Synthesis of MgAC-Fe3O4/TiO2 hybrid nanocomposites via sol-gel chemistry for water treatment by photo-Fenton and photocatalytic reactions

MgAC-Fe3O4/TiO2 hybrid nanocomposites were synthesized in different ratios of MgAC-Fe3O4 and TiO2 precursor. X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray fluorescence spectrometry (XRF), electron spin resonance spectrometry (ESR), Brunauer-Emmett-Teller (BET), photoluminescence (PL), and UV photoelectron spectroscopy (UPS) were used to characterize the nanocomposites. The increase of MgAC-Fe3O4, in the hybrid nanocomposites’ core-shell structure, led to the decrease of anatase TiO2 peaks, thus reducing the photo-Fenton and photocatalytic activities. According to the obtained data, MgAC-Fe3O4 [0.05 g]/TiO2 showed the best photo-Fenton and photocatalytic activities, having removed ~93% of MB (photo-Fenton reaction) and ~80% of phenol (photocatalytic reaction) after 20 and 80 mins, respectively. On the pilot scale (30 L), MgAC-Fe3O4 [0.05 g]/TiO2 was completely removed after 27 and 30 hours by the photo-Fenton and photocatalytic activities, respectively. The synergistic effect gained from the combined photo-Fenton and photocatalytic activities of Fe3O4 and TiO2, respectively, was credited for the performances of the MgAC-Fe3O4/TiO2 hybrid nanocomposites.

www.nature.com/scientificreports www.nature.com/scientificreports/ Fe 3 O 4 NPs have attracted interest due to their considerable magnetic behavior and strong pin polarization 12 . Many methods of Fe 3 O 4 -TiO 2 composite synthesis, such as sol-gel, co-precipitation, hydrothermal, sonochemical, and templates routes, have been reported in the literature 12,13 . Incorporation of Fe 3 O 4 NPs into a TiO 2 matrix can block NP aggregation and improve the durability of catalysts 14,15 . However, due to the small band gap of Fe 3 O 4 NPs (0.1 eV), Fe 3 O 4 -TiO 2 composites will increase the rate of electron-hole pairs recombination, with the result that photocatalysis is usually unchanged or even diminished relative to pure TiO 2 NPs 12 . To overcome this problem, Zheng et al. demonstrated that special structures such as core-shell microspheres in Fe 3 O 4 -TiO 2 composites can delay the recombination of photo-induced electrons 12 ; other researchers have used noble metal (Au or Ag) or rare elements (e.g. Eu) as electron traps to enhance electron-hole separation and facilitate electron excitation by creating a local hole in the electrical field 13,16,17 . In contrast, He et al., after preparing Fe 3 O 4 -TiO 2 core-shell NPs, indicated that Fe 3+ released from Fe 3 O 4 can be doped into TiO 2 NPs to decrease electron-hole pair recombination and thus increase the photocatalytic performance of Fe 3 O 4 -TiO 2 core-shell NPs under visible light 18 . One remarkable report in this research field is that of Sun et al., who found that a small number Fe 3 O 4 NPs loaded onto TiO 2 NPs (Fe/TiO 2 ratio: 1/200) could enhance the degradation of organic dye (Reactive Brilliant Red X3B). They attributed the improved photocatalytic performance of Fe 3 O 4 -TiO 2 to the synergistic contribution of the photocatalytic and Fenton reactions in the composite 19 .
2-D materials have been attracted and extended their applications due to their unique properties 20 . Photocatalytic materials have also been developed based on 2-D materials such as graphene 21 .
From its first introduction by Mann et al., magnesium aminoclay (MgAC), which is also another types of 2-D materials, has attracted interest in its propylamine functionalities, structures, and high dispersity in water [22][23][24] . Use of MgAC's high adsorption utility for heavy metal and organic dye removal has been reported 25 . Besides being utilized as a single agent, MgAC has been conjugated with other materials for environmental-treatment purposes. For example, MgAC has been coated with nZVI for removal of perfluorinated compounds 26 and chromium 27 .
In previous work, we conjugated MgAC with TiO 2 NPs and Fe 3 O 4 NPs by different methods for environmental-treatment 28,29 and microalgae-harvesting purposes 30,31 . The presence of MgAC in composites was demonstrated to improve the photocatalytic behavior of pure TiO 2 NPs 28 as well as the photo-Fenton behavior of Fe 3 O 4 NPs 29 . Based on these successful preliminary results, in the present study, we synthesized MgAC-Fe 3 O 4 / TiO 2 hybrid composites in order to exploit the advantages of both Fe 3 O 4 and TiO 2 in environmental-treatment applications.  4 2− , and HCO 3 − existed in the aqueous solution. Among these three inhibition agents, SO 4 2− and HCO 3 − had stronger inhibitory effects on the photo-Fenton reactions of the hybrid samples than HA, especially when these anionic and organic substances were presence at high concentrations in www.nature.com/scientificreports www.nature.com/scientificreports/ water (Fig. S2) 34 . The presence of these intermediate products in the reactor after stopping photocatalytic reaction could be supported by results of TOC below 35 . The hybrid nanocomposites could not be regenerated by using simple washing methods 34 . In the future works, the more suitable regeneration methods should be considered to improve the recycling performances.

Results and Discussion
photo-fenton and photocatalytic mechanisms. The photocatalytic mechanism of MgAC-Fe 3 O 4 /TiO 2 was attributed to the presence of anatase TiO 2 in the samples, as follows: OH • and • O 2 − , which are produced through the above chain reactions, will degrade pollutant molecules via oxidation and a reduction reaction process, respectively. To investigate the contribution of these ROS to the photocatalytic activity of MgAC-Fe 3 O 4 /TiO 2 , different scavengers have been used: isopropanol for •OH, methanol for both h + and •OH, and p-benzoquinone for • O 2 − 36,37 . In our system, highly reactive •OH was the main actor in the degradation of the pollutant materials, rather than the valence band h + and the conductive band e − (Fig. S4) 36 .
On the other hand, the photo-Fenton activities of the MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites were extremely high, due to the presence of H 2 O 2 in the reaction: As noted just above, pollutants are degraded mainly by • OH, which is generated in a photo-Fenton-like process: direct photolysis of H 2 O 2 , and photocatalytic oxidation of adsorbed H 2 O through holes in the valence band of the TiO 2 surface. Additionally, photo-induced electrons generated from TiO 2 can cause reduction of Fe 3+ to Fe 2+ 38 . The photocatalytic and photo-Fenton activities of MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites can be additionally supported by the presence of MgAC, which, with its high adsorption utility, can bring reactants to the surfaces of photo-Fenton agents 24 www.nature.com/scientificreports www.nature.com/scientificreports/ Photoluminescence spectra could be used to explain the photocatalytic mechanism 39 . From photoluminescence spectra (Fig. S5), it is clearly seen that, due to the presence of TiO 2 in the hybrid composites, the separated electron and holes were kept longer in excited state than original MgAC-Fe 3  widely investigated on the batch scale. However, for the purposes of industrial application, these materials need to be up-scaled to pilot-scale reactors. There are a variety of pilot reactors that have been introduced in the literature 26,27 . In this study, we used a design of pilot reactor that has been introduced in previous reports 28,29 . The design and photos of the pilot reactor are presented in Fig. S6.
Based on the batch-scale results discussed above, MgAC-Fe 3 O 4 [0.05 g]/TiO 2 was mass produced for testing of photocatalytic and photo-Fenton performances on the pilot scale. For photocatalytic performance testing, 30 g of MgAC-Fe 3 O 4 [0.05 g]/TiO 2 was added to the 30 L pilot reactor to obtain a dosage of 1 g/L (similarly to the batch-scale study). After 30 hours, phenol had been nearly completely removed from the aqueous solution at a constant rate of 0.032 (h −1 ) ( Fig. 2a and Table 2). The extended reaction time might be attributable to the heaviness of the materials, which is quickly self-precipitated. However, this phenomenon makes MgAC-Fe 3 O 4 [0.05 g]/ TiO 2 easily recoverable after the reaction. In this study, ~80% of the materials was recovered from the reactor by simply stopping the reactor and allowing self-precipitation to occur for 24 hours without any external force (Fig. S7). To shorten the reaction time, we added 50 mL of H 2 O 2 to obtain ~15 ppm peroxide in the reactor while reducing the dosage of MgAC-Fe 3 O 4 [0.05 g]/TiO 2 from 1 g/L to 0.5 g/L. Under this photo-Fenton condition, phenol was completely removed after 27 hours at a constant rate of increase to 0.036 (h −1 ) ( Fig. 2b and Table 2). It should be noted that in this paper, we present just the preliminary results; the optimal concentrations of photocatalytic materials and H 2 O 2 should be discussed in future work investigating the effect of tap water on degradation rate. There is also a requirement for continual reactor upgrading to prevent quick sedimentation in the photocatalytic reaction and, thereby, improve the performance of MgAC-Fe 3 O 4 [0.05 g]/TiO 2 on the pilot scale.
For phenol of 500 ppm, after 48 h exposure, the LC 50 was 6.39 mg/L (5.36-7.53, 95% CI). At the concentrations of 0, 2.5, 5, 7.5, 10, and 20 ppm, the mortality rates were 0, 15, 20, 50, 80, and 100%, respectively (Fig. S1b). Meanwhile, the treated samples showed a mortality rate of ~10% (Fig. S1c). We suspected that the remnant toxicity had come from the leakage of Fe 3+ ions from the hybrid nanocomposites or the phenol intermediate products after the reactions. So, we conducted both ICP and TOC experiments to investigate the reason behind the toxicity of the treated samples. The ICP results showed that the leakage of iron ions after stoppage of the reaction was negligible (~60 ppb after treatment, lower than the standard 300 ppb for drinking water according to the WHO) 42 . However, the TOC experiments showed that the organic carbon concentration was still very high after the reaction (~50-55%); thus, the toxicity could be attributed to the residual toxic phenol intermediate products (Fig. S8). Therefore, it is necessary to extend the reaction until intermediate products are completely removed, not to the phenol concentration of zero 36 .    (Fig. S10). These results were similar with Sun et al. 19 . Also, the XPS analysis showed that, for MgAC-Fe 3 O 4 [0.05 g]/TiO 2 , besides the peaks at ~532, ~285, ~102, and ~52 eV belonging respectively to O 1 s, C1s, Si 2p, and Mg 2p (Fig. S11a) 27,43 , there were additional peaks at ~463 eV and ~457 eV that were attributed to the Ti 2p 1/2 and Ti 2p 3/2 of Ti 4+ states of stoichiometric TiO 2 , respectively (Fig. S11b) 44 . The Fe 2p 1/2 and Fe 2p 3/2 peaks of Fe 3 O 4 existed at ~723.9 and ~710.2 eV, respectively (Fig. S11c) 45 .
According to the SEM results, MgAC-Fe 3 O 4 [0.05 g]/TiO 2 had an aggregated form, with a diameter ranging between 32.82 and 201.80 nm (Fig. 4a). The morphology of MgAC-Fe 3 O 4 [0.05 g]/TiO 2 was further investigated by TEM and energy-dispersive X-ray mapping analysis (EDX). In the XRD, TEM, and EDX results, where MgAC-Fe 3 O 4 was uniformly distributed in the TiO 2 matrix, MgAC-Fe 3 O 4 /TiO 2 showed a core-shell-like structure, MgAC-Fe 3 O 4 playing the role as the core material and TiO 2 that of the out-layer material (Fig. 4b,c). Lattice fringe spacing of 0.253 nm, belonging to the (311) plane of the Fe 3 O 4 NPs, and 0.355 nm of TiO 2 NPs in HR-TEM image confirmed the presence of these particles in the hybrid nanocomposites (Fig. 4d) 46 (Table S1). XRF and XRD confirmed the effects of MgAC-Fe 3 O 4 loading on the photocatalytic activities of the MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites. Sun (2018) indicated that a ratio between Fe and TiO 2 of 1:200 could enhance the degradation of organic dye, increase the loading amount of Fe 3 O 4 and, thus, decrease the photocalytic activity 19 . The amounts of MgO and SiO in the XRF analysis and of Mg, Si in the EDX mapping analysis were attributed to the presence of amorphous MgAC in the hybrid nanocomposites ( Fig. 5c and Table S1).
As for the ESR spectrum, the two main signals at ~2.0 and ~4.3 could be assigned to different Fe(III) sites 48   www.nature.com/scientificreports www.nature.com/scientificreports/ atoms; the effective positive charge on Fe and Ti is increased, and the effective negative charge on O is decreased. In other words, the electron density around Fe and Ti atoms is decreased and the shielding effect is weakened, which results in increased binding energy.
From the obtained UPS spectra (Fig. S13) (Table S2). It was noted that MgAC-TiO 2 was synthesized by using MgAC and titanium butoxide (TB) as precursors in ethanol media, as previously reported 28 . The decrease of work function could facilitate electron emission and narrow the energy band gap 55 . It was clearly seen that the work function of MgAC-Fe 3 O 4 was lower than that of Fe 3 O 4 , which is around 3.7 eV in the literature 56 ; this reduction of work function could be explained by the higher photo-Fenton activity of MgAC-Fe 3 O 4 than that of commercial Fe 3 O 4 , as was demonstrated in a previous report 29 . The presence of TiO 2 on the surface of MgAC-Fe 3 O 4 continued to decrease the work function, the lowest number belonging to MgAC-Fe 3 O 4 [0.05 g]/TiO 2 (1.87 eV, optimal sample), which was yet larger than that of MgAC-TiO 2 (1.56 eV), which showed photocatalytic activation under visible light activation 28 . It could be concluded that, in general, the photocatalytic activity of the MgAC-Fe 3 O 4 / TiO 2 hybrid nanocomposite is higher than that of MgAC-Fe 3 O 4 (which has only photo-Fenton activity) but lower  www.nature.com/scientificreports www.nature.com/scientificreports/ than that of MgAC-TiO 2 (which also has photocatalytic activity under visible light irradiation). The advantages of MgAC-Fe 3 O 4 /TiO 2 relative to MgAC-TiO 2 come from its recycling utility. The difference of photocatalytic activity between the two might come from their respective synthesis processes and material structures. In the case of MgAC-TiO 2 , nitrogen from the amino-functional groups of MgAC can be doped into the TiO 2 structure via a calcination process, thus inducing photocatalytic activity under visible light 57 . However, in the case of MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites, the nitrogen element is removed by the process of the synthesis of MgAC-Fe 3 O 4 , as suggested above.
Additionally, the surface areas, pore sizes and pore volumes of the materials were investigated to characterize their surfaces. As can be seen in Table 3, the surface area, pore size and pore volume of the MgAC-Fe 3 O 4 sample were the lowest (surface area: 34.790 m 2 /g; pore size: 3.639 nm; pore volume: 0.0356 cm 3 /g). In the literature, the adsorption capacity of Fe 3 O 4 is lower than that of TiO 2 19 . In the present study, the MgAC-TiO 2 sample had the largest surface area, pore size and pore volume (surface area: 87.822 m 2 /g; pore size: 7.595 nm; pore volume: 0.1510 cm 3 /g). Surprisingly, for the MgAC-  Table S3), due to its higher production efficiency (1 g MgAC can produce only 1.08 g of MgAC-TiO 2 , whereas 1 g MgAC-Fe 3 O 4 can produce 4.4 g of MgAC-Fe 3 O 4 [0.05 g]/TiO 2 ). Another advantage of MgAC-Fe 3 O 4 /TiO 2 is its quick sedimentation, which enables ~80% of materials to be recovered after 24 hours of self-precipitation. Moreover, with their ferromagnetism was remained after the synthesis process, these hybrid nanocomposites show their recycling potential. However, the remnant toxicity of the treated sample against Daphnia magna indicated the presence of residual phenol intermediate products. The comparison of our study with some remarkable researches in the literature was briefly summarized in the Table 4.
For industrial-scale application purposes, the toxicity potential of MgAC-Fe 3 O 4 /TiO 2 as well as its photo-Fenton and photocatalytic mechanisms should be more thoroughly investigated. Additionally, it is necessary to find a way to induce the photocatalytic activity of such hybrid nanocomposites under visible light.

Synthesis of magnesium aminoclay (MgAC).
For preparation of MgAC, 1.68 g of MgCl 2 · 6H 2 O was dissolved in 40 mL of ethanol. Then, 2 mL of 3-aminopropyltrielthoxylane (APTES) was added and stirred for 8 hours to form a white suspension. The resultant suspension was then centrifuged and washed with ethanol (3 × 50 mL) before being dried at 40 °C and ground into powder 58 .

Synthesis of magnesium aminoclay-iron oxide (MgAC-Fe 3 o 4 ) nanocomposites.
A total of 0.7 g of MgAC was mixed with 3 g of FeCl 3 · 6H 2 O in 40 mL of DI water, to which mixture 10 mL of NaOH 10 M was added. The solution was stirred for 12 hours and then centrifuged and washed with DI water (3 × 50 mL) and dried at 60 °C to form a brown solid. The brown products were ground into powder and calcinated at 500 °C for 3 hours in a furnace (FU-100TG, Samheung Energy, Korea) under 4% H 2 /Ar (flow rate: 0.15 L/min) to produce MgAC-Fe 3 O 4 nanocomposites 29 .
Synthesis of magnesium aminoclay-iron oxide/tio 2 (MgAC-Fe 3 o 4 /tio 2 ) hybrid nanocomposites. In Table 3. BET surface areas, pore sizes and pore volumes of samples in this study.
www.nature.com/scientificreports www.nature.com/scientificreports/ The degradation rate of the organic compounds was determined by the equation where C is the concentration of MB at time (t), C o is the initial MB concentration, and k is the pseudo-first-order rate constant (min −1 ) 19 . The recycle usage experiments were conducted to check the stability of materials. After photocatalyst materials were separated from degraded solution, they were washed with DI water and ethanol, then dried in the oven at 60 °C for 12 hours. Then the materials is ready for another photocatalytic experiments. This method was repeated for 5 times 53 . preliminary evaluation of photocatalytic and photo-fenton performances of MgAc-fe 3 o 4 / tio 2 hybrid nanocomposites on pilot scale. The photocatalytic performances of the MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites were tested using the systems introduced in a previous report 28 . Briefly, MgAC-Fe 3 O 4 / TiO 2 hybrid nanocomposites were loaded into a reactor (90 cm (width) × 30 cm (depth) × 60 cm (height)) containing 30 L of tap water contaminated with phenol at 3 ppm and stirred with 3 stirrers (GGM speed control motor, Korea) at 145 rpm. After obtainment of equilibrium adsorption, 18 × 365 nm wavelength UV lamps (light intensity: ~610 µW/cm 2 , 65 cm × 3 cm) were turn on. After an interval of 3 hours, 40 mL of treated water was withdrawn, and the remaining concentration of phenol was determined by HPLC. For evaluation of the photo-Fenton  www.nature.com/scientificreports www.nature.com/scientificreports/ performances of the MgAC-Fe 3 O 4 /TiO 2 hybrid nanocomposites, H 2 O 2 was supplied after obtainment of equilibrium adsorption.
For an ecotoxicity test, Daphnia magna was incubated in a 16 h light/8 h darkness cycle at 21 ± 1 °C with M4 medium prepared according to OECD Test Guideline 202 (OECD, 2014) 60 . The culture medium was replaced daily, and Chlorella was fed once a day. The Daphnia test was carried out according to OECD Test Guideline 202, and young daphanids aged less than 24 hours were collected and exposed to the test materials for 48 hours after Chlorella feeding for 2 hours. In the case of 500 ppm phenol feedstock, 5 daphanids were exposed to 25 mL of 0, 2.5, 5, 7.5, 10, and 20 ppm phenol to determine the LC 50 (Fig. S1a). Ecotoxicity tests of photo-Fenton-and photocatalytic-treated samples against Daphnia magna also were performed. The toxicity of the photo-Fenton and photocatalytic samples also were investigated, first, via inductive coupled plasma atomic emission spectroscopy (ICP-AES; Optima 7300 DV, Pelkin Elmer, USA) for leakage of iron ions and, second, via total organic carbon (TOC, Vario TOC Cube, Elementar, Germany) for the presence of intermediate products, after the reaction.
Next, for investigation of the optical properties of the materials, UV photoelectron spectroscopy (UPS; Axis Ultra DLD, Japan) with He I line (21.2 eV) UV source was applied. The work functions of the materials were calculated by the equation where hv is the incident energy (21.2 eV), E cutoff is the secondary electron cutoff energy, and E f is the Fermi energy 55 .

Data Availability
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.