Novel magnetic core–shell Ce–Ti@Fe3O4 nanoparticles as an adsorbent for water contaminants removal

Magnetic core–shell Ce–Ti@Fe3O4 nanoparticles were synthesized by coating cerium titanate on magnetite under mild experimental conditions. Combining magnetism, crystallinity, stability and adsorption capacity, it can be a promising nanomaterial as an adsorbent for anionic water contaminants, exhibiting high removal capacity, from 85% to 100%, for nitrates, phosphates and fluoride.

on nanoparticles (NPs), nanomembranes, carbon nanotubes (CNTs) and nanofibers, among others 4 . Thus, the use of adsorbents nanomaterials has become an interesting way for the removal of various contaminants from drinking water 5 such as of heavy metals 6 and, in a minor extent, nutrients 7 . For instance, the use of CeO 2 , Fe 3 O 4 and TiO 2 NPs and magnetic nanocomposite 7 for the adsorption of cadmium 8 and phosphate 7,9 has been reported by our group. Finally, other metal oxides and metal hydroxides had also been reported 10 for fluoride removal from water as well as bimetallic or mixed oxides 11 . Further, it is worthy to consider the reusability and the regeneration of the adsorbents as well as the trapping of the NPs to prevent its environmental and health safety risks 12 . Thus, the use of magnetic NPs for pollutants removal provides efficient, easy separation, and reusability. The magnetic NPs can be either used directly or as the core material in a coreshell NPs structure 13 . For instance, cerium titanates nanomaterials (Ce 2/3 TiO 3 ) have many applications as photocatalytic and ferroelectric materials 14 . However, its properties exhibits canted-antiferromagnetic order 15 and few studies for improving its properties have been reported 16 . In this work, a novel nanomaterial was synthesized by coating cerium titanate on magnetite NPs using a simple and easy method as one step synthesis under room temperature to obtain magnetic core-shell Ce-Ti@Fe 3 O 4 NPs (shell@core).
Moreover, its efficiency was demonstrated for water remediation as a versatile nanoadsorbent for typical inorganic contaminants.
Briefly, the synthesis of the nanoadsorbent here reported was based in the following procedure: previously of the synthesis of the core-shell Ce-Ti@Fe 3 O 4 NPs, magnetite nanoparticles (Fe 3 O 4 -NPs) were prepared by the coprecipitation method as reported elsewhere 7,17 by using cetyl trimethyl ammonium bromide, CTAB, as dispersant. During the synthesis, the mixture's colour turned from light yellow to red brown and then eventually to black which confirmed the formation of Fe 3 O 4 -NPs. Once the Fe 3 O 4 -NPs were washed and dried, Ce-Ti@Fe 3 O 4 NPs were synthetized. TiCl 4 and Ce(NO 3 ) 3 ·6H 2 O were mixed with 100 mL of Milli-Q water containing the previous formed Fe 3 O 4 -NPs, with a Ti 4+ :Ce 3+ molar ratio of 1:1 and to reach a total molar concentration of 50 mM. Mixing was under agitation at room temperature for 30 min. Next, a slowly dropwise titration with 12.5 %v/v NH 3 solution until pH 7.0 was reached. Then, the Ce-Ti@Fe 3 O 4 NPs produced were washed with ultrapure water and magnetic decantation and finally dried at 80 °C for 24h. The synthetic procedure was adapted from similar works about core-shell magnetic Ti-NPs synthesis 13a . Further details about the materials used and the synthetic protocols are found in Electronic Supplementary Fig.1a, Ce-Ti@Fe 3 O 4 NPs present a particle size within the range of 10 -15 nm. Electron diffraction pattern allows studying the crystal structure of the NPs. Thus, the ED pattern for Ce-Ti@Fe 3 O 4 NPs (Fig. 1b), taken from randomly selected area of the nanomaterial (SAED), exhibits multiple rings consisting of discrete spots, which suggested that the core-shell NP is based on nanocrystals. In addition, EDS provided the metal composition of the samples and Fig. 1c proved the presence of the three components Ce, Ti, and Fe in the material. Moreover, magnetic properties of the Ce-Ti@Fe 3 O 4 nanomaterial were qualitatively tested using a square magnet, showing the strong magnetism of the material (Fig. 1d). This result demonstrates an enhancement on its properties comparing to the literature where Ce 2/3 TiO 3 exhibits antiferromagnetic properties 15 . This also demonstrates that the NPs could be easily recovered from the reaction mixture for further reuse in its application. Further, as shown in Table S1 ( †ESI.2.2), the calculated interplanar distance (d) from the SAED pattern of the Ce-Ti@Fe 3 O 4 NPs and the corresponding Miller indices (h k l diffraction plan) were compared to the standard values 18 . The values concluded that the crystal pattern presented similarity to both cerium titanate and Fe 3 O 4 patterns, which was in agreement with the XRD data further reported.
XRD technique was used to obtain the crystalline structure of the Ce-Ti@Fe 3 O 4 NPs. In a diffraction pattern, the location of the peaks on the bragg angles (2θ scale) can be compared to the reference peaks ( †ESI.2.3). Fig. 2 shows the XRD pattern of the original Ce-Ti@Fe 3 O 4 nanomaterial, which consists of two phases: magnetite and cerium titanate reference patterns that were proved by matching from database (  Fig. 2, one can notice the characteristic diffraction peaks belonging to cubic Fe 3 O 4 . They correspond to (220), (311), (400), (511) and (440) family planes (PDF 89-4319). After cerium titanium oxide coating, the characteristic peaks of cerium titanate were appeared and its bragg angles were found to be close to that of Ce 2/3 TiO 2.98 . Two peaks were observed corresponding to rutile/anatase in the magnetic cerium titanate NPs: (110) peak of anatase and (101) peak of rutile closely positioned at each other, namely at around 2θ = 25.3 and 2θ = 27.4, respectively.  The morphology of the core-shell structure of the Ce-Ti@Fe 3 O 4 nanomaterial was demonstrated by Scanning Transmission Electron Microscopy (STEM) coupled with Electron Energy Loss Spectra (EELS), ( †ESI.2.4). Thus, the images obtained from a HAADF detector provide densitybased contrast, the cores appearing bright due to their higher scattering probability. 19 As illustrated in Fig. 3, the HAADF image of Ce-Ti@Fe 3 O 4 shows that the Fe 3 O 4 NPs was coated with Ce-Ti oxide layer (Fig. 3a), moreover STEM based EDS was used to confirm the elemental distribution of the Ti, Fe and Ce, respectively, as shown in Figure 3b. In addition, the scanning line profile appears as a strong peak corresponding to the Please do not adjust margins Please do not adjust margins position of the bright Fe particle core, whilst the spectra on each side of the core are dominated by Ce and Ti edges as illustrated in Fig.3c.
UV/Vis Absorption and luminescence spectra of Ce-Ti@Fe 3 O 4 NPs were performed to estimate the valence of the cerium in the NP and confirm its speciation ( †ESI.2.5). As reported, the cerium ion in the cerium titanate, Ce 2/3 TiO 3 , is mainly Ce(III). 16 Fig. S.1 ( †ESI.2.5) shows the absorption UV-Vis spectra of the Ce-Ti@Fe 3 O 4 nanomaterial. As shown, two peaks were observed at 250 and 310 nm respectively, which could be attributed to the presence of either Ce(III) or both Ce(III) and Ce(IV) in the nanomaterial 20 . Therefore, because of the overlapping of both bands, it is difficult to determine the species responsible of colour with the colorimetric technique. Thus, luminescence spectroscopy is necessary to obtain information about the valence of the cerium. Ce(III) ions presents a characteristic intense blue emission upon UV excitation. 21  nanomaterial. The excitation spectrum for λ em =363 nm shows a band at 258 nm and the emission spectrum recorded upon λ exc =266 nm shows the characteristic emission band at 325 nm, which corresponds to the transition to ground state to excited state as compared to literature. 20,22,23 The slightly shift of the wavelength comparing to the literature could be attributed to the lower temperature of the synthesis of the materials andto the presence of magnetite NPs in the core.
In this work, we also investigated the adsorption capacity and removal efficiency of the nanomaterials for different water contaminants: fluoride, nitrate, phosphates and cadmium. The adsorption experiments procedures are based on batch adsorption tests to determine the adsorption efficiency by the synthesized NPs ( †ESI.4). Residual contaminant concentration in the solution after 24h of adsorption (Ce) was determined by the corresponding analytical method detailed in †ESI.3 and the equilibrium adsorption capacity (Qe) of the adsorbent was calculated as Equation S.1 ( †ESI.4). Adsorption experiments were performed using different initial concentrations for each contaminant that are based either on the reported typical concentration in water or on the maximum contaminated level (MCL). For instance, phosphate initial concentration tested was 10 mg/L due to municipal wastewater may contain 4-15 mg/L, and domestic wastewater may contains 10-30 mg/L 7 . Furthermore, 10 mg/L was selected as initial fluoride concentration because the maximum contaminated level in water is 1.5 mg/L 24 . In the case of fluoride,it has been reported that its concentration in groundwater ranges from well under 1.0 mg/L to more than 35.0 mg/L in several regions. 25 In addition, the initial nitrate concentration tested is 50 mg/L according to the WHO guidelines, where the MCL is 50 mg/L 24b . Also, 10 mg/L of initial cadmium concentration was selected as wastewater contains 10 -100 mg/L of cadmium contaminant 26 . All the experiments were performed at pH 7 as a typical value in real media. Table 1 shows the equilibrium adsorption capacities after 24h, the percentage of removal of each contaminant using a concentration of 1g/L of the nanoadsorbent for all the cases. It is shown that the Ce-Ti@Fe 3 O 4 NPs have a potential effect for removal anionic contaminants (i.e. fluoride, nitrate and phosphates) from 85% removal for nitrate to 100% for phosphates and fluoride. However, it presents low removal for cationic contaminants such as cadmium (45% removal). The differences obtained on the adsorption process for the different contaminants tested may be discussed in terms of the physicochemical properties of the material and thus, the adsorption mechanism of the Ce-Ti@Fe 3 O 4 material could be hypothesized. On the one hand, the metal oxides NPs present a relatively negative charge (hydroxyl groups, OH-) on the oxide surface due to its hydrolysis in aqueous media. On the other hand, as discussed in this study, Ce-Ti@Fe 3 O 4 NPs consist of Ce(III) ions and also it has a potential of +165 mV at pH 7. Thus, different sorption processes could take place for these contaminants tested. The adsorption mechanism for the anions could be attributed to two explanations: i) electrostatic attraction (i.e. chemisorption) 27 and ii) surface ion-exchange process (i.e. physisorption) 28 . In the first case, phosphate removal could be explained in terms of the formation of cerium phosphate, as reported. 29 In the second case, the OHgroups on the adsorbent surface played a dominant role. Therefore, fluoride or nitrate removal may suffer a surface ionexchange process based on the exchange of the OH-group with the contaminant anion 28 . Also, electrostatic force could take place between the anions and the positive charge from the quaternary ammonium group from the cross-linker CTAB, used in the synthetic protocol as stabilizer. It has been reported that heavy metal adsorption onto NPs is an emerging technique for the removal of these pollutants due to its suitable electric charge given by an adequate Z-potential 30 . However, Ce-Ti@Fe 3 O 4 material shows a low removal for cadmium due to the positive charge of the NPs surface (positive potential) that contributes to a weak electrostatic interaction between cadmium and the OH-from the oxide surface of the NPs. Further, the efficiency of this material was demonstrated by comparing the equilibrium adsorption capacities with those reported in literature (Table SI.

Conclusions
Magnetic core-shell Ce-Ti@Fe 3 O 4 nanoparticles were designed and synthesized by incorporating magnetite into Ce-Ti oxide nanoparticles by mild experimental conditions. The resulting magnetic and core-shell nanomaterial exhibited a suitable composition, crystallinity and magnetic properties to be functional as nanoadsorbent for the removal of inorganic pollutants from aqueous media. Remarkably, the adsorption capacity and removal efficiency at pH 7 for anionic contaminants such as nitrates, phosphates and fluoride was from 85 to 100% under the experimental conditions. In comparison with other materials reported in literature, this nanoadsorbent is highly competitive, as it has high adsorption capacity and it is easy to recover from the reaction mixture for further reuse due to its magnetic properties.
The author, Ahmad Abo Markeb, appreciated and would like to thank the Ministry of Higher Education of Egypt for the Ph.D external mission grant. Special thanks are given to Servei de Microscopia from Universitat Autònoma de Barcelona and from the Institut Català de Nanociencia i Nanotecnologia.

ESI.2.4. Scanning Transmission Electron Microscopy (STEM) coupled with Electron Energy Loss Spectra (EELS)
The morphology of the core-shell nanocomposite was estimated by STEM coupled with HAADF detector and EELS. Images were acquired using an FEI Tecnai G2 F20 microscope operated at 200 kV and equipped with a GIF Quantum energy filter. All spectra were recorded using a convergence semiangle of about 12 mrad and a collection semiangle of about 40 mrad. EDX spectra were obtained using an EDAX super ultrathin window (SUTW) X-ray detector. The sample was first dispersed in ethanol and sonicated, then deposited onto the copper microscopy grid coated with an amorphous carbon film. By imaging with the electrons that have an energy loss corresponding to core losses of particular elements using STEM, one can obtain elemental information with high spatial resolution. A full energy loss spectrum from a series of points across the particle in a STEM configuration, which allows the extraction of linear compositional variation was used to obtain chemical information about the nanostructure. Analyses were performed at Institut Català de Nanociència i Nanotecnologia (ICN2), Spain.

ESI.2.5. UV/VIS and luminescence spectra analysis
Absorption and luminescence spectra of Ce- Chromeleon® software was used to acquire data and control the instrumentation.
Standard error in the measurements is < 0.1%.
A stock solution of each contaminant was prepared by dissolving the appropriate amount of its precursor in ultrapure water. All working contaminants solutions for calibration curve and adsorption studies were prepared by diluting the stock solution.
Calibration standards and samples were filtered using 0.45 µm Nylon membrane filter before injection.

ESI.3.2. UV-Vis for Cadmium analysis
Calibration curves for cadmium were constructed using 99.995% cadmium(II) chloride by using a colorimetric method, based on the reaction of cadmium with dithizone to form a complex that is extracted with chloroform. Then the absorbance is measured at 518nm 7 . Cadmium stock solution was prepared by dissolving the appropriate amount in 5 % nitric acid. The cadmium solutions for calibration curve and adsorption studies were prepared by diluting the stock solution. wastewater may contains 10-30 mg/L 1a . Furthermore, 10 mg/L initial fluoride concentration was selected due the maximum contaminated level in water is 1.5 mg/L 8 but, it has been reported that the fluoride concentrations in groundwater range from well under 1.0 mg/L to more than 35.0 mg/L in several regions of India. 9 In addition, the initial nitrate concentration tested is 50 mg/L due to according WHO guidline the MCL is 50 mg/L 8b . Also, 10 mg/L intial cadmium coencentration was selected due to wastewater contains 10 -100 mg/L of cadmium contaminant 10 . All the experiments were performed at pH 7 as a typical value in real media.