites is doped them with another material such as SiO2 nanoparticles , Ag  and C  to them, which were recently studied by researchers.
The main reason for the electron recombination in the nanocomposite is the proximity of energy level of Fe3O4 as conductor and the lowest unoccupied molecular orbital (LUMO) of the TiO2 as semi-conductor transferred easily on the Fe3O4 nanoparticle surface and then will be recombined. According to studies, this issue can be solved by the addition of nanoparticles such as graphene or carbon nanotubes to the nanocomposite. According to Fig. 1, the excited electrons will be transferred on the Fe3O4 surface and then will be transferred on MWCNT due to the lower energy of the MWCNT’s work function.
According to studies, MWCNTs have a high ability to accept and scavenge electrons. Thus, by reducing the electron recombination the photocatalytic activity will increase [14, 15, 16]. In a study using TiO2-Fe3O4/graphene nanocomposite, methyl blue dye photodegraded with 98.8% efficiency under UV light . Also, the dye photodegraded with 100% efficiency under xenon light in another study . Other similar studies have been done in the recent years consisting of photodegradation of RhB dye using of Graphene–TiO2–Fe3O4 , photodegradation of methyl Orange using of Fe-doped TiO2 under visible light  and photodegradation of phenol using of Fe3O4/C/TiO2 .
In the present study, the magnetic and photocatalytic TiO2/Fe3O4/MWCNT nanocomposite was prepared using a simple, cost-effective and energy-efficient method to make nano photocatalyst materials with high photodegradation ability, reusable as well as the ability to be produced it industry. Other purposes of this study are to investigate the morphology, functional and photocatalytic activity of the nanocomposite in laboratory and real conditions.
Carbon Nanotubes (MWCNTs) with an average diameter of 8 nm, length of 30 μm and purity of 95%. Iron (III) Chloride Hexahydrate (FeCl3, 6H2O), Iron (II) Sulfate Heptahydrate (FeSO4, 7H2O), Ammonium hydroxide (NH4OH) and Hydrochloric acid were used in the synthesis of iron oxide nanoparticles. TiCl4, ethanol (CH3CH2OH), and Milli-Q water were used to prepare the TiO2 nanoparticles and ammonium Sulfate ((NH4) 2SO4) was used for the preparation of Fe3O4/TiO2 nanocomposites. Also, sulfuric acid (H2SO4) was used to activate the MWCNT carbon nanotubes. All chemicals were analytical grade and were purchased from Merck.
Synthesis of TiO2
The TiO2 anatase phase nanoparticles were synthesized using the hydrothermal method from TiCl4 aqueous solution. For pure TiCl4 and diluted it to 1 M was dissolved into the chloridric acid. In the following 100 ml of the TiCl4 1 M solution was added to 200 ml of Milli-Q water. Drops of ammonium hydroxide (NH4OH) were added slowly into the solution of titanium tetra chloride (TiCl4) in order to adjust the pH value to 7 then the Milli-Q water was added again to increase the solution volume to 500 ml so as to adjust the concentration of Ti (IV) to 0.2 M. After stirring vigorously, the solution was made to settle for 12 hours. Then, the precipitate was centrifuged. The obtained sol was washed with deionized water to remove the chloride ions. Then, the precipitate was dried at 200°C to remove the absorbed water for 3 hours and finally amorphous TiO2 was obtained. The obtained amorphous TiO2 was calcinated at the temperatures of 300°C (heating rate 30 °C /min) for 4 hours .
Synthesis of Fe3O4
3.01 g FeSO4·۷H2O and 2.04 g FeCl3.6H2O were dissolved in deionized water. Both solutions then mixed together in an ultrasonic bath simultaneous nitrogen gas with a flow rate of 0.5 L min-1 and 1 atm pressure was introducing into the solution. The reaction process started when drops of 1M sodium hydroxide (NaOH) were added into the solution, then a black suspension appeared indicating the formation of magnetite. The precipitated nanomaterials were separated using the external magnetic bar and finally collected with a centrifuge and dried in Temperature at 80oC for one day .
Preparation of TiO2/Fe3O4
At first 100 mg of prepared Fe3O4 nanoparticle was dispersed into the 200 mL deionized water. On other beakers, 1 g TiO2 was dispersed into 100 mL of 0.02 M (NH4)2SO4 and mixed with the magnetic nanoparticle suspension. The suspension was sonicated for 30 minutes at room temperature then centrifuged at 4000 rpm to obtained solid material. The obtained solid was dried in oven for 18 hours, at a temperature of 100o C. the nanocomposite was obtained and ready for further experiments .
Surface modification of MWCNT
MWCNT were modified to introduce surface functional groups, e.g. -OH and -COOH. Briefly, 1.0 g MWCNT were added into 50 mL 1.7 M ammonium persulfate and 2.0 M H2SO4 mixture solution and was then placed in an ultrasonic bath for 3 hours. Upon completion, the mixture was kept in an oven at 50oC for 17 hours. Finally, the functionalized MWCNT was rinsed by 100 mL deionized water three times and dried in an oven at 60oC overnight .
Synthesis of TiO2/Fe3O4/MWCNT nanocomposite
In order to synthesis the electromagnetic and reusable TiO2/Fe3O4/MWCNT nanocomposite at first, the functionalized carbon nanotubes, in different weights 1, 2.5, 5, and 10 individually were dispersed into 100 mg ethanol (CH3CH2OH) then was placed in an ultrasonic bath for 30 minutes. In another container, 10 mg of the prepared Fe3O4 magnetic nanoparticles was dispersed in 200 ml of deionized water, then the solution was sonicated for 30 minutes, in the following the prepared Fe3O4 solutions were added to a nanotube carbon solution, as a result the nanoparticles iron covered the surface of carbon nanotubes. After washing the mixture with deionized water, it was placed inside the oven at 60 ° C for 24 hours to calcination procedure. In the next step, the prepared Fe3O4/MWCNT nanocomposite was added to 200 mg deionized water and was sonicated for 30 minutes. In the last step, by using a non-homogeneous sonochemical method, the surface of the Fe3O4/MWCNT nanocomposite was coated with TiO2 nanoparticles (In Fig. 2, a schematic representation of how nanoparticles are deposited in TiO2/Fe3O4/MWCNT nanocomposites). It is known that Fe3O4 and TiO2 have a different isoelectrostatics point . Based on the measurements of zeta potential, the surface charge of Fe3O4 changes from (+) to (-) at pH 8.5 while that of TiO2 changes from (+) to (-) at about pH 5. At pH range above 5 and below 8, Fe3O4 possesses positive surface charge whereas TiO2 has negative charge so that an attractive force would adhere to two particles together.
The crystal phase of the synthesized nanocomposites was characterized by X-ray diffraction (Philips X’Pert MPD). The elemental composition was studied by Dispersive X-ray Spectroscopy (EDS) analysis. The morphology and structure of the samples were investigated by Field Emission Scanning Electron Microscopy (FESEM, MIRA3, TESCAN), UV–Vis Diffuse Reflectance Spectra (DRS) of the samples were recorded on a UV–Vis spectrophotometer (Avaspec-2048-TEC) with an integrating sphere attachment. Zeta potential of nanoparticle samples was investigated by Zeta potential analyzer (SZ-100z, Horiba Jobin Jyovin). Spectra of the samples were measured using UV–Vis spectrophotometer (Hach - DR4000) in the range of 180 – 800 nm. Magnetometer (VSM) analysis was carried out on a Lakeshore V8 M7410 magnetometer. Ultrasonic bath, (Fungilab, UE-6SFD) to synthesis and disperse the nanomaterials.
Photo reactor and photocatalytic activity experiment
The photocatalytic activity of the synthesized nanocomposites was evaluated for four different dyes including Acid red 14, Methyl orange, Reactive Red 76 and Acid blue 19. Table 1, shows the specification of the evaluated dyes. To experiment the photocatalytic activity used a reactor (Fig. 3) consists of a glass reservoir of 600 cc and a cross-sectional area of 24/50 cm which surrounded by three electromagnets, to isolate and the maximum use of light, the reactor was embedded in an isolated optical box. For the visible light source, a 55 W Xenon lamp was used due to the closest radiation spectrum to sunlight .
In each experiment, 50 mg of nanocomposite was added to 50 mL dyes solution in the reactor at room temperature. Before irradiation, the suspension was stirred in the dark for 30 min to measurement the adsorption-. After the start of photocatalytic reaction in specified times, 5 mL of the suspension was taken and then the nanocomposite particles were separated by a magnetic to measurement dye concentration. Also, reusability of TiO2/Fe3O4/MWCNT nanocomposites was tested to evaluate the durability and stability, for this the nanocomposite was deposited and separated by a magnet within 1 minute, then washed with distilled water and dried at 60 ° C for 2 hours,
To ensure the accuracy of the results and to minimize the human error factor and equipment, the tests were performed with three times repeated and the means of data was used as results experiments.
RESULTS AND DISCUSSION
Determination of the optimal ratio of composition of TiO2/Fe3O4/CNT
To investigate the effect of the compounds ratio on dye removal, TiO2/Fe3O4/CNT nanocomposites in four ratios (100: 10: 10), (100: 10: 5), (2.5: 100: 10) and (100: 10: 1) were made and investigated under visible light. According to Fig. 4, after 5.5 hours the amounts of dye removal efficiency were 56.66, 60.27, 85.96 and 77.69%, respectively. The maximum efficiency was related to the nanocomposite with a ratio of 2.5: 100: 10.
DRS analysis results
Light absorption properties and energy gap on each of TiO2, TiO2/Fe3O4, and TiO2/Fe3O4/MWCNT were studied by DRS analysis. The results of the DRS is shown in Fig. 5. At wavelength of 200 to 400 nm the TiO2 nanoparticles have the highest light absorption whereas in TiO2/Fe3O4  and TiO2/Fe3O4/MWCNT composite the highest light absorption at a wavelength of 400 to 800 nm were observed.
In Fig. 6, the sharp edge gradient indicates the transition of electrons from the energy gap between the valence layer and the conduction layer, which is possible to determine the energy gap by using the equation 1 for each of the nanocomposites. The results are presented in Table 2.
(E) = h*C/λ (1)
In equation 1, h= 6.626-34 J/s (Planck’s constant), C is the velocity of light in vacuum condition, which is equal to 3×10+8 m/s and λ is the wavelength of the light at the edge of the absorption.
According to Fig. 6, bond gap value for the TiO2 nanoparticles is 3.22 eV which after addition of Fe3O4 nanoparticles to TiO2 band gap were shifted to the red and was measured in the range of 2.2 eV. It is also observed that with the addition of MWCNT, the energy gap shifted to blue (2.94 eV) that the highest absorption spectrum was in the visible range about 436 nm.
XRD measurements were carried out to investigate the crystal identity of TiO2/Fe3O4 and TiO2/Fe3O4/MWCNT nanocomposites. Fig. 7 shows the XRD pattern TiO2, TiO2/Fe3O4, and TiO2/Fe3O4/MWCNT. The diffraction peaks at 2θ = 25.37°， 37.81°， 48.21°， 54.01°， 55.11°，63.01°，69.02° and 75.25° are determined The presence of TiO2 nanoparticles in the anatase phase . In the TiO2/Fe3O4 and TiO2/Fe3O4/MWCNT diffraction peaks at 2θ = 30.21°， 35.51°， 43.31°， 53.21°，57.21° and 63.01° indicate the presence of Fe3O4 superparamagnetic nanoparticles .
SEM and EDX analysis
FESEM and EDS tests were performed to investigate the morphology and ensure the formation of nanocomposites. In Fig. 8, FESEM images of Fe3O4, MWCNT nanoparticles, TiO2/Fe3O4, TiO2/MWCNT, and TiO2/Fe3O4 /MWCNT nanocomposite are presented.
The average diameter of the nanoparticles of iron oxide is between 14 to 35 nm (Fig. 8-a), and the average particle diameter in TiO2/Fe3O4 nanocomposite is between 19 to 25 nm (Fig. 8-b), Compared to a similar study reported by Harifi et al. (2014) , they are of a lower average value, which indicates the high quality of nanocomposite production.
In Fig. 7, TiO2/Fe3O4/MWCNT nanocomposite FESEM images can be seen in two different parts, the particle size is in the nanoscale size and despite the presence of carbon nanotubes, there is not much accumulation and particles are separated, also is observed Fe3O4 and TiO2 nanoparticles are well covered the carbon nanotubes. In Fig. 8-F, it is observed that the diameter of the nanotube has been increased, which can be due to the coating of two layers of nanoparticles. TiO2 nanoparticles were bonded to the surface of the MWCNT also the nanoparticles themselves are almost continuously bonded so that there are no uncovered parts on the surface of the nanotubes.
The EDS analysis was performed to determine the ratio of particles in nanocomposite particles in optimal conditions. Fig. 9, shows the qualitative results of the EDS analysis that the presence of energy peaks of 4.56, 6.18 and 0.27 Kev, respectively, indicate the presence of Ti, Fe and C elements in the nanocomposite. Also, in Table 3, the weight percent of each of the elements are provided. the weight percent of C, O, Ti and Fe elements were 5.48, 41.46, 46.84 and 22.6 percent respectively, which is approximately equal to the ratio of nanocomposite in the specified state (100/10/2.5: TiO2/Fe3O4/MWCNT).
Investigating the Magnetism of Nanocomposites
To determine the magnetic strength of the nanocomposite, the VSM analysis at room temperature was performed on Fe3O4 nanoparticles and TiO2/Fe3O4/MWCNT nanocomposites in weight ratio (2.5: 100: 10) and their magnetic curves were plotted, which can be seen in Figs. 10 and 11. The results show that Fe3O4 nanoparticles have a superparamagnetic property, and despite the low weight ratio of nanoparticles of Fe3O4 in the composition, the TiO2/Fe3O4/MWCNT nanocomposites were well-magnetized.
Photocatalysis activity of nanocomposites
To evaluate the photocatalytic degradation by the nanocomposite, the dye degradation in the specified conditions was investigated by Fe3O4/TiO2 nanocomposite and TiO2 nanoparticles, as shown in Fig. 12. The dye degradation efficiency by Fe3O4/TiO2/MWCNT was much higher than Fe3O4/TiO2 nanocomposites and was equal to the efficiency by TiO2 nanoparticles alone and slightly different in the elimination process.
An increase in the dye degradation efficiency by the nanocomposite could be due to the reduction of the energy gap caused by the MWCNT pegging to the TiO2 nanoparticles or the scavenger role of MWCNT and accumulation of electrons on MWCNT particles and the prevention of recombination of the free hole in TiO2 .
According to Fig. 12, we could say that Fe3O4/TiO2/MWCNT nanocomposites have higher initial absorption than TiO2 nanoparticles alone, which can be due to the presence of MWCNTs that is an excellent absorbent of organic and inorganic materials .
Also initially, the start of the light emission to the reactor photodegradation rate by the TiO2 nanoparticles is more than the Fe3O4/TiO2/MWCNT nanocomposite, but after a while, this difference disappears, partly because of the dye absorbed on the surface of the Fe3O4/TiO2/MWCNT nanocomposite which reduces photon absorption.
To further investigation of the photocatalytic activity of the nanocomposite, degradation of other dyes including acid blue 19, reactive red 77 and methyl orange were measured under optimal conditions; the results are shown in Fig. 13. The rate of degradation was measured after 4 hours from the onset of light exposure for acid blue 19, reactive red 77, and methyl orange equal to 21.25, 97.87 and 13.44 percent, respectively. The results show that the nanocomposite is ineffective in the degradation of Methyl orange dye.
To find out this issue, the principal properties of the dye, as well as the degradation of dye by another nanoparticle, were investigated. The results are shown in Fig. 14.
In a study by Dorraj et al., 2017 , the degradation of Methyl Orange under xenon light radiation and the use of TiO2 nanoparticles were investigated, the efficiency less than 10% was reported. Also, Zhang et al., , study about the degradation of Methyl Orange by nanocomposite P25 TiO2 (Degussa), the similar results and very low efficiency was reported again . In the investigation of the degradation efficiency of methyl Orange dyes by TiO2 and Fe3O4/TiO2, the degradation rate after 4 hours was 19.04 and 19.02%, respectively.
To the structural analysis of the methyl orange dye and its difference with another dye, the absorption spectra of acid red 14, acid blue 19, reactive red 77 and methyl orange were investigated. According to Fig. 15, its optical absorption was much higher than of acid red 14. Most of the photons are absorbed by dye. It will no longer be absorbed by photocatalyst to stimulate electrons This can be attributed to the low dye degradation by the nanocomposite and the total photocatalytic reaction.
Investigation of the mechanism of dye removal
Fig. 16. Shows a typical time-dependent UV-Vis spectrum of AR14 solution during photoirradiation. The absorption peaks, corresponding to dye, diminished and finally disappeared under reaction which indicated that the dye had been degraded. The UV–Vis spectrum of AR14 mainly consisted of three well-resolved bands. They were 514, 322 and 220 nm, respectively. The lowest energy absorption band (514 nm) was assigned to the n-Π* transition of (–N=N–) group and the peak observed in ultraviolet spectra absorption was assigned to the Π-Π* related to naphthalene rings bonded. According to Fig. 16. adsorption values during photocatalytic reaction decreased and after 4 and 6 hours, minor and Mainor peak have disappeared, respectively which can be attributed to the collapse of azo bonds due to oxidation.
Investigation of photocatalytic properties of nanocomposite in real conditions under sunlight
The energy from sunlight is recognized as a sustainable, clean and free source, and the benefits of this sustainable energy source are essential. One of the main goals of this research is the use of sunlight source for the degradation of water pollutants, so the rate and amount of pollutant removal under sunlight was investigated in real terms.
An experiment was conducted to investigate the efficiency of photodegradation of acid red 14 under sunlight on December 18 and 19, 2017, and in the geographical location of 3837845/51, 7219535/35 (Tarbiat Modares University, Tehran, Iran) for 8 hours and in optimal condition. According to the information extracted from the World Meteorological Site (https://www.wmo.int), the average cloudiness of the weather was 5 and the humidity was 36 percent. Also the intensity of sunlight was measured every hour, with an average of 280 W/m2. Results of the study of the removal efficiency of dye under direct sunlight are shown in Fig. 17. The dye removal efficiency after 6 hours under direct sunlight for TiO2/Fe3O4/MWCNT, TiO2/MWCNT, TiO2/Fe3O4, and TiO2 Nanocatalyst was 86.27, 76.88, 19.19 and 89.53 percent, respectively
The result indicates that, the high removal efficiency of the dye by the TiO2/Fe3O4/MWCNT nanocomposite under sunlight. Which confirms the economics of the process of removing pollutants from water with nanocomposites.
Reusability of TiO2/Fe3O4/MWCNT nanocomposite
Sustainability and reusability of a photocatalyst is a major factor in its applicability because it will be very effective at the final cost of sewage treatment. Reusability of TiO2/Fe3O4/MWCNT nanocomposites to evaluate the durability and stability under optimal conditions are presented in Fig. 18. According to the results after 3 cycles, there was no significant change in the efficiency of removing any of the dyes by the nanocomposite. For example, in the photocatalytic removal of AR14 after 4 hours, only 2.43% reduction was observed in the removal efficiency, but from period 4 to Then, the amount of reduction slightly increased after 4 hours compared to the first period, 7.79% reduction was observe in the dye removal efficiency, which could be due to various reasons, including loss of part of the nanocomposite during its recovery  or remaining a portion of the pollutant on active site of nanocomposite surfaces  . However, within 6 hours, light irradiation and in the fifth cycle 90.86% of the dye, was eliminated which demonstrates the high ability of nanocomposites to recycle and reuse, reducing the cost of removing dyes, solving the problem of its release in the environment, and hence increasing the chance of industrial use of nanocomposite. Also, in Fig. 19, the final efficiency of the removal of AR14 dye after 6 hours of light exposure as a result of 5 consecutive cycles is observed. Fig. 20 shows the difference between TiO2/Fe3O4/MWCNT and TiO2/MWCNT in sewage.
In this study, with a simple and inexpensive method, TiO2 nanoparticles could be magnetically extracted by superparamagnetic Fe3O4 nanoparticles. Also, the addition of carbon nanotubes prevented the reduction of photocatalytic activity of nanocomposite, so that dye degradation during 6 hours under sunlight in optimum condition by TiO2 alone, Fe3O4/TiO2 and TiO2/Fe3O4/CNT, was measured 89.53, 21.19 and 86., respectively. To investigate the application of nanocomposite in removal of others pollutants, the removal efficiency of MO, RR76 and AB19 dyes were measured 22.16, 99.51 and 66.02 respectively. Also, its reusability was investigated in 5 cycles that only 7.79% reduction in degradation efficiency was observed. According to the results, the stability of the solar source and the ability to recycle and reusability TiO2/Fe3O4/CNT nanocomposites was founded the nanocomposite is useful in removing non-biodegradable pollutants.
The study was supported by a grant from the Tarbiat Modares University (TMU). The authors have no conflicts of interest to declare.
CONFLICTS OF INTEREST
There are no conflicts to declare.