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

Abstract

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 (m_ads = 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⁺.

1. 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 (Pb²⁺), chromium (total chromium (CrT)), and mercury (Hg²⁺) 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 (Hg₂⁺²), and mercuric (Hg²⁺), along with organic mercury-containing methyl and ethyl groups. It is pertinent that the highly notorious form of methylmercury is caused by the methylation of inorganic [15] and elemental mercury [17] that is present in the aquatic environment by sulfate-reducing bacteria such as Desulfovibrio desulfuricans. Methylmercury can bioaccumulate and biomagnify in the oceanic food chain to reach 10⁶ times the concentrations that have caused several tragedies in the past, such as the Minamata tragedy in Japan [15] Negative toxic effects of versatile and highly mobile stable forms of chromium (Cr³⁺ and Cr⁶⁺) are a constant threat to humans and the environment. Depending on the pH, they can be present in acidic media (pH 0–4) in the form of soluble complexes (Cr³⁺, [Cr (H₂O)₆³⁺]), while near neutral (pH 6–9) inert precipitates (Cr (OH₃) (s)) can easily be adsorbed on solid media [18].

HM ions are present in different concentrations (trace and shock concentrations) and forms [19,20], in combined industrial and municipal wastewater streams. Disposal of treated wastewater into the environment necessitates adjusting its pH to neutral. Especially for recycling of HM ions and water reuse [19,21], this means raising costs and increases the complexity of the pretreatment process, as well as increasing the addition of excessive amounts of chemicals [19,21].

Therefore, it is necessary to act sustainably and environmentally consciously by removing HMs from the highly polluted wastewaters using technology/methods that allow the removal/recycling of HM ions [4].

Currently, different conventional methods are used to remove HM ions from water/industrial wastewater [5,11], such as precipitation [22], electrochemical removal [23], ion exchange [24], membrane filtration [25,26], coagulation [27], flocculation [28], and sorption on natural materials [29]. Although these methods are efficient in removing HM ions, they do not allow the recycling and reuse of HM ions. Some already well-established methods for the removal of HM ions produce toxic coproducts and large amounts of waste sludge [30], e.g., membrane filtration and coagulation/flocculation [31]. Moreover, these methods are often costly and energy-inefficient [30].

On the other hand, the adsorption method is well known, efficient, and used for the removal of HM ions due to its low adsorbent and operational costs and simple principle [5,11,12]. Adsorption can be performed with various natural materials [29] and other hybrid materials based on silica and iron oxide NPs (γ-Fe₂O₃, Fe₃O₄), as well as their functionalized forms [32,33,34,35,36]. The most commonly used adsorbent of HM ions from industrial and leachate wastewaters is activated carbon [31]. Activated carbon is efficient in the removal of HM ions from wastewaters, due to its high specific surface area, micropore volume, and pore volume [31,33,37,38,39,40,41]. At the same time, limitations of its use include non-selectivity and high material price. Furthermore, activated carbon does not enable the recycling and regeneration of HM ions and the adsorbent itself, and for now it does not enable the circular economy approach [31,33].

Due to the increased need for recycling of municipal and industrial wastewaters [42] research in nanotechnology is investing in the preparation and testing of functionalized (nano)materials that can improve the recycling of specific HM ions [34,36,43,44].

Maghemite (γ-Fe₂O₃) is a member of the family of iron oxides. It has a cubic spinel ferrite structure, and it is ferrimagnetic. When reduced to particle dimensions smaller than a certain domain—i.e., becoming a single domain—ferrimagnetic materials exhibit superparamagnetic behavior, which means that when an external magnetic field is applied, they magnetize, but when the magnetic field is removed, they no longer exhibit either residual magnetism or coercivity.

Such superparamagnetic nanoparticles, if they are surface-functionalized, provide promising applications in the adsorption of heavy metals from aqueous media, as they enable more efficient separation and recovery of heavy metals from the contaminated aqueous medium using an external magnetic field [45]. On the other hand, surface modification of superparamagnetic γ-Fe₂O₃ nanoparticles with TEOS and APTMS precursors improves their stability, prevents them from agglomerating, and increases their surface functionality by increasing the number of adsorption sites (-NH₂), facilitating and accelerating diffusion pathways for heavy metal pollutants [46,47]. Despite all of the advantages of superparamagnetic γ-Fe₂O₃ 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 γFe₂O₃ 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’_xSi(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, (CH₃O)₃-Si-(CH₂)₃-NH₂)) molecules on the surface of the γFe₂O₃ particles would polymerize into a highly homogeneous crosslinked SiO₂ coating with functional amino (-NH₂) groups present. However, the presence of a non-hydrolyzable fraction in the AMPTS structure ((CH₃O)₃-Si-(CH₂)₃-NH₂)) 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 (-NH₂) groups with an insufficient surface coverage of the γFe₂O₃ nanoparticles [50,51,52].

In contrast to APTMS, under base-catalyzed conditions, the reactivity of tetraethoxysilane (TEOS, Si(OCH₂CH₃)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 ((CH₃O)₃-Si-(CH₂)₃-NH₂)) was used as a supplier of the -NH₂ functional groups.

In this way, it was possible to create uniform spherical γFe₂O₃@SiO₂-NH₂ core–shell structures with the presence of amino (-NH₂) 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 H₂CrO₄⁰, HCrO₄⁻, CrO₄²⁻, and Cr₂O₇²⁻ [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 Pb²⁺ 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, Hg₂²⁺, composed of two singly charged ions; and a doubly charged Hg²⁺. 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 Hg₂(OH)₂²⁺ and Hg(OH)^3– at 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, CrO₄²⁻, 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 Fe₃O₄@SiO₂-NH₂ magnetic nanoparticles [54,61].

Furthermore, it is generally known that iron oxides (γFe₂O₃, Fe₃O₄, 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 (γFe₂O₃@SiO₂-NH₂) [61].

Iron oxide nanoparticles—i.e., goethite (α-FeOOH), hematite (α-Fe₂O₃), magnetite (Fe₃O₄), maghemite (γ-Fe₂O₃) [66,67,68,69,70], —show moderate affinity towards HM ions on their surface. They appear more applicable if the surface is stabilized [71,72,73], and enlarged by coating with silica NPs (-SiO₂), whereby agglomeration is prevented [74]. Additionally, by using different functional groups [71,72,73]—e.g., amino (-NH₂) [34,66,67,68,75,76], mercapto (-SH) [77,78], carboxy (-COOH) [79]—the adsorption efficiency and adsorption capacity of HM ions can be improved [33,34,71].

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

Table 1. Comparison of adsorption capacity and desorption efficiency for tested MNPs and amino-functionalized MNPs at the optimal model solution pH for adsorbing Pb²⁺ ions.

Adsorbent	HM Ions	Tested pH	Adsorption Capacity	Desorption Efficiency	Reference
γ-Fe2O3 NPs	Pb2+	7.5	10.55 mg/g	-	[35]
Fe3O4@SiO2 NPs		6.0	14.9 mg/g	95.7%	[82]
NH2-functionalized Fe2O3/chitosan NPs		5.0	32.46 mg/g	-	[36]
NH2-functionalized Fe2O3 NPs		5.0	39.30 mg/g	-	[36]
Magnetic composite of activated carbon and superparamagnetic Fe3O4 NPs (Fe3O4@C magnetic composite)		6.0	41.7 mg/g	>77%	[83]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		4.0	53.5 mg/g	90.7%	This work
Amino-functionalized graphene oxide (GO-NH2)		5.0	53.9 mg/g	-	[4]
Fe3O4@SiO2–NH2 NPs		6.2	0.37 mmol/g 76.66 mg/g *	-	[43]
Amino-functionalized Fe3O4@mesoporous SiO2 core-shell composite microspheres		5.5	82.29 mg/g	-	[35]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		7.0	83.6 mg/g	91.3%	This work
Polyethylenimine (PEI)-functionalized Fe3O4 magnetic nanoparticles (MNPs)		pH 5.0	60.98 mg/g		[84]
Composite beads of Zea mays rachis (ZMR) and sodium alginate (AL)		pH 5.0	60 mg/g		[85]
Carbon-doped TiO2 (C-TiO2)		pH 6.5	28.7 mg/g		[86]

* calculated.

,

Table 2. Comparison of adsorption capacity and desorption efficiency for tested MNPs and amino-functionalized MNPs at optimal model solution pH for adsorbing CrT/Cr³⁺/Cr⁶⁺/Cr(VI) ions.

Adsorbent	HM Ions	Tested pH	Adsorption Capacity	Desorption Efficiency	Reference
Magnetic magnetite NPs (Fe3O4)	CrT/Cr3+/Cr6+/Cr(VI)	4.0	8.67 mg/g	>75%	[87]
Iron oxide magnetic NPs (MNPs)		2.5	15.0 mg/g	≅100%	[72]
Maghemite NPs (γ-Fe2O3)		2.5	19.2 mg/g	87.7%	[88]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		4.0	24.0 mg/g	-	This work
Amino-functionalized magnetite NPs (NH2-Fe3O4)		3.0	24.25 mg/g	98.02%	[89]
APTES@TEOS@MNP		2.5	35.0 mg/g	≅100%	[72]
NH2-functionalized nanomagnetic polymer adsorbents (EDA-NMPs)		2.5	37.6 mg/g	-	[44]
NH2-functionalized nanomagnetic polymer adsorbents (DETA-NMPs)		2.5	37.9 mg/g	-	[44]
NH2-functionalized nanomagnetic polymer adsorbents (TETA-NMPs)		2.5	38.5 mg/g	-	[44]
NH2-functionalized nanomagnetic polymer adsorbents (TEPA-NMPs)		2.0	40.0 mg/g	-	[44]
Amino-functionalized graphene oxide (GO-NH2)		2.0	90.4 mg/g	-	[4]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		7.0	90.4 mg/g	96.7%	This work
Carbon-encapsulated hematite nanocubes (αFe2O3@C)		pH 3	76.92 mg/g		[90]
Activated carbons		pH 2	4.35 mg/g		[91]

and Table 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., Pb²⁺, CrT/Cr³⁺/Cr⁶⁺, and Hg²⁺).

Table 1 shows comparison of the adsorption capacities and desorption efficiency for Pb²⁺ ions by non-functionalized and functionalized MNPs. It can be seen that the adsorption of Pb²⁺ 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-Fe₂O₃ NPs via the wet chemical method and tested adsorption at pH 7.5, while Nicola et al. (2020) [82] synthesized Fe₃O₄@SiO₂ 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 SiO₂-stabilized magnetic nanomaterials [82]. Nicola et al. (2020) [82] also tested the desorption efficiency of Pb²⁺ ions with 5% HCl, and the final desorption efficiency was evaluated as 95.7% [82]. Qian et al. (2019) [36] stated that the adsorption capacity of materials functionalized with chitosan and with an amino group (NH₂-functionalized Fe₂O₃/chitosan NPs) at pH 5.0 was not significantly better compared to NH₂-functionalized Fe₂O₃ materials. Higher adsorption capacity with Fe₃O₄ NPs coated with activated carbon was achieved at pH 6.0 [83], while Huang et al. (2020) [5] reached 53.9 mg/g with amino-functionalized graphene oxide at pH 5.0. Wang et al. (2010) [43] and Tang et al. (2013) [34] stated that even higher adsorption capacity at pH 6.2 (76.66 mg/g) can be achieved by pre-stabilization with SiO₂ and amino-functionalization of magnetic materials (Fe₃O₄@SiO₂–NH₂ NPs). Tang et al. (2013) [34] achieved 82.29 mg/g with amino-functionalized Fe₃O₄@mesoporous SiO₂ core–shell composite microspheres at pH 5.5. In polyethylenimine (PEI)-functionalized Fe₃O₄ magnetic nanoparticles (MNPs) (pH 5.0), adsorption capacity of 60.98 mg/g for Pb²⁺ ions was reported [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 TiO₂ (C-TiO₂) at pH 6.5 for the adsorption of Pb²⁺. The comparison of adsorption capacities showed that the adsorption capacity of Pb²⁺ ions depends on the pH of the medium, stabilization, and, to a large extent, the presence of -NH₂ groups.

From the literature, it can be observed that higher adsorption of Cr³⁺/Cr⁶⁺/CrT ions was achieved using amino (-NH₂)-functionalized MNPs (Table 2). The highest adsorption of Cr³⁺ ions by bare MNPs was reported in the literature [72,87,88]. Zhang et al. (2020) [11] used bare magnetic magnetite NPs (Fe₃O₄) for the adsorption of Cr³⁺ ions at pH 4.0 and achieved an adsorption capacity of 8.67 mg/g. Additionally, in acidic media (pH 2.5), Gallo-Cordova et al. (2019) [72] performed adsorption of Cr³⁺ ions using bare iron oxide magnetic NPs and reported an adsorption capacity of 15.0 mg/g, while Hu et al. (2005) [17] achieved a very low adsorption capacity of 19.2 mg/g using maghemite NPs (γ-Fe₂O₃). Zhang et al. (2020) [11], Gallo-Cordova et al. (2019) [72], and Hu et al. (2005) [17] also performed desorption with NaOH, achieving desorption efficiency of >75%, ≅100%, and 87.7%, respectively. Other studies used amino-functionalized MNPs and achieved higher adsorption capacity in acidic media (pH 2.0 to 3.0). Adsorption of Cr³⁺ ions at pH 3.0 was performed using amino-functionalized magnetite NPs (NH₂-Fe₃O₄) [89]. Baghani et al. (2016) [89] achieved an adsorption capacity of 24.25 mg/g and desorption efficiency of 98.02%. Even better adsorption (i.e., 35.0 mg/g) was reported by Gallo-Cordova et al. (2019) [72] using APTES@TEOS@MNP at pH 2.5, and the desorption efficiency was also high (≅100%). Zhao et al. (2010) [44] prepared NH₂-functionalized nanomagnetic polymer adsorbents. Functionalization was performed with different precursors (i.e., EDA-, DETA-, TETA-, TEPA-). A maximum adsorption capacity of 38.5 mg/g at pH 2.5 was achieved using TETA-functionalized nanomagnetic polymer adsorbents. In another study, 40.0 mg/g of Cr³⁺ ions was adsorbed at pH 2.0 by TEPA-functionalized nanomagnetic polymer adsorbents [64]. Huang et al. (2020) [5] reported the adsorption properties of functionalized non-magnetic materials at pH 2.0. Using amino-functionalized graphene oxide (GO-NH₂), Huang et al. (2020) [5] achieved 90.4 mg/g, which is the same adsorption capacity that we achieved with NH₂-functionalized γ-Fe₂O₃ NPs (γ-Fe₂O₃@NH₂ NPs), but at alkaline pH (7.0). The maximum adsorption capacity for chromium(VI) ions was 76.92 mg/g at pH 3.0 when adsorbent carbon-encapsulated hematite nanocubes (αFe₂O₃@C) were used [90]. Puszkarewicz and Kaleta (2019) [91] used activated carbon as an adsorbent, and the maximum adsorption capacity for chromium(VI) ions was 4.35 mg/g at pH 2 [91].

The maximum adsorption capacity for Hg²⁺ (Table 3) was 32.88 mg/g at pH 5.0 using carboxyl-terminated hyperbranched poly(amidoamine) dendrimers grafted onto superparamagnetic NPs (CT-HPMNPs) as adsorbents, and the maximum desorption efficiency was ≅85% (using HNO₃ acid) [92]. Wang et al. (2013) [93] used rhodamine-hydrazide-modified Fe₃O₄ as an adsorbent, and the maximum adsorption capacity for Hg²⁺ was 37.4 mg/g at pH 7.5 [93]. Bolivar et al. (2018) [80] performed a study of Hg²⁺ ion adsorption, in which Fe₃O₄ nanoparticles coated with amino organic ligands and yam peel biomass displayed a maximum Hg²⁺ adsorption capacity of 60 mg/g at pH 7.0 [94]. The maximum adsorption capacity for Hg²⁺ was 50 mg/g at pH 7.0 when an adsorbent nanocomposite based on Fe₃O₄ nanoparticles, chitosan nanoparticles, and polythiophene was used [95]. Dun Chen et al. (2016) [96] studied the adsorption of Hg²⁺ using magnetic adsorbents (Fe₃O₄@SiO₂-NH-HCGs; HCG = py (2-pyridinyl); pyd (3-pyridazinyl)) formed by grafting of different heterocyclic groups onto amino groups via substitution reaction. The maximum adsorption capacity for Fe₃O₄@SiO₂-NH-HCG- (pyd) and Fe₃O₄@SiO₂-NH-HCG- (py) was 77 mg/g and 56 mg/g at pH 7.0 [96], respectively. For both adsorption materials, HCl was used as the desorption eluent, and the stated desorption efficiency was 95% [96]. Hao et al. (2021) [97] performed a study of Hg²⁺ ion adsorption with Armeniaca sibirica shell activated carbon (ASSAC) magnetized by nanoparticles (Fe₃O₄/ASSAC), showing a maximum adsorption capacity of 97.1 mg/g at pH 2. At pH 5.5, Zhang et al. (2016) [98] studied the adsorption of Hg²⁺ ions with activated carbon (XLAC) derived from Xanthoceras sorbifolia Bunge hull as an adsorbent, showing a maximum adsorption capacity of 235.6 mg·g⁻¹. A maximum adsorption capacity of 162 mg g⁻¹ for Hg²⁺ ions at pH 5.0 was achieved using cadmium sulfide nanoparticles doped in a nanoadsorbent fabricated from polycaprolactam (nylon 6) nanofibers (CdS/N6) [99].

Table 3. Comparison of adsorption capacity and desorption efficiency for tested MNPs and amino-functionalized MNPs at optimal model solution pH for adsorbing Hg²⁺ ions.

Adsorbent	HM Ions	Tested pH	Adsorption Capacity	Desorption Efficiency	Reference
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)	Hg2+	4.0	16.2 mg/g	100%	This work
CT-HPMNPs		5.0	32.88 mg/g	≅85%	[92]
Rhodamine-hydrazide-modified Fe3O4		7.5	37.4 mg/g	-	[93]
Nanocomposite based on Fe3O4 nanoparticles, chitosan nanoparticles, and polythiophene		7.0	50 mg/g	-	[95]
Fe3O4@SiO2-NH-HCG- (py)		7.0	56 mg/g	95%	[96]
Fe3O4 nanoparticle coated with amino organic ligands and yam peel biomass		7.0	60 mg/g	-	[94]
Fe3O4@SiO2-NH-HCG- (pyd)		7.0	77 mg/g	95%	[96]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		7.0	85.6 mg/g	100%	This work
Armeniaca sibirica shell activated carbon (ASSAC) magnetized by nanoparticles (Fe3O4/ASSAC)		pH 2	97.1 mg/g		[97]
Activated carbon (XLAC) derived from Xanthoceras sorbifolia Bunge hull		pH 5.5	235.6 mg·g		[98]
Cadmium sulfide nanoparticles doped in a nanoadsorbent fabricated from polycaprolactam (nylon 6) nanofibers (CdS/N6)		pH 5	162 mg g		[99]

Table 3. Comparison of adsorption capacity and desorption efficiency for tested MNPs and amino-functionalized MNPs at optimal model solution pH for adsorbing Hg²⁺ ions.

Adsorbent	HM Ions	Tested pH	Adsorption Capacity	Desorption Efficiency	Reference
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)	Hg2+	4.0	16.2 mg/g	100%	This work
CT-HPMNPs		5.0	32.88 mg/g	≅85%	[92]
Rhodamine-hydrazide-modified Fe3O4		7.5	37.4 mg/g	-	[93]
Nanocomposite based on Fe3O4 nanoparticles, chitosan nanoparticles, and polythiophene		7.0	50 mg/g	-	[95]
Fe3O4@SiO2-NH-HCG- (py)		7.0	56 mg/g	95%	[96]
Fe3O4 nanoparticle coated with amino organic ligands and yam peel biomass		7.0	60 mg/g	-	[94]
Fe3O4@SiO2-NH-HCG- (pyd)		7.0	77 mg/g	95%	[96]
NH2-functionalized γ-Fe2O3 NPs (γ-Fe2O3@NH2 NPs)		7.0	85.6 mg/g	100%	This work
Armeniaca sibirica shell activated carbon (ASSAC) magnetized by nanoparticles (Fe3O4/ASSAC)		pH 2	97.1 mg/g		[97]
Activated carbon (XLAC) derived from Xanthoceras sorbifolia Bunge hull		pH 5.5	235.6 mg·g		[98]
Cadmium sulfide nanoparticles doped in a nanoadsorbent fabricated from polycaprolactam (nylon 6) nanofibers (CdS/N6)		pH 5	162 mg g		[99]

Adsorption has predominantly been investigated using -NH₂ [36], -SiO₂ [82], and -SH [94]-functionalized Fe₃O₄ or γ-Fe₂O₃ NPs [67] prepared by different approaches, in model water media of various pH values [34,66], from pH 2.0 to 8.0.

There are not many previous studies [32,66] on testing adsorption by γ-Fe₂O₃ NPs functionalized with an amino (-NH₂) 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-Fe₂O₃ MNPs (γ-Fe₂O₃@NH₂ 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 Pb²⁺, CrT, and Hg²⁺ ions using different amounts (m_ads = 20/45/90 mg) of the γ-Fe₂O₃@NH₂ 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, -NH₂-functionalized γ-Fe₂O₃ MNPs were characterized with different methods, such as FTIR, BET, TEM, and TGA. Zeta potential changes in γ-Fe₂O₃@NH₂ NPs were analyzed to understand the mechanisms taking place during the adsorption and desorption process of Pb²⁺ ions. Moreover, to evaluate the adsorbent regeneration, desorption with 0.1 M HNO₃ was tested, which is of great importance for the reuse of adsorption materials and recycling of heavy metals. The prepared γ-Fe₂O₃ and functionalized γ-Fe₂O₃@NH₂ MNPs were also characterized by X-ray powder diffractometry (XRD).

2. Results and Discussion

2.1. Properties of the Prepared γ-Fe₂O₃@NH₂ NPs

This section explains the characterization of the synthesized, stabilized, and functionalized MNPs (γ-Fe₂O₃@NH₂ NPs). In addition, the adsorption mechanisms and the results of batch adsorption and desorption experiments are also discussed.

2.1.1. Crystallographic Properties

The prepared γ-Fe₂O₃ and functionalized γ-Fe₂O₃@NH₂ MNPs were characterized by X-ray powder diffractometry (XRD) (Figure 1). In the X-ray powder diffraction pattern in Figure 1, the presence of diffraction peaks at 2θ of 30.2°, 35.5°, 43.2°, 53.6°, 57.2°, and 62.9°—which correspond to the cubic crystal planes of (220), (311), (400), (422), (511), and (440), respectively—are characteristic of the spinel crystal structure (JPCD Card 39-1346). The spinel crystal structure is evident for both samples—γ-Fe₂O₃ and functionalized γ-Fe₂O₃@NH₂ MNPs—while the presence of a broad amorphous diffraction peak for the functionalized γ-Fe₂O₃@NH₂ MNPs, which appears at a low diffraction angle 2θ of 20°, is due to the presence of the amorphous SiO₂ surface layer, indicating that the crystalline cubic spinel γ-Fe₂O₃ magnetic cores were surface-modified [78]. The average size of the γ-Fe₂O₃ crystalline magnetic cores was estimated to be 13 nm, using the Debye–Scherrer equation [100,101].

2.1.2. 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].

2.1.3. 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 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].

2.1.4. 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 m²/g for γ-Fe₂O₃ and 41.3 m²/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 cm³/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 cm³/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, indicating a decrease in the porosity of the as-prepared γ-Fe₂O₃ 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-Fe₂O₃@NH₂ NPs, with a total pore volume of 0.090762 cm³/g, while for the BJH desorption the average pore size for the γ-Fe₂O₃@NH₂ NPs increased to 6.0 nm, with a total pore volume of 0.090311 cm³/g.

According to the specific surface area (BET) at a relative pressure (p/p⁰) of 0.3, the calculated average particle size was 11.6 nm for γ-Fe₂O₃ and 27.9 nm for γ-Fe₂O₃@NH₂ NPs [109,110]

2.1.5. 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.

2.1.6. Zeta Potential

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.

2.2. 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 Pb²⁺, CrT, and Hg²⁺ 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 Pb²⁺ 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]:

$$-\mathrm{N}{\mathrm{H}}_2+{\mathrm{H}}_3{\mathrm{O}}^{+}\to -\mathrm{N}{\mathrm{H}}_3^{+}+{\mathrm{H}}_2\mathrm{O}$$

(1)

$$-\mathrm{N}{\mathrm{H}}_2+\mathrm{O}{\mathrm{H}}^{-} \to -\mathrm{N}{\mathrm{H}}^{-}+{\mathrm{H}}_2\mathrm{O}$$

(2)

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 attraction between Pb²⁺ ions and the negatively charged material surface and, consequently, high adsorption efficiency [82,111]. On the other hand, acidic conditions cause the transformation of -NH₂ groups into -NH₃⁺ form, resulting in fewer available active sites for Pb²⁺ ions. Because of that, the adsorption efficiency of Pb²⁺ ions drops under acidic conditions (pH < 7) [5,111].

The solution pH value is also a key factor in the adsorption efficiency of CrT ions. Adsorption efficiency generally decreases with increasing pH values [5]. In acidic conditions, CrT ions are mainly present as H₂CrO₄⁰, HCrO₄⁻, and Cr₂O₇⁻ species [68,72,81,113], while -NH₂ is present in protonated form, i.e., -NH₃⁺ form (Equation (3)). Consequently, the γ-Fe₂O₃@NH₂ NPs’ surfaces are positively charged [5,81,114]. Strong electrostatic attraction occurs in such cases, and chromium species can be easily captured on the amino-functionalized adsorbent surface [72,81].

$$R-\mathrm{N}{\mathrm{H}}_2+{\mathrm{H}}^{+}\to R-\mathrm{N}{\mathrm{H}}_3^{+}$$

(3)

$$-\mathrm{N}{\mathrm{H}}_3^{+}+\mathrm{H}\mathrm{C}\mathrm{r}{\mathrm{O}}_4^{-}\to -\mathrm{N}{\mathrm{H}}_3^{+}-\mathrm{H}\mathrm{C}\mathrm{r}{\mathrm{O}}_4^{-}$$

(4)

$$-\mathrm{N}{\mathrm{H}}_3^{+}+\mathrm{C}{\mathrm{r}}_2{\mathrm{O}}_7^{2-}\to -\mathrm{N}{\mathrm{H}}_3^{+}-\mathrm{C}{\mathrm{r}}_2{\mathrm{O}}_7^{2-}$$

(5)

On the other hand, at alkaline pH, negatively charged chromate ions (CrO₄²⁻) are the predominant form [81,87,113]. At pH > 7, γ-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ NPs was 8.76; at lower pH, amino groups on the material’s surface were mainly present in protonated form (-NH₃⁺).

We tested the adsorption of CrT ions at pH 4 and 7. At pH 4, the functional groups were mainly in -NH₃⁺ 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-Fe₂O₃@NH₂ NPs, so their adsorption capacity was very high. The adsorption efficiency at pH 4 was low due to the instability of γ-Fe₂O₃@NH₂ NPs in acidic conditions—the adsorbent material is soluble in acidic media, i.e., at pH < 4.

Hg²⁺ readily reacts with OH⁻ to form Hg₂(OH)₂ precipitates under alkaline conditions. The adsorption of Hg²⁺ 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 Hg²⁺ ions favors neutral and basic pH. The rationale for more adsorption of Hg²⁺ ions using amino groups at neutral and basic pH is that the amino group obtains a net positive charge at acidic pH and the Hg²⁺ 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].

2.3. Effects of pH

Batch adsorption experiments of Pb²⁺, CrT, and Hg²⁺ ions for different adsorption times were performed at two pH values, i.e., pH 4 and 7 (Figure 8, Figure 9, Figure 10 and Figure 11). The results show that the adsorption of Pb²⁺ and Hg²⁺ ions is more efficient at pH 7. Such results were expected, due to the opposite charges of the Pb²⁺ ions and the surface of the adsorption material. The opposite surface charges caused strong electrostatic interactions and high material uptake.

The adsorption capacity at pH 4 and 7 slowly increased with longer specific adsorption times. At pH 4, the maximal adsorption capacity of Pb²⁺ 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 Pb²⁺ ions slowly increased with a longer adsorption time. This indicates that the adsorption of Pb²⁺ 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-Fe₂O₃@NH₂ 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 Hg²⁺ 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, Hg²⁺ 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.

2.4. Effect of Adsorbent Mass

The adsorption of CrT ions on 45 mg of γ-Fe₂O₃@NH₂ 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-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 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-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ NPs, we concluded that 45 mg was the optimal mass of adsorbent.

2.5. Effects of Anions (NO₃⁻, Cl⁻, SO₄²⁻)

The adsorption of CrT ions showed high adsorption efficiency (99.9%) after 1 min. For these reasons, we also tested the impacts of different anions (i.e., NO₃⁻, Cl⁻, and SO₄²⁻) on the adsorption efficiency of CrT ions. Experiments were performed under further optimal determined conditions (pH = 7, m_ads = 45 mg, c = 200 mg CrT/L), and only for 1, 5, 10, 15, 30, 45, and 60 min at RT.

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).

2.6. Desorption

To verify the possibility of recycling HM ions and reusing adsorption materials, desorption of Pb²⁺, CrT, and Hg²⁺ 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-Fe₂O₃@NH₂ 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 HNO₃ was more efficient for Hg²⁺, CrT, and Pb²⁺ ions. For Pb²⁺ ions, the desorption efficiency was 91.3%; for CrT ions it was 96.7%; and for Hg²⁺ 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 Hg²⁺ 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.

3. Methods and Materials

3.1. Materials

For the lab-scale synthesis, stabilization, and functionalization of γ-Fe₂O₃ MNPs, iron(II) sulfate heptahydrate (FeSO₄ 7H₂O, 99.5%, 278.01 g/mol, CAS 7782-63-0, Honeywell International Inc., Charlotte, NC, USA), iron(III) sulfate hydrate (Fe₂(SO₄)₃ xH₂O, 97%, 399.88 g/mol, CAS 15244-10-7, Honeywell International Inc., Charlotte, NC, USA), tetraethyl orthosilicate TEOS (C₆H₂₀O₄Si, 99%, 208.33 g/mol, CAS 78-10-4, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), (3-aminopropyl)trimethoxysilane APTMS (C₆H₁₇NO₃Si, 97%, 179.29 g/mol, CAS 13822-56-5, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), ammonium hydroxide (NH₄OH, 25%, 35.05 g/mol, CAS 1336-21-6, GramMol, Zagreb, Croatia), ethanol (C₂H₅OH, 96%, 46.07 g/mol, CAS 64-17-5, GramMol, Zagreb, Croatia), and 2-propanol (C₃H₇OH, 99.9%, 60.10 g/mol, CAS 67-63-0, GramMol, Zagreb, Croatia) were used. For the batch adsorption experiments, aqueous solutions were prepared from lead(II) nitrate (Pb(NO₃)₂, 99.9%, 331.21 g/mol, CAS 10099-74-8, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), chromium(III) nitrate nonahydrate (Cr(NO₃)₃ 9H₂O, 99%, 400.15 g/mol, CAS 7789-02-8, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), mercury(II) nitrate monohydrate (Hg(NO₃)₂ H₂O, ≥99.99%, 342.62 g/mol, CAS 7783-34-8, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), chromium(III) chloride hexahydrate (CrCl₃ 6H₂O, ≥98%, 266.45 g/mol, CAS 10060-12-5, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany), and chromium(III) sulfate hydrate (Cr₂(SO₄)₃ xH₂O, 99%, 392.17 g/mol, CAS 15244-38-9, Sigma-Aldrich, Merck Group KGaA, Darmstadt, Germany). Chemicals were used as purchased. For the preparation of suspensions and solutions, deionized water (dH₂O) was used.

3.2. Synthesis of MNPs

The synthesis of the maghemite (γ-Fe₂O₃) MNPs was carried out as described in our previous studies [32]. First, 30 mL of NH₄OH 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 FeSO₄ 7H₂O and Fe₂(SO₄)₃ xH₂O 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 dH₂O, centrifuged at 3500 rpm for 15 min and, finally, separated and allowed to settle on the external magnet overnight.

3.3. Stabilization of MNPs

The rinsed γ-Fe₂O₃ MNPs were suspended in 100 mL of NH₄OH 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.

3.4. Functionalization of MNPs

Amino-functionalization of the γ-Fe₂O₃ MNPs was carried out with tetraethyl orthosilicate (TEOS) and (3-aminopropyl)trimethoxysilane (APTMS) precursors, with a water-to-TEOS molar ratio R = 293 and TEOS-to-APTMS molar ratio P = 1:2. After the stabilization phase, 4 mL of γ-Fe₂O₃ colloidal solution was added to the mixture of 21.6 mol% 2-propanol, 2.2 mol% NH₄OH, 15.1 mol% dH₂O, 0.25 mol% TEOS, and 0.36 mol% APTMS, and then mixed vigorously for 5 min. After 24 h of reaction (at RT under constant stirring at 220 rpm), the colloidal solution was intensively washed with ethanol and dH₂O and centrifuged at 3500 rpm for 5 min to remove agglomerates of non-functionalized γ-Fe₂O₃ MNPs.

3.5. Characterization of Amino-Functionalized γ-Fe₂O₃ MNPs

Characterization of the lab-scale amino-functionalized γ-Fe₂O₃ MNPs (γ-Fe₂O₃@NH₂ NPs) was performed using the appropriate method after each preparation phase (i.e., synthesis, stabilization, functionalization). For characterization purposes, the γ-Fe₂O₃@NH₂ MNPs were dried at 90 °C overnight, and the mass of the obtained dried particles was determined. The prepared γ-Fe₂O₃ and functionalized γ-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ 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.

3.6. Adsorption of Heavy Metal Ions in Aqueous Solutions

Batch adsorption tests of Pb²⁺, CrT, and Hg²⁺ ions were performed. The initial concentration of 200 mg/L of HM ions in the model water solutions was prepared by dissolving Pb (NO₃)₂, Cr (NO₃)₃ 9H₂O, and Hg (NO₃)₂ H₂O 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., NO₃⁻, Cl⁻, and SO₄²⁻) on the adsorption of CrT ions was investigated. For Pb²⁺ and Hg²⁺ 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 HNO₃ to simulate an acidic environment, which does not affect the γ-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ NPs were weighed. Then, 20 mL of the prepared model salt solution was added to the γ-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ 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-Fe₂O₃@NH₂ NPs were washed two times with 10 mL of dH₂O.

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 VISION, SPECTRO Analytical Instruments GmbH, Kleve, Germany) for Hg²⁺. For both analyses, the supernatants were acidified with HNO₃ (0.5 mL of acid to 10 mL of the supernatant sample).

The adsorption efficiency (R %) was calculated using Equation (6) [87]:

$$\mathrm{R}=\frac{\left({\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{t}}\right)}{{\mathrm{C}}_0}\times 100$$

(6)

where R (%) is the adsorption efficiency, C₀ (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]:

$${\mathrm{q}}_{\mathrm{t}}=\left({\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{t}}\right)\times \frac{\mathrm{V}}{\mathrm{m}}$$

(7)

where q_t (mg/g) is the adsorption capacity, C₀ (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.

3.7. Desorption of HM Ions and Regeneration Experiments

The desorption experiments for Pb²⁺, CrT, and Hg²⁺ ions were conducted to evaluate the recyclability of γ-Fe₂O₃@NH₂ 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 HNO₃ 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 Pb²⁺ and CrT, and with ICP-OES for Hg²⁺ ions.

4. Conclusions

In this study, stabilized and amino-functionalized magnetic nanoparticles (γ-Fe₂O₃@NH₂ NPs) with a diameter of 17 ± 1 nm were synthesized, characterized, and used as adsorbents for Pb²⁺, CrT, and Hg²⁺ ions. Adsorbent characterization showed that γ-Fe₂O₃@NH₂ 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., Pb²⁺, CrT, and Hg²⁺. The adsorption efficiency was the highest and quickest for CrT > Hg²⁺ > Pb²⁺ ions. The maximal adsorption capacity for Hg²⁺ ions was achieved in 30 min, at 85.6 mg/g; for CrT and Pb²⁺ 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-Fe₂O₃@NH₂ 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., NO₃⁻, Cl⁻, and SO₄²⁻—showed no effect on the adsorption efficiency of CrT ions.

A study of the desorption process with 0.1 M HNO₃ for 1 h showed the possibility of reusing γ-Fe₂O₃@NH₂ NPs. Desorption was effective only for γ-Fe₂O₃@NH₂ NPs after the adsorption process was performed at neutral pH. We observed excellent desorption efficiency for Hg²⁺ (100%), CrT (96.7%), and Pb²⁺ (91.3%) ions.

The adsorption–desorption results showed that γ-Fe₂O₃@NH₂ 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.