Photocatalytic removal of methylene blue using titania- and silica-coated magnetic nanoparticles

The scope of this investigation is the photocatalytic degradation performance of newly synthesized nanoparticles (NPs); namely; Fe3O4; Fe3O4@TiO2 and Fe3O4@SiO2. Non-thermal synthesis methods are used to synthesize the NPs and to explore the ferromagnetic properties of the photocatalysts. The synthesized NPs are characterized using TEM, XRD, FTIR, TGA, VSM, and surface area analysis techniques. The photocatalytic activities of Fe3O4 and Fe3O4@SiO2 NPs, put under solar irradiation, and Fe3O4@TiO2 NPs, put under UV irradiation, are examined. The efficiency in degradation of Methylene Blue (MB) pollutant is shown to be the best for Fe3O4@SiO2 NPs, then in Fe3O4 NPs, and lastly in Fe3O4@TiO2 NPs. The silica (SiO2) coating on Fe3O4 NPs significantly enhances the light absorption and is found to improve the MB degradation rate and the photoinduced charge generation and separation (i.e. it enhances the exciton lifetime). That makes the Fe3O4@SiO2 NPs promising candidates for organic pollutants removal in various environment-related applications.


Introduction
Amongst side effects of the industrial revolution is the release of huge amount of wastewater. Wastewater contains several hazardous substances that pollute the environment. For instance, the textile dyes are one of the largest group of chemicals being produced all over the world [1,2]. Namely, methyl red, methylene blue (MB), methyl orange and malachite green, that cause various kinds of human cancers, coming major challenging concerns for public health safty [3,4]. However, these dyes show high stability and resistance to biodegradation [5]. Consequently, several physical and chemical methods are undertaken to remove dyes from industrial effluents. Among the attempted methods, one can mention: adsorption, coagulation, flocculation, electrocoagulation [6][7][8] and photodegradation [9][10][11]. The methods of wastewater treatment aim usually to adsorb or to head off pollution; but only photocatalytic degradation can transfer organic wastewater into CO 2 and H 2 O without any intermediate [12,13]. Nonetheless, photocatalyst is collected from the solution by using a magnet, which is a promising green mechanism in environment technology. Hence, producing a new photocatalyst comprising magnetic property, in the form of core-shell structure is of emerging interest. Recently, Magnetite (Fe 3 O 4 ) nanoparticles (NPs) are discovered to be ideal materials as a core for a photocatalyst, because of many reasons: (i) its low toxicity, and bio-compatibility [14], (ii) high magnetization and high coercivity, (iii) relatively low Curie temperature [15][16][17][18], (iv) superior photons' absoption capability, and (v) strong bonds between the molecules due to the dipole-dipole interactions [19]. Furthermore, it is easy to control the mobility of Fe 3 O 4 NPs through the application of magnetic field.

Synthesis and characterization of magnetic photocatalysts
The Fe 3 O 4 NPs are prepared separately then coated with either a titania or silica layer. A chemical coprecipitation method is used to synthesize the Fe 3 O 4 NPs [36]. Typically, dissolve 0.02 Mole of FeCl 3 4 NPs are formed as a black precipitate. Then, add 11 ml of polyethylene glycol-4000 (PEG) solution dropwise for 1 h. PEG stabilizes the surface of Fe 3 O 4 NPs, thus NPs keep dispersed in the solution, for further coating with titania or silica. Furthermore, PEG plays positive role in the passivation of surfaces of Fe 3 O 4 NPs and provide covalent bonds to suppress and saturate the dangling bonds. Consequently, PEG removes the defect-related gap states and enhances the optical qualities which are, as a matter of fact, needed to enhance the photoreaction and photoconductivity effects [37][38][39][40]. Finally, collect the Fe 3 O 4 NPs by a magnet.
Regarding the coating process, TiO 2 is applied on the surface of Fe 3 O 4 NPs using a non-thermal method [41], Typically, disperse 10 mg of the as synthesized Fe 3 O 4 NPs in 25 ml ethanol by ultrasonication for 30 min. Then add TiCl 4 /ethanol diluted solution (0.5 ml/2 ml). Keep the stirring for 1 h, until the color of the mixture change to gray indicating the formation of Fe 3 O 4 @TiO 2 NPs. Finally, collect the Fe 3 O 4 @TiO 2 NPs by a magnet. On the other hand, SiO 2 is applied on the surface of Fe 3 O 4 NPs using a modified Stöber method [36]. In brief, disperse 100 mg of the as synthesized Fe 3 O 4 NPs in 75 ml ethanol by ultrasonication, add 6 ml of NH 4 OH and 200 μl Tetraethyl ortho-silicate dropwise for 2 h. Finally, collect the Fe 3 O 4 @SiO 2 NPs by a magnet.
The synthesized photocatalysts are characterized using several techniques including transmission electron microscopy (TEM), dynamic light scattering (DLS) by Zetasizer Nano ZS, Fourier transform infrared spectroscopy (FTIR), powder x-ray diffraction (XRD), thermogravimetric analysis (TGA) and UV-visible spectroscopy. Brunauer-Emmett-Teller (BET) method is used to measure the specific surface area of the dried photocatalyst powders. Barrett-Joyner-Halenda (BJH) method is used to measure the pore size distribution. SQUID Vibrating sample magnetometer (VSM) is used to measure the magnetic properties of the photocatalysts. Dynamic light scattering (DLS) technique used to measure the zeta potential of the photocatalysts in H 2 O suspension, at 25°C, in which Zeta Sizer Nano ZS machine is utilized.

Photocatalytic experiments
Kinetics adsorption is investigated in dark before the photocatalytic experiment, using 300 ml of Methylene blue (MB) dye solution, MB concentration is 70 ppm (mg L −1 ). 40 mg of the synthesized magnetic photocatalyst is added to MB. Thus, Fe 3 O 4 /MB, Fe 3 O 4 -TiO 2 /MB and Fe 3 O 4 -SiO 2 /MB containers are kept closed in darkness under continuous shaking. A sample is withdrawn every 20 min, photocatalyst is collected and detached from the sample by a magnet, and the remaining concentration of MB is measured by spectrophotometer.
As the adsorption-desorption equilibrium in the mixture of the photocatalyst and the MB dye is attained, photocatalytic experiments are carried out, under solar radiation in October 2017. The temperature and average solar light intensity at that time and site are 34°C and 5320 Wh/m/day, respectively. Samples are kept in closed glass containers and placed under sunlight, except for the Fe 3 O 4 -TiO 2 /MB sample, are placed under UV light of 250 nm wavelength. Then collected at several times. The MB concentration is calculated as the absorbance is monitored at 664 nm wavelength, using a UV-vis spectrophotometer. The experimental measurements are conducted in duplicate and the average of two values is recorded.

Results and discussion
3.1. Morphology and structure of the photocatalysts Figure 1 displays the TEM images of the Fe 3 O 4 NPs without coating and as they are coated with SiO 2 and TiO 2 layer. The results indicate that all Fe 3 O 4 , Fe 3 O 4 @TiO 2 , and Fe 3 O 4 @SiO 2 NPs have almost a spherical shape. Figure 1(a) shows that Fe 3 O 4 NPs have sizes in the range between 2 and 10 nm. Figure 1(b) shows nanoclusters of Fe 3 O 4 , are successfully coated with a thin layer of titania leading to core-shell structure of Fe 3 O 4 @TiO 2 NPs, having sizes in the range between 50 and 100 nm. Figures 1(d)-(f) shows that Fe 3 O 4 NPs clustered into bigger size than the previous particles (figures 1(b), (c)), then coated with a thin layer of SiO 2 leading to core/shell structure of Fe 3 O 4 @SiO 2 NPs with sizes in the range between 400 and 500 nm. Almost same particles' sizes distribution data are obtained using DLS technique, where the obtained average size of Fe 3 O 4 NPs, Fe 3 O 4 @TiO 2 , and Fe 3 O 4 -SiO 2 are 7.8 nm, 73 nm 475 nm, respectively, more details are in [36,41].
XRD patterns as 2θ in the range of 10°-80°for the Fe 3 O 4 NPs before and after the TiO 2 and SiO 2 coating are shown in figure 2. The XRD pattern of Fe 3 O 4 demonstrates sharp and intense diffraction peaks, since each plane of the NPs reflects the incident x-ray at certain scattering angle 2θ, each peak at around 2θ is corresponding to one lattice plane of Fe 3 O 4 as follow; 30°for (2 2 0), 36°for (3 1 1), 43°for (4 0 0), 53°for (4 2 2), 57°for (5 1 1), 63°f or (4 4 0) and 75°for (533) plane. Indicating a crystallized cubic inverse spinel structure of Fe 3 O 4 . The core/shell Fe 3 O 4 @TiO 2 NPs exhibits diffraction peaks as follow; 25.3°for (1 0 1), 37.8°for (0 0 4), 48.2°for (2 0 0), 54.2°for (1 0 5), 55.3°for (2 1 1), 69°for (116) and 70°for (220) plane of titania [42]. These peaks demonstrate that titania has anatase tetragonal structure, the existence of other few peaks with extremely low intensity are assigned to Fe 3 O 4 cores. The core/shell Fe 3 O 4 @SiO 2 NPs exhibits a broad diffraction peak at 10°-30°assigned to the amorphous SiO 2 [43,44], in addition to the existence of some peaks from Fe 3 O 4 cores with less intensity than     [46], verifying the titania-coated Fe 3 O 4 structure. Fe 3 O 4 @SiO 2 spectrum shows a strong band at 960 cm −1 attributed to Si-OH bond. The peaks at 797 cm −1 and 464 cm −1 resulted from vibrational modes in and out of plane bending of the Si-O-Si, respectively. The band at 570 cm −1 is ascribed to Si-O-Fe bond [47], verifying the silica-coated Fe 3 O 4 structure. Figure 4 shows the TGA curve of the Fe 3 O 4 NPs, which gives an approximation of the weight loss of the sample with temperature, to evaluate the mass percentage of PEG on the surface of Fe 3 O 4 NPs. The first decomposition stage occurs at 30°C up to ≈150°C, where a steep weight loss of 2% is observed, owing to the release of any remaining water and alcohol on the particles' surface. The second stage shows a major weight loss with 6% change over a wide range of temperature between 150°C and 450°C. The weight loss at this stage is due to the PEG decomposition. which indicates a very thin layer of PEG on Fe 3 O 4 NPs surface. At temperatures in  the range of 450°C-700°C polyethylene glycol is completely dehydrated leading to about 1% weight loss, then no further weight reduction is noticed because of the high thermal stability of Fe 3 O 4 . Figure 5(a) shows UV-vis light absorption spectra of the prepared photocatalysts. Obviously, Fe 3 O 4 and Fe 3 O 4 @TiO 2 NPs exhibite absorption in UV light region. However, Fe 3 O 4 @SiO 2 exhibite absorption in the full light spectrum (i.e., no clear absorption peak). According to the band theory, the optical band gap can be estimated using the Tauc's formula [48]: Where α is the absorption coefficient (cm −1 ), B is a characteristic constant and n is a parameter identifying the electronic transition type (e.g., n=2 for direct transition), hν is the photon energy (eV), E g is the energy band gap (eV). The E g can be calculated by extrapolating the straight line fit of the graphical representation of (αhν) 2 versus hν to hν-axis, as shown in figures 5(b), (c). The linearity of the curves usually confirms the existence of direct band gap transitions.
The measureable E g for Fe 3 O 4 NPs and Fe 3 O 4 @TiO 2 NPs are 3.3 eV and 4.1 eV, respectively, as shown in figures 5(b), (c). The E g values of these NPs are much larger than the E g of their bulk, which is attributed to the quantum confinement effects, thus, as crystal size decreases the band gap energy increases. Additionally, a very low amount of titania on the surface of Fe 3 O 4 NPs can lead to the difference in energy levels of the energy molecular quantum states within Fe 3 O 4 . In addition to quantum-confinement effects, PEG and titania possess energy levels close to the interface states.

Surface area analysis
Specific surface areas of the photocatalysts are measured using the BET method, by measuring the amount of the adsorbed N 2 gas on the surface of the NPs [49]. Figure 6(a) shows the N 2 adsorption/desorption isotherms of the Fe 3 O 4 , Fe 3 O 4 @TiO 2 and Fe 3 O 4 @SiO 2 NPs. The isotherms exhibit type IV, indicates the mesoporous (i.e. 2-50 nm pore size) character of the synthesized photocatalysts. Figure 6(b) shows their corresponding pore-size distributions which are measured using the BJH method [50]. The values of BET surface area, BJH cumulative pore volume and average pore size are shown in table 1.
Fe 3 O 4 @SiO 2 NPs shows the highest surface-to-volume ratio and largest pore size whereas the Fe 3 O 4 @TiO 2 shows the least surface-to-volume ratio with poor pore size. Furthermore, Fe 3 O 4 NPs exhibit surface area of 111 m 2 .g −1 which is less than the surface area of the silica-coated Fe 3 O 4 (138 m 2 .g −1 ). These results confirm that the silica coating increases the specific surface area of the NPs which is, in turn, beneficial for the photodegradation process.  However, it is evident that the M s values at 300 K is less than the M s at 2 K due to thermal agitations of the NPs' magnetic moments. Furthermore, the M s values are considerably reduced compared to the M s for bulk Fe 3 O 4 (around 90 emu g −1 at 300 K) [51]. The decrease in M s of magnetic NPs is due to disorder of the surface spin and finite size effects. The NPs have much higher surface-to-volume ratio than the bulk material. That leads to the residence of a huge fraction of the total number of atoms on the surface of the particle. However, the environments of the atoms in core are different than those on the surface of the particle. The surface of the particle can experience several defects resulting in spin canting or disordered surface spins and spin-glass behavior, due to changes in the (i) atomic coordination, (ii) lattice disorder, (iii) dangling bonds [52]. Thus, the net magnetization of the NPs are smaller than that of the bulk material. Figure 7(b) shows the ZFC and field-cooled (FC) magnetization-temperature (M-T) measurements, conducted on the NPs. In the FC measurements, the NPs are heated up to 300 K, then the temperature is cooled again from 300 K to 2 K at a constant applied magnetic field of 50 Oe. The ZFC/FC curves of the Fe 3 O 4 and   Fe 3 O 4 @TiO 2 show that as the temperature increases from 0 to 300 K, the ZFC magnetization increases first and then decreases after reaching a maximum at 300 K, which corresponds to the blocking temperature (TB). This result further proves that the as-synthesized NPs show a superparamagnetic behavior at room temperature. Whereas the FC magnetization increases as the temperature decreases to 200 K. A small dep is noticed around 200 K, then the magnetization becomes nearly constant as the temperature decreases to 2 K, which is an evident on the existence of super or surface spin-glass structure [52][53][54][55][56]. The super-spin-glass behavior is resulted from the large interparticle interactions whereas the surface spin-glass behavior can be caused by frozen disordered surface spins [54,55]. M-H and M-T curves show that the magnetization values of Fe 3 O 4 @TiO 2 , considerably smaller than those of Fe 3 O 4 because of the nonmagnetic titania coating which surrounding the Fe 3 O 4 cores. Additionally, this reduction of magnetization in the coated sample can be attributed to super-spin-glass structures; each agglomerate of Fe 3 O 4 NPs of sizes around 7 nm is coated with TiO 2 . This huge number of the Fe 3 O 4 NPs (core) leads to a strong interparticle interaction which forms super-spin-glass structures. Similar magnetic properties for Fe 3 O 4 @SiO 2 are expected, since the Fe 3 O 4 NPs are synthesized the same way, and the nonmagnetic silica coating is surrounding an agglomerate of a huge number of Fe 3 O 4 NPs (cores). However, the as-synthesized NPs exhibit strong responsive to an external magnet in which they can be collected easily from the solution, washed and reused for another purification cycle. Once the magnet is removed, the NPs do not retain any net magnetization, indicating the superparamagnetic characters.

Adsorption performance
Adsorption kinetics is a more pronounced mechanism in the absence of light. Adsorption of a model pollutant

Photocatalytic activity
As the adsorption-desorption equilibrium is attained among the particles and the dye, the photoactivity of the as synthesized NPs is investigated by photodegradation of MB. After reaching the equilibrium adsorption state in the dark. Figures 8(a) Figure 8(d) shows that the concentration of MB decreases significantly by about 97% within 140 min.
The performance of Fe 3 O 4 @TiO 2 is evaluated under solar irradiation for 5 h. However, unnoticeable change in the solution colour is observed, hence insignificant change in the MB concentration is obtained. That can be attributed to the fact that TiO 2 (on the surface) is a wide band gap semiconductor that can be excited by light of high energy (i.e., UV, short wavelength). Furthermore, the reduction of the particles' size to the nanoscale increases the band gap value, consequently, the elaboration of the excitation energy, due to quantumconfinement effects. When the band gap becomes larger, electrons require more energy to be excited to the conduction band.
Photocatalytic experimental results of MB degradation on Fe 3 O 4 , Fe 3 O 4 @TiO 2 and Fe 3 O 4 @SiO 2 NPs are fitted to three models, to understand the controlling mechanism of the degradation processes as shown in figure 9. (i) The zero-order kinetic model can be expressed by [57] C C Where, C o is the initial MB concentration, C t the concentration of MB at time t, and k is the degradation rate coefficient. C t /C o vs t plot gives a straight line, and its slope is −k/C o , as shown in figure 9(a).   NPs as a photocatalyst is best described by the second order model, with k value of 0.0063 min −1 being the least, which indicate that the degradation rate of MB is boosted by using Fe 3 O 4 @SiO 2 NPs. The removal efficiency (E) or de-colouration rate is calculated using the following formula [58]: Under solar light irradiation, Fe 3 O 4 photocatalyst could decompose 89% of MB after 140 min irradiation time. When Fe 3 O 4 NPs coated with silica layer and used as a photocatalyst, an increase of MB photodegradation efficiency is observed (97%). However, as Fe 3 O 4 NPs coated with titania layer the exhibited MB photodegradation efficiency is 56%, under UV irradiation as shown in figure 10(a).
The photocatalytic experiments reveal that the silica coating could importantly improve the photocatalytic performance of Fe 3 O 4 photocatalyst. The Fe 3 O 4 @SiO 2 yield the highest MB degradation efficiency, which is about 1.7 times the degradation efficiency of the Fe 3 O 4 @TiO 2 . TEM images show that Fe 3 O 4 nanocluster surface is completely shielded with titania layer. Which is confirmed by the high reduction of the Fe 3 O 4 XRD peaks in Fe 3 O 4 @TiO 2 pattern ( figure 3). Thus, the photodegradation activity is mainly obtained by TiO 2 layer that absorbs UV light. Conversely, in Fe 3 O 4 @SiO 2 photocatalyst, the photocatalytic degradation is attained by the core of the particles (Fe 3 O 4 ) and the silica coating serve as a protective layer to enhance the stability of the photocatalyst in the aqueous solution and, furthermore, to enhance the optical absorption.
Recyclability is the main advantage of the magnetic photocatalysts. The reusability evaluation is carried out for four times as shown in figure 10(b). The photocatalytic efficiency of the recycled Fe 3 O 4 NPs after four cycles reduced from 89% to 62%, this is can be attributed to the Fe 3 O 4 NPs' tendency to agglomerate, due to the magnetic dipole interaction, that leads to the reduction of the surface area, and the degradation efficiency. However, the recycled Fe 3 O 4 @TiO 2 and Fe 3 O 4 @SiO 2 NPs showed insignificant change in the degradation efficiency for four cycles, demonstrating their stablity and effectiveness in the photodrgradation of organic pollutants from water. The high effeciency obtained by Fe 3 O 4 @SiO 2 for four cycles indicates that Fe 3 O 4 @SiO 2 is a promising photocatalyst.
Furthermore, the stability of the photocatalysts are checked by measuring the surface charge of the NPs. The measured zeta potential values are about +7, −43 and −54 mV for Fe 3 O 4 , Fe 3 O 4 @TiO 2 and Fe 3 O 4 @SiO 2 NPs, respectively. The small positive charge on the surface of Fe 3 O 4 indicates their affinity to agglomerate into clusters and their low stability. Whereas, the huge negative charge on the silica and titania surface generates coulomb repulsion among the particles to boost the inter-particle spacing. Leading to a high stable and homogenous colloidal solution. Thus, the high reusability of of the Fe 3 O 4 @TiO 2 and Fe 3 O 4 @SiO 2 NPs is promising. Figures 11(a), (b) shows MB aqueous solution before and after solar irradiation, respectively. The fading of the dark blue color indicates the degradation of the MB dye, where the Fe 3 O 4 @SiO 2 magnetic photocatalysts are collected by the external magnet to be used for another purification cycle. Figure 8(c) demonstrates the mechanism of MB degradation, using Fe 3 O 4 @TiO 2 as a photocatalyst. First, titania layer absorbs photons (UV radiation) with energy larger than its bandgap, so electron-hole pairs (excitons) are generated. Then the electron and hole in the exciton separate apart, the free electrons transfer towards the conduction band and the holes to the valence band at the interface (PEG layer), where electric feild exists, due to depletion zone of core-shell interface. Thus, the existence of an extremely thin layer of PEG with a large bandgap insulating material, Consequently, e − and h + transfer to the photocatalyst surface. The e − reduce O 2 to superoxide radicals O .

( )
Simultaneously, the h + oxidize H 2 O to form hydroxyl radicals (OH · ), that works on MB degradation or h + directly oxidize MB. The formed reactive species, O , 2 -· OH · , and h + initiate the redox reactions and degradation of MB into CO 2 , H 2 O, or inorganic ions. As shown in figure 8(c). Finally, the photocatalyst is extracted from the media by an external magnet [46,59].
Similar mechanism of MB degradation using Fe 3 O 4 @SiO 2 as photocatalyst, the silica layer enhances the transfer of the photoinduced charge to the photocatalyst surface, and improves the light harnessing ability of Fe 3 O 4 , which could be beneficial for photocatalytic performance. This mechanism is most effective when photocatalytic oxidation is conducted in the presence of water (as it is done in our experiment). Furthermore, photodegradation is boosted by enhancing the transportation of e − and h + , to increase the effectiveness of charge separation and to increase the lifetime of the photo-generated excitons.

Conclusions
Magnetic photocatalysts, namely Fe 3 O 4 @TiO 2 NPs and Fe 3 O 4 @SiO 2 NPs, are synthesized using a non-thermal method and a modified Stöber method, respectively. Their photocatalytic degradation performance is investigated using MB dye. The as synthesized NPs have a spherical shape with core/shell characteristic, and the typical size of Fe 3 O 4 @TiO 2 NPs (73 nm) is smaller than Fe 3 O 4 @SiO 2 NPs (475 nm). TiO 2 and SiO 2 are finecoated on the surface of Fe 3 O 4 NPs with thickness of few tens of nanometers. However, TiO 2 layer is thicker than SiO 2 layer, which affects their photodegradation mechanism. The as synthesized NPs show perfect superparamgnetic behavior, hence they are believed to be promising for wide range of engineering applications. The photocatalytic activity results show that Fe 3 O 4 @SiO 2 NPs, under solar irradiation, exhibit a distinguished degradation efficiency of MB as high as 97%. This later efficiency is higher than Fe 3 O 4 NPs (89%), and is much higher than the degradation efficiency obtained by Fe 3 O 4 @TiO 2 NPs (56%) under UV irradiation. The improved photocatalytic performance of Fe 3 O 4 @SiO 2 NPs can be attributed to the increased surface area, enhanced optical absorption, the effective generation of excitons in the Fe 3 O 4 core, dissociation of the photogenerated charges, then the separation of the e − and h + at the core/shell interface, and the enhancement of exciton's life time. Moreover, the kinetics of degradation of MB for Fe 3 O 4 @SiO 2 NPs photocatalyst is fitted by the zero-order model, and exhibit the highest degradation rate of MB. Therefore, the Fe 3 O 4 @SiO 2 NPs can be extensively used in several environmental applications for organic contaminant treatment.