Removal of Pb2+, CrT, and Hg2+ Ions from Aqueous Solutions Using Amino-Functionalized Magnetic Nanoparticles

In this paper, a circular economy approach with the adsorption and desorption of heavy metal (HM) ions—i.e., lead (Pb2+), chromium (CrT), and mercury (Hg2+)—from aqueous solutions was studied. Specific and selective binding of HM ions was performed on stabilized and amino-functionalized iron oxide magnetic nanoparticles (γ-Fe2O3@NH2 NPs) from an aqueous solution at pH 4 and 7. For this purpose, γ-Fe2O3@NH2 NPs were characterized by thermogravimetric analysis (TGA), Fourier-transform infrared spectroscopy (FTIR), specific surface area (BET), transmission electron microscopy (TEM), EDXS, and zeta potential measurements (ζ). The effects of different adsorbent amounts (mads = 20/45/90 mg) and the type of anions (NO3−, Cl−, SO42−) on adsorption efficiency were also tested. The desorption was performed with 0.1 M HNO3. The results showed improvement of adsorption efficiency for CrT, Pb2+, and Hg2+ ions at pH 7 by 45 mg of g-Fe2O3@NH2 NPs, and the sequence was as follows: CrT > Hg2+ > Pb2+, with adsorption capacities of 90.4 mg/g, 85.6 mg/g, and 83.6 mg/g, respectively. The desorption results showed the possibility for the reuse of γ-Fe2O3@NH2 NPs with HNO3, as the desorption efficiency was 100% for Hg2+ ions, 96.7% for CrT, and 91.3% for Pb2+.


Introduction
Today, Europe is facing limited stocks of raw materials (RMs), such as heavy metal ions (HM ions) and rare-earth elements (REEs) [1][2][3][4], Even more obviously, in the context of the COVID-19 pandemic, Europe's economy is facing an even larger lack of RMs and HM ions. Moreover, the European Union (EU)'s industry is dependent on imports of large amounts of RMs from the Asian market [1][2][3]. Therefore, the EU Commission was forced to prepare a list of critical raw materials (CRMs) [2,3,5,6], with sustainable strategies to foresee a circular economy based on recycling and reuse of critical REEs [2].
Lead (Pb 2+ ), chromium (total chromium (CrT)), and mercury (Hg 2+ ) ions are listed among the top 20 most hazardous substances [7,8] (accessed on 30 August 2022), since large amounts of HM ions are released into the environment due to agriculture and specific industries, such as the automotive, textile, mining, dye, and electroplating industries, among others [5,[9][10][11]. HM ions dissolved in water are already toxic in small quantities and non-biodegradable, and some are carcinogenic and bioaccumulative, so they need to be treated as priority pollutants and efficiently cleaned [10][11][12].
Among these HMs, mercury has taken the spotlight because it is a global pollutant [13][14][15][16]. Mercury exists in various forms in the natural environment, such as mercurous (Hg 2 +2 ), and mercuric (Hg 2+ ), along with organic mercury-containing methyl and pathways for heavy metal pollutants [46,47]. Despite all of the advantages of superparamagnetic γ-Fe 2 O 3 nanocomposites for use in environmental technologies, the policy debate on their safety should not be ignored. Their toxicity is still an open question, even though much research has recently been carried out on this topic [48,49]. Surface functionalization of γFe 2 O 3 nanoparticles was performed via a sol-gel method involving base-catalyzed hydrolysis and co-condensation of tetra-coordinated alkoxysilanes in an alcohol medium. Tetra-coordinated silanes can be described by the general chemical formula R' x Si(OR) (4−x) , 0 < x < 3, where OR is the hydrolyzable part (e.g., methoxy, ethoxy, etc.) and R' is the non-hydrolyzable part of the structure with functional substituents (e.g., amino, mercapto, carboxy, etc.).
Ideally, it would be expected that the 3-aminopropyltrimethoxysilane (APTMS, (CH 3 O) 3 -Si-(CH 2 ) 3 -NH 2 )) molecules on the surface of the γFe 2 O 3 particles would polymerize into a highly homogeneous crosslinked SiO 2 coating with functional amino (-NH 2 ) groups present. However, the presence of a non-hydrolyzable fraction in the AMPTS structure ((CH 3 O) 3 -Si-(CH 2 ) 3 -NH 2 )) causes steric hindrance, and the electron density on the silicon (Si) atom increases due to the inductive (+I) effect, which decreases the rate of hydrolysis and condensation of the APTMS and increases its tendency for homocondensation. The chemical reactivity is thus slowed down, leading to an undesired heterogeneous distribution of functional amino (-NH 2 ) groups with an insufficient surface coverage of the γFe 2 O 3 nanoparticles [50][51][52].
In contrast to APTMS, under base-catalyzed conditions, the reactivity of tetraethoxysilane (TEOS, Si(OCH 2 CH 3 ) 4 ) is enhanced due to the number and nature of the alkoxide (i.e., ethoxy) groups, which have a key influence on the crosslinking rate. This higher reactivity of TEOS can be attributed to the inductive stabilization of positively charged intermediates and transition states in the hydrolysis and condensation reactions by the ethoxy groups [53]. Therefore, TEOS was used as a crosslinker and APTMS ((CH 3 O) 3 -Si-(CH 2 ) 3 -NH 2 )) was used as a supplier of the -NH 2 functional groups.
In this way, it was possible to create uniform spherical γFe 2 O 3 @SiO 2 -NH 2 core-shell structures with the presence of amino (-NH 2 ) functional groups on the surface of the nanoparticles, which are required for the subsequent binding of heavy metal ions from water [54].
The adsorption process of heavy metal ions for an adsorbent is highly dependent on the initial pH of the solution, owing to its remarkable effect on the speciation of metal ions [5].
If we take a closer look at the speciation of Cr, Pb, and Hg, we can find that at an acidic pH value, the predominant Cr(VI) species consist of H 2 CrO 4 0 , HCrO 4 − , CrO 4 2− , and Cr 2 O 7 2− [5,55] while Cr(III) remains relatively stable in acidic media and is more likely to be oxidized to chromate in alkaline media [56]. For Pb(II) in the pH range from 2 to 6, the dominant form is positively charged Pb 2+ species, while when the pH values increase above 7, other Pb(II) species-including Pb(OH) + , Pb(OH) 2 , and PbO-are usually present [57].
Mercury has two common cations in aqueous solutions: a di-ion, Hg 2 2+ , composed of two singly charged ions; and a doubly charged Hg 2+ . Diagrams of Eh-pH indicate that Hg(I) is stable only within a narrow band of Eh values in acidic solutions, while Hg(II) is the dominant form of the Hg species in most aqueous solutions [58]. The hydrolysis reactions of Hg(II) are significant at pH > 1, and different hydrolyzed forms can be formed depending on the aqueous mercury concentration [59] At low aqueous mercury concentrations, the dominant hydrolysis species formed are HgOH + and Hg(OH) 2(aq) , while at higher mercury concentrations the formation of Hg 2 (OH) 2 2+ and Hg(OH) 3at pH > 13 has been reported [60].
Insoluble metal species will usually not form at pH < 7.2 as long as their concentration is below the solubility limit [61][62][63]. Therefore, at acidic pH values, the removal of positive heavy metal ions is mainly accomplished by adsorption. In contrast, at higher solution pH values, the precipitation of metal hydroxides or even oxides (e.g., Pb(OH) 2 , PbO, CrO 4 2− , etc.) can occur as a consequence of the low solubility of metal ions [57]. Therefore, at higher pH values, precipitation of insoluble species may take place at the same time alongside adsorption in the process of heavy metal removal, negatively affecting the adsorption efficiency [64].
Many studies have shown that metal ions start to precipitate as hydroxides or oxides when the solution pH is above 7.2. To avoid precipitation of the metal ions, all adsorption experiments should be conducted at a pH below 7.2 [56,[61][62][63].
Moreover, the adsorption capacity of heavy metal ions decreases with increasing pH values. Specifically, it was shown that the maximum adsorption capacity of Cr(VI) is observed at a pH of 2 [5]. Moreover, the optimal pH for adsorbing Pb(II) was shown to be around 5.5 [65], whereas it was about 6 for Fe 3 O 4 @SiO 2 -NH 2 magnetic nanoparticles [54,61].
Furthermore, it is generally known that iron oxides (γFe 2 O 3 , Fe 3 O 4 , etc.) suffer from a tendency to aggregate and decompose in acid-regenerated solutions; thus, to avoid the risk of potential dissolution of iron oxide cores at low pH, in this study, we instead used them in adsorption processes at pH > 3, despite silica shell protection (γFe 2 O 3 @SiO 2 -NH 2 ) [61].
Adsorption studies of HM ions from model water by various magnetic nanoparticles (MNPs) and functionalized magnetic nanoparticles (F-MNPs) show that the maximum adsorption capacity of specific HM ions-i.e., for lead [44,61], mercury [80], and chromium [5,81]-can be obtained in less acidic pH.
In Tables 1-3, the adsorption capacity and desorption efficiency are compared for the tested MNPs and amino-functionalized MNPs at the optimal model solution pH values for adsorbing individual HM ions (e.g., Pb 2+ , CrT/Cr 3+ /Cr 6+ , and Hg 2+ ). Table 1 shows comparison of the adsorption capacities and desorption efficiency for Pb 2+ ions by non-functionalized and functionalized MNPs. It can be seen that the adsorption of Pb 2+ ions was tested mostly at acidic pH, and that the adsorption capacity is higher for the cases of functionalized magnetic nanomaterials. Ahmadi et al. (2014) [35] prepared γ-Fe 2 O 3 NPs via the wet chemical method and tested adsorption at pH 7.5, while Nicola et al. (2020) [82] synthesized Fe 3 O 4 @SiO 2 NPs and found that the adsorption capacity on non-functionalized MNPs was relatively low at pH 6.0 (10.55 mg/g) but a shade higher (14.9 mg/g) for SiO 2 -stabilized magnetic nanomaterials [82]. Nicola et al. (2020) [82] also tested the desorption efficiency of Pb 2+ ions with 5% HCl, and the final desorption efficiency was evaluated as 95.7% [82]. Qian [84]. A maximum adsorption capacity of 60 mg/g at pH 5.0 was achieved using composite beads of Zea mays rachis (ZMR) and sodium alginate (AL) as adsorbents [85]. Luo et al. (2021) [86] reported the adsorption of 28.7 mg/g by carbon-doped TiO 2 (C-TiO 2 ) at pH 6.5 for the adsorption of Pb 2+ . The comparison of adsorption capacities showed that the adsorption capacity of Pb 2+ ions depends on the pH of the medium, stabilization, and, to a large extent, the presence of -NH 2 groups. Composite beads of Zea mays rachis (ZMR) and sodium alginate (AL) pH 5.0 60 mg/g [85] Carbon-doped TiO 2 (C-TiO 2 ) pH 6.5 28.7 mg/g [86] * calculated.
There are not many previous studies [32,66] on testing adsorption by γ-Fe 2 O 3 NPs functionalized with an amino (-NH 2 ) group-specifically, by (3-aminopropyl)trimethoxysilane (APTMS) precursors-and to the best of our knowledge, far less research has been conducted on desorption approaches to date.
Although iron oxide and hybrid iron oxide NPs can be removed from aqueous solutions with an outer magnet, their recycling and regeneration possibilities after adsorption have not been sufficiently explored to fill gaps in the circular economy [32,47,87].
Due to these facts, our challenge was to synthesize and investigate the potential of amino-functionalized γ-Fe 2 O 3 MNPs (γ-Fe 2 O 3 @NH 2 NPs), which would allow efficient adsorption and recycling of HM ions at the shock load concentrations present in the model water, preferably at neutral pH, without pretreatment. To compare adsorption efficiencies and capacities, we tested significant concentrations of Pb 2+ , CrT, and Hg 2+ ions using different amounts (m ads = 20/45/90 mg) of the γ-Fe 2 O 3 @NH 2 adsorbent NPs at two different pH values of the initial aqueous solution, i.e., at pH = 7.0, as well as at pH = 4.0. Furthermore, before the performance of adsorption tests, -NH 2 -functionalized γ-Fe 2 O 3 MNPs were characterized with different methods, such as FTIR, BET, TEM, and TGA. Zeta potential changes in γ-Fe 2 O 3 @NH 2 NPs were analyzed to understand the mechanisms taking place during the adsorption and desorption process of Pb 2+ ions. Moreover, to evaluate the adsorbent regeneration, desorption with 0.1 M HNO 3 was tested, which is of great importance for the reuse of adsorption materials and recycling of heavy metals. The prepared γ-Fe 2 O 3 and functionalized γ-Fe 2 O 3 @NH 2 MNPs were also characterized by X-ray powder diffractometry (XRD).

Properties of the Prepared γ-Fe 2 O 3 @NH 2 NPs
This section explains the characterization of the synthesized, stabilized, and functionalized MNPs (γ-Fe 2 O 3 @NH 2 NPs). In addition, the adsorption mechanisms and the results of batch adsorption and desorption experiments are also discussed.

Thermogravimetric Properties
The thermal stability of γ-Fe2O3@NH2 NPs was determined via thermogravimetric analysis (TGA). The results of mass loss during the TGA analysis indicate the possible presence of -NH2 functional groups on the surface of the F-MNPs. Upon heating up to 180 °C, the measured mass loss corresponds to the evaporation of absorbed moisture and NH4OH residue. Further weight loss at heating up to 700 °C is due to the removal of aminopropyl (NH2(CH2)3-) groups from the nanoparticles' surfaces and the consequence of cracking of the remaining siloxane groups (Si-O-Si) [75]. The TGA curve ( Figure 2) shows that the synthesized, stabilized, and functionalized MNPs have good thermal stability. The weight loss during the TGA analysis was 10.3%.
The thermal stability of the particle samples analyzed was in accordance with previous results in the literature for other functionalized NPs [102][103][104][105].

Thermogravimetric Properties
The thermal stability of γ-Fe 2 O 3 @NH 2 NPs was determined via thermogravimetric analysis (TGA). The results of mass loss during the TGA analysis indicate the possible presence of -NH 2 functional groups on the surface of the F-MNPs. Upon heating up to 180 • C, the measured mass loss corresponds to the evaporation of absorbed moisture and NH 4 OH residue. Further weight loss at heating up to 700 • C is due to the removal of aminopropyl (NH 2 (CH 2 ) 3 -) groups from the nanoparticles' surfaces and the consequence of cracking of the remaining siloxane groups (Si-O-Si) [75]. The TGA curve ( Figure 2) shows that the synthesized, stabilized, and functionalized MNPs have good thermal stability. The weight loss during the TGA analysis was 10.3%.
The thermal stability of the particle samples analyzed was in accordance with previous results in the literature for other functionalized NPs [102][103][104][105].

FTIR Spectroscopy
An FTIR analysis of γ-Fe2O3 and γ-Fe2O3@NH2 NPs was performed comparatively to identify the presence of characteristic functional groups related to the amino-silane coating of the γ-Fe2O3 surfaces. The FTIR spectra of γ-Fe2O3 and γ-Fe2O3@NH2 NPs, as well as those of pure TEOS and APTMS precursors, are shown in Figure 3a.
The functional amino-silane-coated γ-Fe2O3 nanoparticles were derived during the sol-gel process from the mixture of TEOS and APTMS precursors according to the experimental details described in Section 2.4. In contrast to the TEOS precursor (Si(OCH2CH3)4), the APTMS precursor ((CH3O)3Si(CH2)3NH2) included a short aliphatic chain (-(CH2)3-) and a terminal amino (-NH2) group in its structure. Thus, the main difference in the FTIR spectra of the TEOS and APTMS precursors is the presence of primary amino (N-H) vibrations in the range of 3400-3300 cm −1 of the APTMS spectra, while both spectra are identical to the occurrence of C-H vibrations in the range of 3000-2800 cm −1 and Si-O-Si vibrations in the range of 1100-1000 cm −1 , which are common characteristics of alkoxysilanes.
As shown in Figure 3a, the formation of the γ-Fe2O3 structure is closely related to the occurrence of Fe-O bending and stretching vibrations in the range of 650-550 cm −1 . The broad band at 3406 cm −1 observed for the γ-Fe2O3 NPs in the wavenumber region 3550-3200 cm −1 can be assigned to intermolecular O-H stretching ( Figure 3a).
As opposed to γ-Fe2O3 NPs, asymmetric stretching vibrations of Si-O-Si bonds at 1050 cm −1 indicate the formation of a silica (SiO2) shell in the γ-Fe2O3@NH2 samples. Moreover, two weak bands can be observed for the γ-Fe2O3@NH2 samples in Figure 3a, characteristic of primary amines, due to the asymmetric and symmetric N-H vibrations in the range of 3400-3300 cm −1 -more precisely, at 3356 cm −1 and 3281 cm −1 , respectively. These primary amino peaks in the source spectra of γ-Fe2O3@NH2 NPs were not sufficiently visible, but enlarged individual peak areas confirmed their presence (Figure 3b). Specifically, the primary amine (NH2) vibrations occurred in the same wavenumber region as the intermolecular O-H stretching [106]. Because the polarity of the N-H bonds in amines is weaker than that of the O-H bonds, the absorption band of N-H is not as intense as that of O-H, which usually shows stronger and broader absorption bands that are much easier to

FTIR Spectroscopy
An FTIR analysis of γ-Fe 2 O 3 and γ-Fe 2 O 3 @NH 2 NPs was performed comparatively to identify the presence of characteristic functional groups related to the amino-silane coating of the γ-Fe 2 O 3 surfaces. The FTIR spectra of γ-Fe 2 O 3 and γ-Fe 2 O 3 @NH 2 NPs, as well as those of pure TEOS and APTMS precursors, are shown in Figure 3a.
The functional amino-silane-coated γ-Fe 2 O 3 nanoparticles were derived during the solgel process from the mixture of TEOS and APTMS precursors according to the experimental details described in Section 2.4. In contrast to the TEOS precursor (Si(OCH 2 CH 3 ) 4 ), the APTMS precursor ((CH 3 O) 3 Si(CH 2 ) 3 NH 2 ) included a short aliphatic chain (-(CH 2 ) 3 -) and a terminal amino (-NH 2 ) group in its structure. Thus, the main difference in the FTIR spectra of the TEOS and APTMS precursors is the presence of primary amino (N-H) vibrations in the range of 3400-3300 cm −1 of the APTMS spectra, while both spectra are identical to the occurrence of C-H vibrations in the range of 3000-2800 cm −1 and Si-O-Si vibrations in the range of 1100-1000 cm −1 , which are common characteristics of alkoxysilanes.
As shown in Figure 3a, the formation of the γ-Fe 2 O 3 structure is closely related to the occurrence of Fe-O bending and stretching vibrations in the range of 650-550 cm −1 . The broad band at 3406 cm −1 observed for the γ-Fe 2 O 3 NPs in the wavenumber region 3550-3200 cm −1 can be assigned to intermolecular O-H stretching ( Figure 3a).
As opposed to γ-Fe 2 O 3 NPs, asymmetric stretching vibrations of Si-O-Si bonds at 1050 cm −1 indicate the formation of a silica (SiO 2 ) shell in the γ-Fe 2 O 3 @NH 2 samples. Moreover, two weak bands can be observed for the γ-Fe 2 O 3 @NH 2 samples in Figure 3a, characteristic of primary amines, due to the asymmetric and symmetric N-H vibrations in the range of 3400-3300 cm −1 -more precisely, at 3356 cm −1 and 3281 cm −1 , respectively. These primary amino peaks in the source spectra of γ-Fe 2 O 3 @NH 2 NPs were not sufficiently visible, but enlarged individual peak areas confirmed their presence (Figure 3b). Specifically, the primary amine (NH 2 ) vibrations occurred in the same wavenumber region as the intermolecular O-H stretching [106]. Because the polarity of the N-H bonds in amines is weaker than that of the O-H bonds, the absorption band of N-H is not as intense as that of O-H, which usually shows stronger and broader absorption bands that are much easier to identify. Primary amines have also a medium-to-strong absorption band in the wavenumber region 1650-1580 cm −1 , which was identified at 1598 cm −1 for the γ-Fe 2 O 3 @NH 2 NPs [107]. identify. Primary amines have also a medium-to-strong absorption band in the wavenumber region 1650-1580 cm −1 , which was identified at 1598 cm −1 for the γ-Fe2O3@NH2 NPs [107].

Specific Surface Area
The specific surface area of the prepared γ-Fe2O3 and γ-Fe2O3@NH2 MNPs was measured by the Brunauer-Emmett-Teller (BET) method. The obtained BET curves are shown in Figure 4.
The BET analysis showed a specific surface area of 99.9 m 2 /g for γ-Fe2O3 and 41.3 m 2 /g for γ-Fe2O3@NH2. According to the Barrett-Joyner-Halenda (BJH) adsorption method, the average pore size was found to be 6.4 nm for the γ-Fe2O3 NPs, with a total pore volume of 0.378037 cm 3 /g, while for the BJH desorption the average pore size for the γ-Fe2O3 NPs increased to 6.7 nm, with a total pore volume of 0.407662 cm 3 /g, suggesting a mesoporous structure of the γ-Fe2O3 sample, with a typical type IV experimental N2 gas isotherm according to the IUPAC classification [108], as shown in Figure 4. In contrast to γ-Fe2O3, the γ-Fe2O3@NH2 sample showed a BET isotherm with a narrower hysteresis, in-   The BET analysis showed a specific surface area of 99.9 m 2 /g for γ-Fe 2 O 3 and 41.3 m 2 /g for γ-Fe 2 O 3 @NH 2 . According to the Barrett-Joyner-Halenda (BJH) adsorption method, the average pore size was found to be 6.4 nm for the γ-Fe 2 O 3 NPs, with a total pore volume of 0.378037 cm 3 /g, while for the BJH desorption the average pore size for the γ-Fe 2 O 3 NPs increased to 6.7 nm, with a total pore volume of 0.407662 cm 3 /g, suggesting a mesoporous structure of the γ-Fe 2 O 3 sample, with a typical type IV experimental N 2 gas isotherm according to the IUPAC classification [108], as shown in Figure 4. In contrast to γ-Fe 2 O 3 , the γ-Fe 2 O 3 @NH 2 sample showed a BET isotherm with a narrower hysteresis, indicating a decrease in the porosity of the as-prepared γ-Fe 2 O 3 sample, most likely due to the presence of the homogeneous silicate coating. For BJH adsorption, the average pore size was found to be 5.8 nm for the γ-Fe 2 O 3 @NH 2 NPs, with a total pore volume of 0.090762 cm 3 /g, while for the BJH desorption the average pore size for the γ-Fe 2 O 3 @NH 2 NPs increased to 6.0 nm, with a total pore volume of 0.090311 cm 3 /g.
According to the specific surface area (BET) at a relative pressure (p/p 0 ) of 0.3, the calculated average particle size was 11.6 nm for γ-Fe 2 O 3 and 27.9 nm for γ-Fe 2 O 3 @NH 2 NPs [109,110]

Morphological Properties
The results of the TEM analysis (Figure 5a) represent the relatively spherical morphology of the γ-Fe 2 O 3 MNPs, with a particle size distribution of 13 ± 1 nm, while the particle size distribution of the functionalized γ-Fe 2 O 3 @NH 2 MNPs was 17 ± 1 nm (magnetic core 13 ± 1 nm and surface coating 4 ± 1 nm). The electron diffraction pattern of the γ-Fe 2 O 3 MNPs inset in Figure 5b indicates the crystalline nature of the as-prepared powders, with concentric diffraction rings characteristic of a cubic spinel crystal structure.

Morphological Properties
The results of the TEM analysis (Figure 5a) represent the relatively spherical morphology of the γ-Fe2O3 MNPs, with a particle size distribution of 13 ± 1 nm, while the particle size distribution of the functionalized γ-Fe2O3@NH2 MNPs was 17 ± 1 nm (magnetic core 13 ± 1 nm and surface coating 4 ± 1 nm). The electron diffraction pattern of the γ-Fe2O3 MNPs inset in Figure 5b indicates the crystalline nature of the as-prepared powders, with concentric diffraction rings characteristic of a cubic spinel crystal structure.
The EDXS spectra of the γ-Fe2O3 and γ-Fe2O3@NH2 MNPs are shown in Figure 6a,b, respectively. Strong peaks for iron (Fe) and oxygen (O) can be seen in the EDXS spectrum in Figure 6a, indicating the formation of the γ-Fe2O3 MNPs. In contrast, the EDXS spectrum of the γ-Fe2O3@NH2 MNPs shows that they contain significant amounts of silicon (Si), alongside iron (Fe) and oxygen (O), suggesting the success of the surface functionalization of γ-Fe2O3 MNPs with APTMS precursor molecules and, thus, the formation of the γ-Fe2O3@NH2 MNPs. The lack of a nitrogen (N) peak is expected, due to its low Z-number and overlapping with the K-alpha C and O peaks. The larger peaks towards the right in both EDXS spectra are the copper (Cu) signals sourced from the TEM copper-grid-supported transparent carbon foil. The EDXS spectra of the γ-Fe 2 O 3 and γ-Fe 2 O 3 @NH 2 MNPs are shown in Figure 6a,b, respectively. Strong peaks for iron (Fe) and oxygen (O) can be seen in the EDXS spectrum in Figure 6a, indicating the formation of the γ-Fe 2 O 3 MNPs. In contrast, the EDXS spectrum of the γ-Fe 2 O 3 @NH 2 MNPs shows that they contain significant amounts of silicon (Si), alongside iron (Fe) and oxygen (O), suggesting the success of the surface functionalization of γ-Fe 2 O 3 MNPs with APTMS precursor molecules and, thus, the formation of the γ-Fe 2 O 3 @NH 2 MNPs. The lack of a nitrogen (N) peak is expected, due to its low Z-number and overlapping with the K-alpha C and O peaks. The larger peaks towards the right in both EDXS spectra are the copper (Cu) signals sourced from the TEM copper-grid-supported transparent carbon foil.

Zeta Potential
The zeta potential was measured for bare MNPs (γ-Fe 2 O 3 ) and amino-functionalized MNPs (γ-Fe 2 O 3 @NH 2 ), as depicted in Figure 7. For bare, stabilized MNPs, the zeta potential is positive at low pH due to the presence of OH 2 + . As the pH of the solution increases, the potential decreases and approaches negative potential at high pH, due to the presence of O − . The measured isoelectric point of the bare MNPs was 8.76 (measured potential −0.743 mV). At this value, the concentration of protonated and deprotonated amino groups is the same. Meanwhile, the measured isoelectric point for functionalized MNPs was at pH 12.1 (measured potential +0.161 mV) =, indicating successful MNP functionalization. This difference in the isoelectric point is due to the presence of amino groups on MNPs, resulting in a functionalized magnetic nanomaterial with a negative charge above pH = 12.1. The zeta potential was measured for bare MNPs (γ-Fe2O3) and amino-functionalized MNPs (γ-Fe2O3@NH2), as depicted in Figure 7. For bare, stabilized MNPs, the zeta potential is positive at low pH due to the presence of OH2 + . As the pH of the solution increases, the potential decreases and approaches negative potential at high pH, due to the presence of O − . The measured isoelectric point of the bare MNPs was 8.76 (measured potential −0.743 mV). At this value, the concentration of protonated and deprotonated amino groups is the same. Meanwhile, the measured isoelectric point for functionalized MNPs was at pH 12.1 (measured potential +0.161 mV) =, indicating successful MNP functionalization. This difference in the isoelectric point is due to the presence of amino groups on MNPs, resulting in a functionalized magnetic nanomaterial with a negative charge above pH = 12.1 .

Adsorption Mechanisms
The solution pH is a key parameter of the effectiveness of HM ions' adsorption. HM ions have specific forms at different pH values; moreover, the adsorbent surface charge and protonation degree of the adsorbent surface coating (i.e., amino groups) are dependent on the pH [111,112]. In general, HM ions' adsorption on γ-Fe2O3@NH2 NPs includes three sorption mechanisms, i.e., ion exchange, surface complexation, and electrostatic attraction [5]; the specific adsorption mechanism predominantly depends on the solution's pH value [5].
We tested the adsorption of Pb 2+ , CrT, and Hg 2+ ions at different pH values, i.e., pH 4 and 7. At different pH values, adsorption takes place by a different mechanism for each metal ion [5].
The adsorption of Pb 2+ ions is entirely dependent on the pH value [70]. The adsorbent surface is negatively charged at alkaline pH, which indicates the deprotonated form of -NH2 functional groups. The behavior of -NH2 groups on the adsorbent material according to the pH is shown by Equations (1) and (2) [111]: At pH 7, -NH2 groups are deprotonated, causing a negatively charged adsorbent surface, while lead ions are mostly in Pb (OH) + form, which causes high electrostatic

Adsorption Mechanisms
The solution pH is a key parameter of the effectiveness of HM ions' adsorption. HM ions have specific forms at different pH values; moreover, the adsorbent surface charge and protonation degree of the adsorbent surface coating (i.e., amino groups) are dependent on the pH [111,112]. In general, HM ions' adsorption on γ-Fe 2 O 3 @NH 2 NPs includes three sorption mechanisms, i.e., ion exchange, surface complexation, and electrostatic attraction [5]; the specific adsorption mechanism predominantly depends on the solution's pH value [5].
We tested the adsorption of Pb 2+ , CrT, and Hg 2+ ions at different pH values, i.e., pH 4 and 7. At different pH values, adsorption takes place by a different mechanism for each metal ion [5].
The adsorption of Pb 2+ ions is entirely dependent on the pH value [70]. The adsorbent surface is negatively charged at alkaline pH, which indicates the deprotonated form of -NH 2 functional groups. The behavior of -NH 2 groups on the adsorbent material according to the pH is shown by Equations (1) and (2) [111]: At pH 7, -NH 2 groups are deprotonated, causing a negatively charged adsorbent surface, while lead ions are mostly in Pb (OH) + form, which causes high electrostatic attraction between Pb 2+ ions and the negatively charged material surface and, consequently, high adsorption efficiency [82,111]. On the other hand, acidic conditions cause the transformation of -NH 2 groups into -NH 3 + form, resulting in fewer available active sites for Pb 2+ ions. Because of that, the adsorption efficiency of Pb 2+ ions drops under acidic conditions (pH < 7) [5,111].
On the other hand, at alkaline pH, negatively charged chromate ions (CrO 4 2− ) are the predominant form [81,87,113]. At pH > 7, γ-Fe 2 O 3 @NH 2 NPs' surfaces are also negatively charged [5] due to the deprotonated form of the amino functional groups. A double-negative charge of the adsorbent surface and chromate decreases the adsorption efficiency [72].
In our study, zeta potential played an important role in the adsorption mechanism. The zeta potential of our γ-Fe 2 O 3 @NH 2 NPs was 8.76; at lower pH, amino groups on the material's surface were mainly present in protonated form (-NH 3 + ). We tested the adsorption of CrT ions at pH 4 and 7. At pH 4, the functional groups were mainly in -NH 3 + form, while at pH 7 the amino groups were still in protonated form. Consequently, many active sites were present on the surface of the γ-Fe 2 O 3 @NH 2 NPs, so their adsorption capacity was very high. The adsorption efficiency at pH 4 was low due to the instability of γ-Fe 2 O 3 @NH 2 NPs in acidic conditions-the adsorbent material is soluble in acidic media, i.e., at pH < 4.
Hg 2+ readily reacts with OH − to form Hg 2 (OH) 2 precipitates under alkaline conditions. The adsorption of Hg 2+ ions is predominantly influenced by the concentration of hydronium ions in aqueous solutions. The change in adsorption at varying pH levels is because the concentration of surface charges governs the adsorbent particles and the degree of ionization of the ions to be removed [111,115,116]. There is a variety of literature suggesting that the adsorption of Hg 2+ ions favors neutral and basic pH. The rationale for more adsorption of Hg 2+ ions using amino groups at neutral and basic pH is that the amino group obtains a net positive charge at acidic pH and the Hg 2+ ions are also positive; hence, the adsorption is made unfavorable by the repulsive force. The above rationale for mercury species in aqueous solution was theoretically determined as a function of pH by modeling chemical equilibrium using MINEQL+ software (Environmental Research Software, Hallowell, ME, USA) [80,117].

Effects of pH
Batch adsorption experiments of Pb 2+ , CrT, and Hg 2+ ions for different adsorption times were performed at two pH values, i.e., pH 4 and 7 (Figures 8-11). The results show that the adsorption of Pb 2+ and Hg 2+ ions is more efficient at pH 7. Such results were expected, due to the opposite charges of the Pb 2+ ions and the surface of the adsorption material. The opposite surface charges caused strong electrostatic interactions and high material uptake. ciency already after 1 min. When the adsorption time was extended, the efficiency stayed high, and the maximal adsorption capacity (90.4 mg/g) was achieved after 12 h.
For Hg 2+ ions, the maximal adsorption efficiency of 84.3% displayed a corresponding adsorption capacity of 85.6 mg/g, which was reached after 30 min of adsorption time. At pH 4, Hg 2+ ions showed a low adsorption efficiency of 17%, with a corresponding adsorption capacity of 16.2 mg/g at 30 h. As demonstrated in various studies [80,111,[115][116][117], the increase in pH from 4 to 7 also facilitated the maximal adsorption efficiency.

Effect of Adsorbent Mass
The adsorption of CrT ions on 45 mg of γ-Fe2O3@NH2 NPs showed excellent results (qt after 1 min was 81.4 mg/g). To determine the optimal adsorbent mass, we tested different amounts of γ-Fe2O3@NH2 NPs. In adsorption experiments, 20, 45, and 90 mg of γ-Fe2O3@NH2 NPs were investigated under optimal adsorption conditions (pH = 7, c = 200 mg CrT/L and RT). Adsorption tests were performed only for 1, 4, 8, 24, and 30 h, as we expected that longer specific adsorption times would be required with smaller amounts of γ-Fe2O3@NH2 NPs.
The results of CrT ions' adsorption showed excellent adsorption efficiency (>99.2%) for all tested amounts of γ-Fe2O3@NH2 NPs at all tested specific adsorption times ( Figure  12). The only exception was the test using 20 mg of γ-Fe2O3@NH2 NPs after 1 h. The adsorption efficiency on 20 mg of γ-Fe2O3@NH2 NPs was only 35.3%, indicating insufficient The adsorption capacity at pH 4 and 7 slowly increased with longer specific adsorption times. At pH 4, the maximal adsorption capacity of Pb 2+ ions was 53.5 mg/g, which was detected after 30 h. At pH 7, the maximal adsorption capacity of 83.6 mg/g was achieved already after 12 h. At both tested pH values, the adsorption of Pb 2+ ions slowly increased with a longer adsorption time. This indicates that the adsorption of Pb 2+ ions is a slow process but, more importantly, the process is efficient-especially at pH 7.
Adsorption of CrT ions was much faster and very efficient at the same time. At pH 4, the adsorption efficiency was lower than 30%, and the maximal adsorption capacity was 24.0 mg/g. γ-Fe 2 O 3 @NH 2 NPs were less stable in acidic conditions, which was the main reason for the lower material uptake. At pH 7, we achieved 99.9% adsorption efficiency already after 1 min. When the adsorption time was extended, the efficiency stayed high, and the maximal adsorption capacity (90.4 mg/g) was achieved after 12 h.
For Hg 2+ ions, the maximal adsorption efficiency of 84.3% displayed a corresponding adsorption capacity of 85.6 mg/g, which was reached after 30 min of adsorption time. At pH 4, Hg 2+ ions showed a low adsorption efficiency of 17%, with a corresponding adsorption capacity of 16.2 mg/g at 30 h. As demonstrated in various studies [80,111,[115][116][117], the increase in pH from 4 to 7 also facilitated the maximal adsorption efficiency.

Effect of Adsorbent Mass
The adsorption of CrT ions on 45 mg of γ-Fe 2 O 3 @NH 2 NPs showed excellent results (q t after 1 min was 81.4 mg/g). To determine the optimal adsorbent mass, we tested different amounts of γ-Fe 2 O 3 @NH 2 NPs. In adsorption experiments, 20, 45, and 90 mg of γ-Fe 2 O 3 @NH 2 NPs were investigated under optimal adsorption conditions (pH = 7, c = 200 mg CrT/L and RT). Adsorption tests were performed only for 1, 4, 8, 24, and 30 h, as we expected that longer specific adsorption times would be required with smaller amounts of γ-Fe 2 O 3 @NH 2 NPs.
The results of CrT ions' adsorption showed excellent adsorption efficiency (>99.2%) for all tested amounts of γ-Fe 2 O 3 @NH 2 NPs at all tested specific adsorption times ( Figure 12). The only exception was the test using 20 mg of γ-Fe 2 O 3 @NH 2 NPs after 1 h. The adsorption efficiency on 20 mg of γ-Fe 2 O 3 @NH 2 NPs was only 35.3%, indicating insufficient adsorbent mass. Nevertheless, the adsorption efficiency reached 99.9% after 24 h, showing that the adsorption of CrT ions with a smaller amount of γ-Fe 2 O 3 @NH 2 NPs required a longer adsorption time. Meanwhile, the adsorption of CrT ions on 45 and 90 mg was equal; thus, based on the results of adsorption on 20/45/90 mg of γ-Fe 2 O 3 @NH 2 NPs, we concluded that 45 mg was the optimal mass of adsorbent.   The results showed no effects of different anions; furthermore, the adsorption of CrT ions remained quick, and after 1 min the adsorption efficiency rate was 99.9% for all tested anions ( Figure 13). The results showed no effects of different anions; furthermore, the adsorption of CrT ions remained quick, and after 1 min the adsorption efficiency rate was 99.9% for all tested anions ( Figure 13).

Desorption
To verify the possibility of recycling HM ions and reusing adsorption materials, desorption of Pb 2+ , CrT, and Hg 2+ ions was performed. Due to the higher adsorption capacity of metal ions at 1140 and 1800 min, adsorption was tested for longer specific adsorption times. Desorption was performed in one cycle because of material loss during the desorption process. The results of desorption showed that the γ-Fe 2 O 3 @NH 2 NPs enabled high desorption efficiency for the samples on the surface of which the HM ions' adsorption was performed at pH 7. This result indicates the electrostatic binding of HM ions on the adsorption material's surface. Electrostatic binding of HM ions is weaker than covalent interactions, which probably appear at lower pH, i.e., pH 4. For this reason, only desorption results of samples after adsorption was performed at pH 7 are reported in this work ( Figure 14).
The first desorption cycle performed with 0.1 M HNO 3 was more efficient for Hg 2+ , CrT, and Pb 2+ ions. For Pb 2+ ions, the desorption efficiency was 91.3%; for CrT ions it was 96.7%; and for Hg 2+ ions it was 100%. From the obtained results, it was possible to determine that higher desorption efficiency was achieved for all tested HM ions with a longer specific adsorption time (i.e., 30 h).
The desorption efficiency of Hg 2+ ions showed that the γ-Fe 2 O 3 @NH 2 NPs enabled high desorption efficiency (100%) for all samples, with the adsorption process being carried out at pH 4 and 7.
CrT, and Pb 2+ ions. For Pb 2+ ions, the desorption efficiency was 91.3%; for CrT ions it was 96.7%; and for Hg 2+ ions it was 100%. From the obtained results, it was possible to determine that higher desorption efficiency was achieved for all tested HM ions with a longer specific adsorption time (i.e., 30 h).
The desorption efficiency of Hg 2+ ions showed that the γ-Fe2O3@NH2 NPs enabled high desorption efficiency (100%) for all samples, with the adsorption process being carried out at pH 4 and 7.

Materials
For the lab-scale synthesis, stabilization, and functionalization of γ-

Synthesis of MNPs
The synthesis of the maghemite (γ-Fe 2 O 3 ) MNPs was carried out as described in our previous studies [32]. First, 30 mL of NH 4 OH was added to a 100 mL glass flask and heated up to 90 • C in an oil bath, under constant stirring at 220 rpm. Afterward, 50 mL of an aqueous solution prepared at a stoichiometric ratio of 1:2 using FeSO 4 7H 2 O and Fe 2 (SO 4 ) 3 xH 2 O salts was added to the reaction flask. Synthesis then proceeded in alkaline conditions at pH 10 for 1 h at 90 • C. After the reaction was finished, the suspension was cooled down to room temperature (RT), and the magnetic sediment was settled out for 30 min using an external permanent magnet. After magnetic separation, the supernatant was decanted and discarded. The colloid was rinsed several times with dH 2 O, centrifuged at 3500 rpm for 15 min and, finally, separated and allowed to settle on the external magnet overnight.

Stabilization of MNPs
The rinsed γ-Fe 2 O 3 MNPs were suspended in 100 mL of NH 4 OH overnight at RT under constant stirring (330 rpm) for the stabilization process. After 16 h, the stabilized maghemite MNPs were separated into two phases overnight on an external magnet. The upper phase was decanted, and the colloid was prepared for further functionalization.

Characterization of Amino-Functionalized γ-Fe 2 O 3 MNPs
Characterization of the lab-scale amino-functionalized γ-Fe 2 O 3 MNPs (γ-Fe 2 O 3 @NH 2 NPs) was performed using the appropriate method after each preparation phase (i.e., synthesis, stabilization, functionalization). For characterization purposes, the γ-Fe 2 O 3 @NH 2 MNPs were dried at 90 • C overnight, and the mass of the obtained dried particles was determined. The prepared γ-Fe 2 O 3 and functionalized γ-Fe 2 O 3 @NH 2 MNPs were characterized by X-ray powder diffractometry (XRD) (Bruker D4 Endeavor, Bruker, Billerica, MA, USA). The thermogravimetric properties were analyzed with a 4000 TGA PerkinElmer analyzer (PerkinElmer, Waltham, MA, USA), FTIR spectra were recorded with a Spectrum Two (PerkinElmer, Waltham, MA, USA), and specific surface area (BET) was measured with a TriStar II 3020 (Micromeritics Instrument Corporation, Norcross, GA, USA). The morphology of the synthesized γ-Fe 2 O 3 @NH 2 MNPs was investigated using a transmission electron microscope (JEM 2100 JEOL, JEOL Ltd, Musashino Akishima, Tokyo, Japan) equipped with an energy-dispersive X-ray spectroscopy (EDXS) unit and a CCD camera to capture images.

Adsorption of Heavy Metal Ions in Aqueous Solutions
Batch adsorption tests of Pb 2+ , CrT, and Hg 2+ ions were performed. The initial concentration of 200 mg/L of HM ions in the model water solutions was prepared by dissolving Pb (NO 3 ) 2 , Cr (NO 3 ) 3 9H 2 O, and Hg (NO 3 ) 2 H 2 O in a 1 L flask. The adsorption efficiency (R %) and adsorption capacity (q t mg/g) at different pH of the model solution and adsorption at the defined time were calculated using Equations (6) and (7). Additionally, the effect of different adsorbent mass (m ads = 20, 45, 90 mg) was tested, and the impact of various anions (e.g., NO 3 − , Cl − , and SO 4 2− ) on the adsorption of CrT ions was investigated. For Pb 2+ and Hg 2+ ions, only adsorption at different pH values was tested.
The initial pH values of the solutions were measured and set to pH 4 with 0.1 M and 1 M HNO 3 to simulate an acidic environment, which does not affect the γ-Fe 2 O 3 @NH 2 NPs, while 1 M NaOH was used to adjust the pH to 7 to simulate actual wastewater conditions. Adsorption was conducted in 50 mL plastic centrifuges into which 20, 45, or 90 mg of the lab-scale γ-Fe 2 O 3 @NH 2 NPs were weighed. Then, 20 mL of the prepared model salt solution was added to the γ-Fe 2 O 3 @NH 2 NPs for selected specific adsorption times (1,5,10,15,30,60,240,480,720,1140, and 1800 min). All tests were carried out at RT. To separate the γ-Fe 2 O 3 @NH 2 NPs from the supernatant after adsorption, a centrifuge (4500 rpm, 15 min) (UNIVERSAL 320, Andreas Hettich GmbH & Co. KG, Tuttlingen, Germany) and an external magnet were used. The supernatant was decanted into a glass vial; meanwhile, the γ-Fe 2 O 3 @NH 2 NPs were washed two times with 10 mL of dH 2 O.
The concentration of the HM ions in the supernatant was measured via atomic adsorption spectroscopy (AAS PerkinElmer, PerkinElmer, Waltham, MA, USA) and inductively coupled plasma optical emission spectrometry (ICP-OES, SPECTRO CITROS VI-SION, SPECTRO Analytical Instruments GmbH, Kleve, Germany) for Hg 2+ . For both analyses, the supernatants were acidified with HNO 3 (0.5 mL of acid to 10 mL of the supernatant sample).
The adsorption efficiency (R %) was calculated using Equation (6) [87]: where R (%) is the adsorption efficiency, C 0 (mg/L) is the initial concentration of HM ions, and C t (mg/L) is the residual concentration of HM ions. The adsorption capacity was calculated using Equation (7) [87]: where q t (mg/g) is the adsorption capacity, C 0 (mg/L) is the initial concentration of HM ions, C t (mg/L) is the residual concentration of HM ions, V (mL) is the solution volume, and m (mg) is the adsorption material mass.

Desorption of HM Ions and Regeneration Experiments
The desorption experiments for Pb 2+ , CrT, and Hg 2+ ions were conducted to evaluate the recyclability of γ-Fe 2 O 3 @NH 2 NPs. Desorption tests were performed immediately after specific adsorption times-i.e., 1, 1140, and 1800 min-and after rinsing twice with 10 mL of distilled water. Desorption was performed at RT with 20 mL of 0.1 M HNO 3 added to 45 mg of adsorbent material. Desorption was in contrast to adsorption performed in dynamic mode with an IKA MS3 digital shaker (IKA-Werke GmbH & Co. KG, Staufen, Germany) at minimal rpm. The desorption efficiency was evaluated with AAS for Pb 2+ and CrT, and with ICP-OES for Hg 2+ ions.

Conclusions
In this study, stabilized and amino-functionalized magnetic nanoparticles (γ-Fe 2 O 3 @NH 2 NPs) with a diameter of 17 ± 1 nm were synthesized, characterized, and used as adsorbents for Pb 2+ , CrT, and Hg 2+ ions. Adsorbent characterization showed that γ-Fe 2 O 3 @NH 2 NPs have good thermal stability. The particles were successfully stabilized, and amino-functionalization was confirmed with FTIR spectroscopy.
The adsorption process was carried out in aqueous solutions at pH 4 and 7. The adsorption results showed the highest adsorption efficiency and capacity at pH 7 for all investigated heavy metal (HM) ions, i.e., Pb 2+ , CrT, and Hg 2+ . The adsorption efficiency was the highest and quickest for CrT > Hg 2+ > Pb 2+ ions. The maximal adsorption capacity for Hg 2+ ions was achieved in 30 min, at 85.6 mg/g; for CrT and Pb 2+ ions, the maximal adsorption capacities were achieved after 12 h and were 90.4 mg/g and 83.6 mg/g, respectively. Experiments with different amounts of γ-Fe 2 O 3 @NH 2 NPs (20/45/90 mg) showed that the optimal mass of adsorbent was 45 mg. Moreover, under optimal adsorption conditions (pH = 7, m ads = 45 mg, c = 200 mg CrT/L, and RT), different anions-i.e., NO 3 − , Cl − , and SO 4 2− -showed no effect on the adsorption efficiency of CrT ions. A study of the desorption process with 0.1 M HNO 3 for 1 h showed the possibility of reusing γ-Fe 2 O 3 @NH 2 NPs. Desorption was effective only for γ-Fe 2 O 3 @NH 2 NPs after the adsorption process was performed at neutral pH. We observed excellent desorption efficiency for Hg 2+ (100%), CrT (96.7%), and Pb 2+ (91.3%) ions.
The adsorption-desorption results showed that γ-Fe 2 O 3 @NH 2 NPs have great ability and potential for specific and selective binding of HM ions and show excellent potential for real application in the circular economy. For this reason, further investigation of the circular adsorption-desorption process for different HM ions (such as copper, iron, and cadmium) in a single or binary system should be carried out in the near future.
The use of functionalized magnetic nanomaterials as adsorbents in this study showed that they combine the advantages of magnetic properties-which allow the removal of pollutants from water using an external magnetic field-with the properties of other functional materials, improving their adsorption, separation, and regeneration properties. Such adsorption materials are capable of removing the main components of inorganic pollutants, such as heavy metal ions, under different concentrations and pH conditions, due to their chemical and physical stability.
At the same time, this study showed that such functional magnetic nanoparticles, in conjunction with existing treatment technologies, can offer tremendous potential for the effective treatment of water and wastewater. Due to their unique properties related to magnetism and their surface and structural properties, these adsorption materials offer many alternative applications in many other fields. Their use has been growing in recent years, particularly in the recycling of critical materials-including rare-earth elements, which are now used in all high-tech products and are almost impossible to replace because their properties are unique or "rare", which is why they are so highly valued, and their extraction and production pose major problems in terms of environmental pollution. In addition, such functionalized magnetic nanoparticles could also be effectively used to remove organic and biological pollutants such as organic dyes, fluoridated and chlorinated organic compounds, pesticides, bioactive compounds, etc., which are often found in groundwater, drinking water, and wastewater.
Despite the vast potential shown by functionalized magnetic nanomaterials as adsorbents, most of them are still at the laboratory research stage. The lack of legislation and regulation and the issue of toxicity of nanomaterials, which should also not be ignored, represent the major obstacles encountered in the remediation of water and wastewater with nanomaterials, while many other obstacles associated with their use are only temporary, such as high costs and technical handling.
Although many studies have been carried out on the adsorption of heavy metal ions, the mechanism of their interaction with adsorbents is, in many cases, not fully understood. Therefore, more research on the interactions between functionalized magnetic nanomaterials and pollutants is expected in the future, as they are of key importance for the design and improvement of the properties of adsorbents, but the lack of knowledge on their environmental and human impacts has to be taken into account in order to move towards a justification of their use in real environments.