Preparation of multifunctional nanocomposites Fe₃O₄@SiO₂–EDTA and its adsorption of heavy metal ions in water solution

With the rapid development of electroplating equipment, mining, tanneries, batteries, paper, fertilizers, and pesticides, more and more heavy metal wastewater is directly or indirectly discharged into the environment, which has a serious impact on the ecosystem. Unlike organic pollutants, heavy metals are not biodegradable and accumulate in living organisms (Awual et al. 2014; Dong et al. 2015); therefore, heavy metal pollution is a major problem in today's society. There are many methods for removing heavy metal ions, such as ion exchange method (Gode & Pehlivan 2006), chemical precipitation method (Fu & Wang 2011), electrochemical treatment (Nanseu-Njiki et al. 2009), membrane separation method (Landaburu-Aguirre et al. 2010), biochemical method (Wu et al. 2010), adsorption method (Baccar et al. 2009; Ali 2012; Sharififard et al. 2012; Ren et al. 2013; Sun et al. 2013; Kong et al. 2014; Cegłowski & Schroeder 2015) and so on. In these methods, the adsorption method is suitable for low ion concentration wastewater and large adsorption amount, but the selection of adsorbents is particularly important. The adsorbents currently used include carbon-based materials, mineral materials, metal oxide materials, ion-selective materials, and biological materials. Magnetic nanoparticles (MNPs, such as ferroferic oxide, Fe₃O₄) have also been extensively studied due to their superparamagnetism, which make them easy to separate from water (Zargoosh et al. 2014; Zhang et al. 2014). In addition, magnetic materials possess a larger surface area and can be synthesized by simpler methods (Cheng et al. 2012). However, the agglomeration of Fe₃O₄ in water reduces the active sorption sites and weakens its adsorption capacity (Zhang et al. 2014). Many experiments have shown that proper surface functionalization can solve this problem and enhance the stability and performance of Fe₃O₄ MNPs (Kim et al. 2013).

MNPs have undergone many modifications, including thiol functionalization and amino functionalization, and MNPs modified with poly(2-aminoethyl methacrylate hydrochloride), dithiocarbamate, (3-aminopropyl) triethoxysilane, and polydopamine have a distinct affinity to combine with Cu(II), Hg(II), Ag(II), Cd(II), Pb(II), Co(II), Ni(II), and Cr(III) (Zhang et al. 2014; Anbia et al. 2015; Chensi et al. 2020; Mahtab et al. 2020). Thiol-functionalized magnetic nanoparticles (MNPs) also have higher adsorption properties for Hg(II) because Hg(II) can bind strongly with sulfur-containing groups (Saman et al. 2014).

Ethylenediamine tetraacetic acid (EDTA) as a modification material for adsorbents has recently attracted more attention (Ren et al. 2013). EDTA functionalization adsorbent not only possesses good adsorption behaviors, but it also does not cause any secondary pollution. In previous studies, materials assembled with EDTA using Na₂EDTA (EDTA disodium salt) (Ren et al. 2013), EDTA (Liu et al. 2013), EDTA anhydride (Repo et al. 2010) and N-(trimethoxy-silylpropyl)ethylenediamine triacetic acid (EDTA-silane) (Liu et al. 2016) have been widely investigated.

In this study, the EDTA-functionalized Fe₃O₄ (–COOH-modified magnetic materials,) was prepared using a novel method that bonded EDTA to the particle surface by polycondensation of amino and carboxyl groups, and removed Pb(II) and Cu(II) from aqueous solutions under various experimental conditions.

Sodium hydroxide, ferric chloride hexahydrate (FeCl₃·6H₂O), ferrous chloride tetrahydrate (FeCl₂·4H₂O), hydrochloric acid, tetraethyl orthosilicate (TEOS), ammonia solution, Na₂EDTA dihydrate and metal nitrates (Cu(NO₃)₂·5H₂O, Pb(NO₃)₂, Cd(NO₃)₂·4H₂O,) were purchased from Chengdu Chron Chemicals China. 3-Aminopropyltrimethoxysilane (APTMS) was obtained from Tokyo Chemical Industry Japan. N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDCI) and 1-hydroxybenzotriazole monohydrate (HoBt) were supplied by Aladdin Reagent, China. All the reagents were of analytical grades and used as received without further purification. All solutions were prepared using de-ionized (DI) water.

The phase structures were analyzed using XRD (Dang Dong Fang Yuan, DX-2600, China) with Cu Kα radiation. The functional groups' infrared spectra were obtained with a TENSOR 27 Fourier transform infrared spectrometer (FTIR, Bruker Optics, TENSOR 27, Germany). The morphology and the sizes of the reluctant products were characterized using a field emission-scanning electron microscopy (FE-SEM, ZEISS, Germany). Thermogravimetric analysis (TGA) was carried out using a thermal analyzer (Mettler Toledo, TGA, Switzerland) at the heating rate of 15 °C min⁻¹ under a dynamic nitrogen atmosphere flow. Hysteresis loops were recorded at room temperature with a vibrating sample magnetometer (VSM, Quantum Design, USA). An 5100 Synchronous Vertical Dual View inductively coupled plasma–optical emission spectrometer (ICP–OES, Agilent, USA) was used to test the concentrations of metal ions.

CMS was prepared using the raw materials mentioned earlier and the preparation process included three stages (shown in Figure 1). First, 60 mL iron salt solution (containing 5.44 g FeCl₃·6H₂O and 2.00 g FeCl₂·4H₂O) was poured into a 250 mL three-neck round flask under nitrogen flow in ultrasonic bath. Then 100 mL sodium hydroxide solution (1 mol/L) was added drop by drop into the flask with slow mechanical agitation. The mixture was stirred for 0.5 h at 60 °C and left another 2 h under continues nitrogen flow. The black Fe₃O₄ nanoparticles were separated using external magnets and washed with dissolved oxygen-free water to a neutral pH and dried at 60 °C in a vacuum oven.

Second, 0.5 g Fe₃O₄ was suspended in a mixture (150 mL ethyl alcohol and 25 mL DI water) using ultra-sonication for 15 min to obtain a homogeneous Fe₃O₄ dispersion. Then 3 mL of ammonia was injected whilst stirring at 30 °C. After 30 min, 2.5 mL TEOS was added and the reaction was further held under mechanical agitation for 45 min at 30 °C. After that, 0.5 ml APTMS was added and the reaction was held under mechanical agitation for 5 min at 30 °C. The black–brown nanoparticles were then rinsed with ethanol and DI water several times and –NH₂-modified magnetic materials (NMS) was prepared. The obtained product was separated using an external magnet, dried at 60 °C for 24 h, ground into fine powder and stored in a desiccator for further experiments.

Finally, the surface modification of NMS with EDTA was performed as follows: the as-prepared NMS (0.4 g) was dispersed in 160.0 mL DI water solution (containing 5.96 g EDTA 2Na, 1.60 g EDCI, and 0.86 g HoBt) overnight and mechanically stirred. Thus the EDTA²⁻ ions were adsorbed onto the NMS via the electrostatic attraction between the protonation of amino groups of NMS and anionic moiety of carboxyl groups of EDTA; the reaction of the carboxyl groups of EDTA with the amino groups on NMS was completed using a small amount of EDCI as a cross-linker; and the addition of HoBt increased the condensation yield of the reaction. The products were rinsed with DI water several times, dried in the vacuum oven at 50 °C for 12 h, and then stored in the desiccator.

Figure 2 shows the X-ray diffraction (XRD) patterns of the samples, including Fe₃O₄, Fe₃O₄@SiO₂, NMS, and CMS nanoparticles. The XRD results showed that the samples had sharp peaks, which demonstrated that the prepared adsorbents have a higher crystalline nature. The characteristic diffraction peaks that appeared in Fe₃O₄ at 18.3°, 30.1°, 35.4°, 43.1°, 53.4°, 56.9, 62.5°, and 75° could be assigned to [111], [220], [311], [400], [422], [511], and [622] planes of magnetite (JCPDS card no. 19-0629), respectively, confirming that pure Fe₃O₄ has a cubic spinel structure (Larumbe et al. 2012; Tan et al. 2014). No more peaks were assigned to the crystalline form in Fe₃O₄@SiO₂ NMS and CMS These characteristic diffraction peaks observed for the samples of Fe₃O₄@ SiO₂, NMS, and CMS indicated that Fe₃O₄ was present in Fe₃O₄@SiO₂, NMS, and CMS, and the whole preparation processes did not result in the phase change of Fe₃O₄, implying that Fe₃O₄@SiO₂, NMS, and CMS nanoparticles retained the crystalline cubic spinel structure (Kassaee et al. 2011).

Figure 3 showed the FTIR spectra of Fe₃O₄, Fe₃O₄@SiO₂, NMS, and CMS nanoparticles. The typical absorption peak at 577 cm⁻¹ and 1,634 cm⁻¹ is attributed to the stretching vibration of Fe–O bond (Tan et al. 2014). For Fe₃O₄@SiO_2, the Fe–O peak decreased in Fe₃O₄@SiO₂, confirming that the functional groups successfully coated the Fe₃O₄ surface. The presence of the peak at 3,405 cm⁻¹ is due to the stretching vibrations of O–H and N–H bonds (Repo et al. 2011). Si–O–Si bands, which were generated from oligomerization of the silanes, as well as Si–OH were also found in Fe₃O₄@SiO₂. As shown in Figure 3, the peak at 1,085 cm⁻¹ was related to Si–O–C and symmetric Si–O–Si stretching vibrations, and the peaks at 959 cm⁻¹ and 804 cm⁻¹ were related to Si–OH and asymmetric Si–O–Si stretching, respectively (Repo et al. 2011; Dupont et al. 2014; Tan et al. 2014), implying the successful formation of silica layers on Fe₃O₄. For NMS, the presence of two new bands at 2,942 cm⁻¹ and 2,885 cm⁻¹ was associated with stretching of methylene groups, indicating the successful grafting of APTMS (Hou et al. 2010). The spectra of CMS shows a strong absorption peak at 1,405 cm⁻¹ due to the stretching vibration of C=O in symmetrical –COOH groups, and this peak could be also assigned to the C–N bond and γ_CH2 group (Hou et al. 2010; Huang & Keller 2015). Moreover, compared with the two spectra lines of Fe₃O₄@SiO₂ and NMS, CMS has a distinct enhanced absorption peak at 1,634 cm⁻¹, which is caused by the C=O stretching vibrations of the asymmetric COO− (Repo et al. 2011). All these phenomena indicate that EDTA successfully coated the surface of magnetite nanoparticles, and the carboxyl modification was completed. 2,349 cm⁻¹ is an asymmetric stretching peak of CO₂ in the air.

Figure 4 presents the SEM micrographs of Fe₃O₄, Fe₃O₄@SiO₂, NMS, and CMS. All synthesized adsorbents have spherical shapes with individual particles, and there are also distinguished morphological differences. The particle size was in the following order CMS > NMS > Fe₃O₄@SiO₂ > Fe₃O₄, which was ascribed to the groups loaded onto their surfaces.

For further investigation of functionalized Fe₃O₄ properties, TGA analysis of the nanoparticles was performed to determine the quantity of the grafted EDTA materials (shown in Figure 5) (Fan et al. 2011). It was found that the TGA curves of Fe₃O₄ and Fe₃O₄@SiO₂ nanoparticles changed slightly and the TGA curves of NMS and CMS nanoparticles changed rapidly with the temperature increase. For Fe₃O₄ nanoparticles, the weight loss was about 5.06% at temperature ranging from 30 °C to 1,000 °C, which was caused by the escape of physically adsorbed water and/or structure water on the surface (Wang et al. 2010). The TGA curve of Fe₃O₄@SiO₂ showed that the weight loss was about 4.90% below 200 °C due to the release of moisture on the surface of the silica layer, while the weight loss was about 2.98% at temperature ranging from 200 °C to 1,000 °C, which was related to the structure water (Wang et al. 2010). For the TGA curves of NMS and CMS, organic polymers on NMS and CMS began to degrade rapidly above 250 °C in addition to water loss, which is similar to the temperature at which degrading begins for cross-linked magnetic cross-linked magnetic chitosan–diacetylmonoxime Schiff's base resin (CMSO) as reported by Monier et al. (Fan et al. 2011). The organic polymers were almost completely decomposed at 1,000 °C for NMS and 950 °C for CMS, respectively. A weight loss of about 12.80 wt.% (taken off the structure water) from 200 °C to 1,000 °C in NMS, and a weight loss of about 22.52 wt.% (taken off the structure water) from 200 °C to 950 °C in CMS was used to estimate the weight proportions of cross-linked amino-groups on NMS and EDTA-modified carboxyl-group on CMS, respectively. The weight proportion of polymers on CMS (22.52 wt.%) was higher than that on NMS (12.80 wt.%). This result further verified that EDTA was modified successfully on the surface of NMS.

Figure 6 shows the magnetic hysteresis loop of Fe₃O₄, Fe₃O₄@SiO₂, NMS, and CMS nanoparticles. The saturation magnetization was 60.82, 45.52, 22.40, and 20.71 emu g⁻¹, respectively. Although the saturation magnetization values of CMS is lower than that of Fe₃O₄ and Fe₃O₄@SiO₂, CMS could be rapidly separated from their water dispersions within several seconds using the Nd–Fe–B permanent magnet (N35 grade) (seen in inset of Figure 6). These results demonstrated that CMS exhibited enough magnetic response (Fan et al. 2011). Moreover, the saturation magnetization decreased after functionalization, demonstrating a successful grafting of EDTA molecules onto the Fe₃O₄ surface. The weighted percentages of silica coating on Fe₃O₄, cross-linked amino-group on NMS, and EDTA-modified carboxyl-group on CMS calculated from saturation magnetization data were about 25.16%, 38.01%, and 40.79%, respectively. However, the content of EDTA on Fe₃O₄ from TGA data was not consistent with that from magnetic hysteresis loop. This could be attributed to the strong thermal stability of silicon (Liu et al. 2016).

The adsorption properties of CMS for heavy metal are determined by many factors. The effect of adsorbent dosage and pH on removal of heavy metal was investigated in this work.

The effect of dosage of CMS on the removal efficiency and adsorption capacity of Cu(II) and Pb(II) is shown in Figure 7. It was necessary to obtain the optimum adsorbent dose to maximize the interactions between metal ions and the adsorption sites of the adsorbent in the solution Hao et al. (2010). It can be seen that the adsorption capacities of metal ions increase with an increase of dosage of CMS. The removal efficiencies also increase greatly as the dosages of CMS increase from 0.5 g L⁻¹ to 1.0 g L⁻¹, and then reach a plateau when the dosages increase from 1.0 g L⁻¹ to 1.5 g⁻¹. These observations illustrate that the increase of the adsorbent dose may multiply the number of available adsorption sites, leading to the raise of the removal efficiency. However, when nearly all the heavy metals in the aqueous solutions were adsorbed by the adsorbents, the number of unoccupied active adsorption sites increased and could no longer contribute to the removal percentage, resulting in the decrease of adsorption capacities of the adsorbents (Ren et al. 2013). Moreover, high dosage may result in aggregation of the adsorbents (Zhou et al. 2009) and could further reduce the adsorption capacities of adsorbents. Therefore, as presented Figure 7, CMS exhibits better removal percentage and better adsorption capacity at the dose of 1.0 g L⁻¹, which was selected as the optimum adsorbent dose for the following experiments.

Solution pH is an important factor affecting the interaction between sorbents and metal ions. In the water environment, pH variations can significantly affect the protonation of surface groups and speciation of metal ions. In aqueous solutions with a pH above 7, the lead (Pb) hydroxyl components are the dominant species, which would hinder the binding between metal ions and sorbents. At a pH below 7, Pb and copper (Cu) exist as Pb(II) and Cu(II), and can favorably achieve sorption onto an adsorbents' surface. In this study, experiments were performed to investigate the effect of solution pH on adsorption at a wider pH range (1–5). Figure 8 presents the effect of solution pH on Cu(II) and Pb(II) adsorption onto CMS. The adsorbent (CMS) exhibited better adsorption efficiency for Pb(II) at pH 2, but the adsorption capacities decreased with pH from 3 to pH 5. The adsorption capacity of Cu(II) followed an increasing trend from pH 2 to pH 5, as shown in Figure 8.

A novel magnetic adsorbent – EDTA-functionalized Fe₃O₄ (CMS) – was synthesized by polycondensation of amino and carboxyl groups and the effect on removing Pb(II) and Cu(II) from aqueous solutions was investigated. Characterization analyses of CMS proved EDTA successfully loaded onto Fe₃O₄. CMS showed high stability in low pH solution and good removal efficiency for heavy metal ions. The maximum adsorption capacities of CMS were 0.11 mmol g⁻¹ for Cu(II) at pH5 (30 °C) and 0.14 mmol g⁻¹ for Pb(II) ions at pH2 (30 °C). This study indicated that the CMS could be considered as a potential adsorbent to remove Pb(II) and Cu(II) from wastewaters at low pH.