Facile Synthesis of Core-Shell Magnetic Iron Oxide@SiO₂-NH₂ Nanoparticles and Their Application in Rapid Boron Removal from Aqueous Solutions

Abstract

In this study, amino-functionalized magnetic particles (iron oxide@SiO₂-NH₂) with core-shell structures were synthesized and evaluated for rapid boron removal from aqueous solutions. The results showed that the specific surface area of the iron oxide@SiO₂-NH₂ (131.24 m²⋅g⁻¹) increased greatly compared to pure iron oxide (30.98 m²⋅g⁻¹). The adsorption equilibrium was less than 2 h, with an adsorption capacity of 29.76 mg⋅g⁻¹ at pH = 6 at 15 °C. The quasi-second-order kinetic model described the boron adsorption process well, and both the Langmuir and Freundlich models were suitable for characterizing the adsorption isotherms. The zeta potential and XPS analysis before and after adsorption revealed that the main adsorption mechanism was the hydrogen bonding formation between the terminal -NH₂ groups of the adsorbent and the boric acid. In addition, the adsorbent still maintained a high adsorption performance after five adsorption–desorption cycles, which illustrated that the iron oxide@SiO₂-NH₂ may be a potential adsorbent for environmental boron removal treatment.

1. Introduction

Boron and boron compounds have been widely used in glass, ceramics, nuclear power and chemical industries due to their lightweight, heat-resistant, wear-resistant, high-strength, and high thermal neutron absorption characteristics [1]. Boron is also an indispensable trace element for animals, plants and humans [2], but excessive intake of boron can result in serious harm to organisms, such as causing declines in crop production and human nervous system diseases [3,4]. So, the amount of boron allowed in drinking water is no more than 2.4 mg·L⁻¹, limited by the World Health Organization (WHO) [5], and this value is even less than the 1.0 mg·L⁻¹ permitted in both drinking and irrigation water in China [6,7]. Therefore, boron recovery/removal from aqueous solutions (salt lake brine, seawater, drinking water, wastewater, etc.) is crucial for boron resource utilization and environmental treatment.

The main technologies for boron extraction include electro-flocculation [8], membrane filtration [9], chemical precipitation [10], solvent extraction [11], and adsorption (including ion exchange) [12,13]. Of these, adsorption has become one of the most effective methods due to its ease of operation, high selectivity, and low energy consumption. Currently, research on boron adsorption has been focused on various absorbents, such as ion-exchange resins, activated carbon, metal oxides and hydroxides, polymers, metal-organic frameworks (MOFs), and magnetic adsorbents [14,15,16,17,18]; each of these has its own advantages and disadvantages during the boron removal process. Magnetic nanoparticles (MNPs) and their functional composites embody high boron adsorption efficiency, simple magnetic separation, and easy regeneration because of their large surface area, strong magnetic property, low cost, and environmentally friendly nature. Iron oxide nanoparticles are the most frequently used magnetic absorbents for boron absorption [19]. However, these materials are susceptible to autoxidation and aggregation and have fewer surface functional groups that interact with the boron. Thus, functionalized magnetic nanoparticles with chelating species are often necessary. Based on the amphoteric properties of amino groups, amino-based MNPs, performing surface modification by covalently linking organosilane species with surface silanol groups, have been extensively studied and used as absorbents in water treatment [20]. Liu, Z. et al. [21] synthesized Fe₃O₄@SiO₂-NH₂ and modified its surface with polyacrylamide for the removal of heavy meatal ions of Cd²⁺, Pb²⁺_, and Zn²⁺ from wastewater, and the adsorption capacity of this material was in the following order: Pb²⁺ > Cd²⁺ > Zn²⁺. Ghorbani, F. et al. [22] synthesized the core-shell magnetic nanocomposite Fe₃O₄@SiO₂-NH₂ with the silica source extracted from rice husks and a 3-aminopropyltrimethoxysilane functional group, and then it was used it for the adsorption of Methyl Red dye. The maximum adsorption was 125 mg·g⁻¹, which was much higher than that of most other absorbents. Furthermore, the removal efficiency was decreased by only 10% after five regeneration cycles. Donga, C. et al. [23] coupled graphene oxide to Fe₃O₄@SiO₂-NH₂ in order to obtain composites for the adsorption of three heavy metal ions, Pb(II), Cd(II), and Ni(II), and showed that the absorption capacity followed the order Pb(II) > Cd(II) > Ni(II). From the above studies, it is known that amino-based magnetic absorbents are chemically stable, reusable, and show a great application potential for heavy metal removal during water/wastewater treatment processes. However, to the best of our knowledge, boron adsorption from aqueous solutions and its interaction mechanisms with amino-based magnetic absorbents still remains unknown. Therefore, in this work, iron oxide@SiO₂-NH₂ was prepared by a facile sol-gel method and was then applied to remove boron from a solution. The pH effect, adsorption isotherm, kinetics, and regeneration were studied.

2. Materials and Methods

2.1. Materials

Ferric chloride hexahydrate (FeCl₃·6H₂O, Aladdin Biochemical Technology Co., Ltd., Shanghai, China), sodium acetate anhydrous (NaOAc, Shanghai Yuanye Biotechnology Co., Ltd., Shanghai, China), ethylene glycol (EG, J&K Scientific Co., Ltd. Beijing, China), tetraethyl orthosilicate (TEOS, Shanghai McLean Biochemical Technology Co., Ltd., Shanghai, China), 3-Aminopropyltriethoxysilane (APTES, Shanghai McLean Biochemical Technology Co., Ltd., Shanghai, China), ethanol (EtOH, Shanghai Boer Chemical Reagent Co., Ltd., Shanghai, China), ammonia solution (NH₃·H₂O, Tianjin Quartz Zhongpi County Chemical Plant, Tianjin, China), and boric acid (H₃BO₃, Tianjin Fuchen Chemical Reagent factory, Tianjin, China) are analytically pure reagents, and can be used without any purification.

2.2. Synthesis of Iron Oxide MNPs

The iron oxide MNPs were synthesized by the solvothermal method [24] with mirror changes. In brief, 2.0 g of FeCl₃·6H₂O, 2.89 g of NaOAc, and 1.5 g of PEG-4000 were added in 70 mL of ethylene glycol. The mixture was stirred vigorously until it was completely dissolved, and then it was sealed in a Teflon-lined stainless-steel autoclave (100 mL) after 10 mL of H₂O was added and heated at 198 °C for 12 h. When the reactor was cooled down to room temperature, the product was collected with a magnet and washed with ultrapure water. After that, it was dispersed in a mixture of ethanol and water for further use.

2.3. Synthesis of Iron Oxide@SiO₂-NH₂ MNPs

The iron oxide@SiO₂-NH₂ was synthesized by the sol–gel method [25]. Firstly, 0.5 g of freshly prepared iron oxide MNPs was dispersed in a mixed solution of 200 mL of ethanol and 100 mL of water then ultra-sonicated for 3 min. Then, 6 mL of NH₃·H₂O (25–28 wt%) was added to adjust the solution pH to be alkaline, followed by the addition of 2–8 mL of TEOS, and the reaction was carried out under mechanical stirring for 40 min. Finally, the reaction was further performed for 4 h prior to the addition of 1–10 mL APTES. Afterward the precipitate was collected, washed with water, and dried in an oven at 60 ºC overnight. The overall synthesis steps are shown in Scheme 1. During the above process, the amount of TEOS and APTES was investigated to obtain the key preparation parameters for boron absorption properties.

2.4. Characterization of Iron Oxide and Iron Oxide@SiO₂-NH₂ MNPs

The XRD (X-ray diffraction) spectrum was recorded by an XRD diffractometer (D8 Discover, Karlsruhe City, Germany). The operating conditions were 60 kV and 60 mA using a Cu-Kα source (λ = 0.154 nm), and the samples were scanned in the 2θ range of 5–70°. The morphology and structural properties of the as-prepared functionalized MNPs were determined by scanning electron microscopy (SEM, SU8010, Tokyo, Japan), energy dispersive spectroscopy (EDS), and transmission electron microscopy (TEM, FEI Tecnai G2 F30, Hillsboro City, America). The SEM and TEM images were acquired at an accelerating voltage of 30 to 200 kV. The chemical structure was analyzed by Fourier transform infrared spectra (FTIR, iS50, Boston, MA, USA), which were recorded with a resolution of 4 cm⁻¹ in the region of 400–4000 cm⁻¹ by the KBr method. Specific surface area, pore size distribution, and pore volume of the samples were measured by a fully automated specific surface and pore size analyzer (Autosorb IQ2, Boynton Beach, FL, USA). The zeta potential was measured by a Zetasizer Nano ZS analyzer (ZS90, Malvern, UK) in the aqueous solution. And the solution pH was adjusted with hydrochloric acid and sodium hydroxide. In addition, the magnetic property was determined by a vibrating sample magnetometer (VSM, MPMS3, San Diego, CA, USA) at room temperature.

2.5. Adsorption Experiments

Stock solutions of boric acid were prepared with ultrapure water. A certain amount of the iron oxide@SiO₂-NH₂ was added into the boric acid solution and shaken for 12 h in a shaker for the boron adsorption. Then, the boron-loaded adsorbent was separated by magnetic separation. The boron concentration was determined by the mannitol titration method. During the absorption process, the effects of adsorbent dosing, initial solution pH, contact time, and reaction temperature on the adsorption were investigated. The adsorption capacity of boron was calculated by the following equation:

$$q={\displaystyle \frac{\left(c_0-c_{\mathrm{e}}\right)}{m}}V$$

(1)

where c₀ and c_e represent the boron concentration in solution before and after adsorption, respectively. V is the solution volume (L), m represents the adsorbent dosage (g), and q represents the adsorption capacity (mg·g⁻¹).

2.6. Regeneration and Reuse Study

Since the iron oxide@SiO₂-NH₂ MNPs were stable in alkaline conditions, while the optimum solution pH for boron adsorption was near neutral (pH = 6), desorption experiments were conducted by mixing 0.6 g of the boron-loaded iron oxide@SiO₂-NH₂ materials with 30 mL of a sodium hydroxide solution (pH = 10). The mixture was shaken at 250 rpm for 2 h at room temperature and then magnetically separated, and this process was repeated three times to make sure complete adsorbent–desorption was achieved. Then, the adsorbent was regenerated with a mixture of ethanol and ammonia (pH = 11). This material was dried overnight at 60 °C and was reused for further adsorption.

3. Results

3.1. Material Synthesis Process Optimization

The optimization synthesis process was carried out by varying the addition amounts of APTES and TEOS based on the boron adsorption capability of the materials. First, 4 mL of TEOS was used, and the APTES addition amounts varied from 1 to 10 mL. Then, the TEOS addition amounts were tested by fixing the APTES amount. Each experiment was repeated in triplicate. As shown in Figure 1, the boron adsorption capability first increased and then decreased with the increasing addition of both APTES and TEOS. The suitable dosages of APTES and TEOS were 7 mL and 6 mL, respectively.

3.2. Structural and Morphological Characterization

In order to obtain the structural information of the iron oxide and iron oxide@SiO₂-NH₂, XRD characterization was carried out, and the results are shown in Figure 2a. It can be seen that the diffraction peaks of the as-prepared iron oxide at 2θ = 18.32°, 30.16°, 35.46°, 43.11°, 53.46°, 57.00° and 62.54° are corresponding to the (111), (220), (311), (222), (400), (422), (511), and (440) crystal planes, respectively, which agreed well with the pure Fe₃O₄ standard PDF card (JCPDS: 89-0691). Since the γ-Fe₂O₃ and Fe₃O₄ have similar structures, it is not easy to distinguish them from the XRD pattern of the prepared sample. The diffraction peaks of the iron oxide@SiO2 and iron oxide @SiO2-NH2 are identical to the synthesized pure iron oxide except for the broad peak at 2θ = 20–30° of the iron oxide@SiO₂-NH₂, which can be attributed to the formation of the amorphous structure of the SiO₂ coating on the iron oxide surface.

According to the FTIR spectra (Figure 2b), the peak at 575 cm⁻¹ that appeared in the iron oxide, iron oxide@SiO₂, and iron oxide@SiO₂-NH₂ samples was the stretching vibration peak of Fe-O [26]. The bands at about 465 cm⁻¹, 797 cm⁻¹, and 1097 cm⁻¹ in the iron oxide@SiO₂ and iron oxide@SiO₂-NH₂ samples were related to the bending mode and symmetric and asymmetric stretching vibrations of the Si-O-Si band [27], respectively. These indicated that the SiO₂ layer has been successfully coated around the iron oxide MNP surface. The spectrum of the iron oxide@SiO₂-NH₂ showed that the characteristic peak at 1565 cm⁻¹ can be attributed to the functional group of N-H bending vibrations [28]. Moreover, the adsorption peaks at 2923 cm⁻¹ and 2856 cm⁻¹ can be generated by the symmetric or antisymmetric C-H stretching vibrations of methyl or methylene. In addition, the bending and stretching vibrations of -OH bands around 1632 cm⁻¹ and 3438 cm⁻¹ in all samples were associated with the water molecule adsorption. It was noted that the spectrum of the iron oxide@SiO₂-NH₂ at 3438 cm⁻¹ became sharper compared with the iron oxide sample, which may be attributed to the N-H symmetric stretching vibrations.

The SEM image in Figure 3a shows a quasi-spherical structure of the iron oxide MNPs with a diameter of 35–50 nm. After coating the SiO₂ on the iron oxide surface and grafting the amino group (Figure 3b), the morphology remains the same, but the diameter is larger than that of the iron oxide. As seen from the partially enlarged TEM image of the iron oxide@SiO₂-NH₂ (Figure 3b_i), there is a noticeable core-shell structure, with the darker part being the iron oxide core and the lighter-colored part the SiO₂ shell. Furthermore, Figure 4 represents the EDS spectra of the iron oxide@SiO₂-NH₂. As can be seen, it contains elements of C, N, O, Si, and Fe with the atom composition of 26.40% C, 8.63% N, 49.92% O, 12.16% Si, and 2.88% Fe, respectively, which confirms the successful amino functionalization for the nano-iron oxide with surface silanol groups. The lower concentration of Fe in the EDS can be attributed to the inner iron oxide core, which is not as easy to determine as the SiO₂ shell.

Figure 5a shows the N₂ adsorption–desorption isotherm curves of the iron oxide and iron oxide@SiO₂-NH₂ measured at 77 K. It can be seen that both of these samples exhibited typical IV isotherms, indicating that they were mesoporous materials. The Brunauer–Emmett–Teller (BET) surface area of the iron oxide@SiO₂-NH₂ was 131.24 m²·g⁻¹ (Table S1), much larger than the naked iron oxide (30.98 m²·g⁻¹) due to the SiO₂ coating and the amino functionalization. The pore diameter was 6.57 nm (the inset). In addition, the thermogravimetric analyses of the iron oxide and iron oxide@SiO₂-NH₂ are shown in Figure 5b, and the iron oxide displays only a gradual weight loss of 5% at 25–300 °C. It may be attributed to the loss of absorbed water. While the iron oxide@SiO₂-NH₂ showed two significant weight losses, the first one is about 7% below 300 ◦C, which can be assigned to the weight loss of the absorbed water, and the second one was around 17% in the range of 300–800 °C. This can be related to the decomposition of amino groups of the iron oxide@SiO₂-NH₂ MNPs at higher temperatures.

3.3. VSM and Zeta Potential Analyses

The magnetic property of the prepared iron oxide@SiO₂-NH₂ MNPs was measured by a VSM, as shown in Figure 6a. According to the hysteresis loops, the saturation magnetization value was 30.70 emu·g⁻¹, which is similar to the literature value of 33.19 emu·g⁻¹ [29] but is much lower than that of the naked iron oxide (56.55 emu·g⁻¹) [29] due to the surface polymer functionalization. The inset shows that this iron oxide@SiO₂-NH₂ composite has a good magnetic response and can be easily collected within 1 min by magnet separation. In Figure 6b, the zeta potential of the iron oxide@SiO₂-NH₂ was also conducted at a solution pH ranging from 3 to 10. According to the results, the isoelectric point value was 4.59, which indicated that when the pH was lower than 4.59, the surface of the iron oxide@SiO₂-NH₂ was positively charged; on the contrary, it became negative. Figure S1 shows the zeta potentials of the bare iron oxide and the iron oxide@SiO₂ MNPs at a solution pH of 4 to 7. As can be seen from Figure S1, the surface of the bare iron oxide MNPs was positive, while it became negative after coating with SiO₂ due to the dissociation of the silicon hydroxyl. Generally, the isoelectric point of SiO₂ in a solution is around 2–3 [30], and the terminal -NH₂ is easy to be protonated into positive -NH₃⁺ groups at lower solution pH. Therefore, the isoelectric point of the iron oxide@SiO₂-NH₂ was affected both by the silica shell and the terminal -NH₂ groups. This explains why the isoelectric point of the iron oxide@SiO₂-NH₂ was less than 7, and its value may have also varied with the grafting ratio of amino groups on the silica shell.

3.4. Adsorption Studies

3.4.1. Factors Affecting Adsorption

Firstly, the dosage effect of the adsorbent on boron adsorption is shown in Figure 7a; it ranged from 0.2 to 5 g·L⁻¹ at 25 °C, and the initial boron concentration was 0.3 mol·L⁻¹. It can be seen that the adsorption capacity increased with the increase in the dosage and then gradually stayed stable when the added dosage was more than 5 g·L⁻¹. This can be attributed to the availability of more adsorbent adsorption sites for the boron adsorption. However, when the effective active sites reached a state of equilibrium due to the utilization of all the boron in the solution at a higher dosage, the adsorption capacity barely increased and stayed constant. The optimum dosage was found to be 5 g·L⁻¹ and was used for the following experiments.

The influence of the initial solution pH from 3 to 10 on the adsorption capacity was also observed and is shown in Figure 7b. It is noted that the adsorption capacity of the iron oxide@SiO₂-NH₂ for boron firstly increased at an acidic pH and then decreased at an alkaline pH. The optimum pH value was about 6. This can be interpreted by the surface charge variation of the adsorbent and solution structure changes of boric acid with changing pH.

The pH dependency of boron species in the 0.3 mol·L⁻¹ boric acid solution is shown in Figure S2 [31]. As can be seen from Figure S2, more than 90% of the boron species was the neutral boric acid when the pH was below 6, while it gradually changed into borate anions with the increasing pH value. As we know, borate anions in a solution tend to react with the ammonium ion NH₄⁺ to produce ammonium borate, such as (NH₄)₂B₄O₇, which is used for the Kjeldahl nitrogen determination [32]. In this study, the terminal -NH₂ groups of the iron oxide@SiO₂-NH₂ would be protonated into -NH₃⁺ groups at the acidic condition (pH ˂ 5), which is not favorable for the main boron species of boric acid adsorption in a solution. Moreover, a lower pH condition is also adverse to the stability of the iron oxide@SiO₂-NH₂ and decreases the adsorption capability, while at the alkaline condition (pH > 7), the surface of the iron oxide@SiO₂-NH₂ MNPs became negative. There was an electrostatic repulsion with the main borate anions in the solution, and the adsorption capability decreased sharply, especially for a pH of more than 10. Therefore, a near-neutral pH was suitable for the stability of -NH₂ groups in the solution, which may adsorb the boric acid by hydrogen bonding interactions between the -NH₂ groups and the boric acid.

In addition, the effects of the initial boron concentration and the contact time on the adsorption are shown in Figure 7c–d, respectively. The adsorption capacity showed a steady increase at a lower boron concentration (0.05 to 0.3 mol·L⁻¹), and a further concentration increase had no significant effect on the adsorption due to the saturation adsorption sites on the adsorbent surface. In Figure 7d, it is noticeable that the adsorption reached its equilibrium quickly after 2 h of contact time.

3.4.2. Adsorption Mechanism

To investigate the possible adsorption mechanisms of boron on the surface of the iron oxide@SiO₂-NH₂ adsorbent, XPS analysis of the adsorbent before and after boron adsorption was carried out, and the spectra are shown in Figure 8. The binding energy of the XPS spectra was calibrated with the standard C 1s peak at 284.6 eV. The Fe element was not detected due to the silica shell on the surface of the iron oxide particle. As can be seen from Figure 8a,b, it is clear that a weak binding energy peak of B 1s at 192.2 eV was observed after adsorption, indicating the successful boron adsorption on the surface of the adsorbent. Compared with the B 1s peak of the pure boric acid at 193.4 eV in Figure S3, there was a blue shift after boron absorption on the magnetic adsorbent. Furthermore, the weak binding energy peak at 197.6 eV before adsorption corresponded to the Cl 2p residual chlorides introduced by the ferric chloride hexahydrate in the preparation of the nanoparticles. In Figure 8c, the N 1s XPS spectra can be deconvoluted into two peaks of the N-C/H at 399.2 eV and the ammonium salt at 401.1 eV [33]. It is worth noting that the N-C/H peak significantly blue-shifted to a higher binding energy of 400.1 eV than the ammonium salt peak after boron adsorption. This may indicate that some possible hydrogen bonds were formed between the nitrogen atom of the terminal-NH₂ groups and the hydroxyl of the boric acid during the boron adsorption process.

3.4.3. Adsorption Kinetics, Isotherms, and Thermodynamic Study

The adsorption kinetics of boron on the iron oxide@SiO₂-NH₂ adsorbent was studied using the pseudo-first-order kinetic [34] and the pseudo-second-order kinetic [35] models, given as follows:

Pseudo-first-order model:

$$\ln \left(q_{\mathrm{e}}-q_t\right)=\ln q_{\mathrm{e}}-k_{1}t$$

(2)

Pseudo-second-order model:

$${\displaystyle \frac{t}{q_t}}={\displaystyle \frac{1}{k_2{q}_{\mathrm{e}}^2}}+{\displaystyle \frac{t}{q_{\mathrm{e}}}}$$

(3)

where q_e and qt are the adsorption capacity (mg·g⁻¹) at equilibrium and time of t, respectively, and k₁ and k₂ are the rate constants (g·mg⁻¹·h⁻¹).

The fitting results are shown in Figure 9, and the parameters calculated were listed in Table S1. Based on the high correlation coefficient R² = 0.9998, it was known that the adsorption followed the pseudo-second-order kinetic model, which indicated that the chemisorption process was mainly controlled by the rate-limiting step.

The isotherm adsorption process of boron at temperatures of 288.15, 298.15, 308.15, and 318.15 K was investigated using the Langmuir [36] and Freundlich [37] isotherm models expressed as the following:

Langmuir:

$$q_{\mathrm{e}}={\displaystyle \frac{q_{\max }·K_{\mathrm{L}}·c_{\mathrm{e}}}{1+c_{\mathrm{e}}·K_{\mathrm{L}}}}$$

(4)

Freundlich:

$$q_{\mathrm{e}}=K_{\mathrm{F}}·c_e^{1/n}$$

(5)

where q_max, q_e are the maximum and equilibrium adsorption capacity adsorption capacity (mg·g⁻¹), respectively, c_e is the boron concentration at equilibrium (mg·L⁻¹), K_L, K_F are the adsorption equilibrium constants (L·mg⁻¹) and the adsorption capacity constant (mg·g⁻¹), respectively, and n is the heterogeneity factor indicating the strength of adsorption.

The fitting curves of the Langmuir isothermal model and Freundlich isothermal model are shown in Figure 10, and the corresponding parameters are listed in Table S2. In Figure 10a, it can be seen that the boron adsorption capacity adsorbed by the iron oxide@SiO₂-NH₂ decreased with the rising of the temperature. This phenomenon is different from other boron adsorbents reported previously [38,39] and can be ascribed to the anti-correlation between the hydrogen bonding and temperature. As shown above, the hydrogen bonding was the main interaction during the boron adsorption process, and it gradually became weaker and weaker with the rising of the temperature, which was not favorable for boron adsorption. According to Figure 10b,c, it is noted that both the Langmuir and Freundlich models could describe the boron adsorption isotherms well since the correlation coefficient R² > 0.96 and the q_max for the iron oxide@SiO₂-NH₂ was 51.19 mg·g⁻¹ at room temperature. But compared with the R² of the Langmuir model, the value of the Freundlich model was slightly higher and may be more suitable for the adsorption isotherms.

3.5. Regeneration Study

Five adsorption–desorption cycles were performed to study the regeneration and reusability of the adsorbent. Based on the alkaline synthesis condition of the iron oxide@SiO₂-NH₂ MNPs, the sodium hydroxide solution (pH = 10) was used as an eluent to desorb the boron-loaded iron oxide@SiO₂-NH₂ MNPs and then regenerated by the mixture of ethanol and ammonia with a pH of 11. The results (Figure 11) showed that there was a slight absorption reduction during the adsorption–desorption process due to the decrease in adsorption sites on the adsorbent. Nevertheless, the iron oxide@SiO₂-NH₂ MNP adsorbent could still maintain a high adsorption capability after five cycles. These results indicated that the iron oxide@SiO₂-NH₂ MNPs have acceptable reusability and stability during the adsorption–desorption process and can be used as potential adsorbents in environmental boron removal treatment.

4. Conclusions

The amino-modified magnetic core-shell iron oxide@SiO₂–NH₂ nanoparticles, synthesized in this study via a facile one-step method, can be used for rapid boron removal from aqueous solutions. The adsorption equilibrium was less than 2 h and the desorption happened easily. The adsorption kinetics of boron can be well described by pseudo-second-order kinetics. Furthermore, the boron adsorption mechanism was mainly due to the hydrogen bonding interaction among boric acid molecules and the terminal -NH₂ groups. The Langmuir and Freundlich models could both describe the boron adsorption isotherms well. Based on the five adsorption–desorption experimental cycles, the regenerated adsorbent still exhibited good adsorption affinity for boron in an aqueous solution, which highlighted that these amino-functionalized magnetic nanoparticles could be used as efficient and recyclable adsorbents for boron recovery from water and wastewater.