826 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... DOI: 10.17344/acsi.2022.7570 Scientific paper Assessment of the Capability of Magnetic Nanoparticles to Recover Neodymium Ions from Aqueous Solution Ana Ambrož,1 Irena Ban2 and Thomas Luxbacher1,3,* 1 University of Maribor, Faculty of Chemistry and Chemical Engineering, Laboratory for Water Biophysics and Membrane Processes, Smetanova 17, 2000 Maribor, Slovenia 2 University of Maribor, Faculty of Chemistry and Chemical Engineering, Laboratory for Inorganic Chemistry, Smetanova 17, 2000 Maribor, Slovenia 3 Anton Paar GmbH, Anton-Paar-Strasse 20, 8054 Graz, Austria * Corresponding author: E-mail: thomas.luxbacher@anton-paar.com Received: 10-03-2022 Abstract Magnetic nanoparticles (MNPs) have received increasing attention for various applications due to their fast synthesis, versatile functionalization, and recyclability by the application of a magnetic field. The high surface-to-volume ratio of MNP dispersions has suggested their use as an adsorbent for the removal of heavy metal ions. We investigated the ap- plicability of MNPs composed of a maghemite core surrounded by a silica shell functionalized with aminopropylsilane, γ-Fe2O3-NH4OH@SiO2(APTMS), for the removal of neodymium ions (Nd3+) from aqueous solution. The MNPs were characterized for their size, composition, surface functionality and charge. Despite of the promising properties of MNPs, their removal from the aqueous dispersion with an external magnet was not sufficient to reliably quantify the adsorption of Nd3+ by UV-Vis spectroscopy. Keywords: Rare earth elements; Maghemite; Nanoparticles; UV-Vis spectroscopy, Adsorption 1. Introduction The rare earth elements (REEs) are a group of 17 strongly related heavy elements that comprise scandium (Sc), yttrium (Y), lanthanum (La) and the f-block elements known as the lanthanide group, cerium (Ce) through lute- tium (Lu). In addition to being of great value to general geochemistry investigations, the REEs also have high com- mercial value and a wide scope of applications. They are used worldwide in various electronic and optical products, in advanced technologies, medical devices, military de- fence systems, as well as in the field of clean energy.1,2 RE- Es have been on the list of critical raw materials since 2010, and pressure on already limited resources is still increasing with the growth of the global population, industrializa- tion, and digitalization. With regard to high economic im- portance and high supply risk, neodymium (Nd), europi- um (Eu), terbium (Tb), dysprosium (Dy), and yttrium (Y) are considered most critical.3 With increasing demand and production, the amount of electronic waste containing REEs in various concentrations is also expanding. Sustain- ability in REE supply and proper treatment of end-of-life electrical and electronic compounds are crucial to achiev- ing climate neutrality.3 However, recycling of REEs pre- sents many challenges. Firstly, REEs are usually present in small amounts in tiny electronic parts of gadgets like mo- bile phones. In some materials like touch screens, these metals are evenly distributed making them much more difficult to extract.4 Secondly, due to the low yield and high cost of recycling processes, REEs are not recycled in large quantities, regardless of the end use. However, if REEs’ prices rise, recycling may become feasible.4,5 At present, the main focus is on the direct recycling of scrap and the urban mining, and subsequent recycling of end-of-life REE-containing products.6 The conventional processes for the separation and recovery of REEs mainly include precipitation, ion ex- change, coagulation, flocculation, liquid–liquid extrac- tion, biosorption, and adsorption. Among these methods, adsorption offers an efficient, environmentally friendly, and economical procedure for the removal of rare earth ions.7,8 Among different adsorbents, magnetic nanoparti- 827Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... cles show a significant potential due to their high surface area and their response to an external magnetic field, which eases separation from a supernatant solution after completing the adsorption process.9–11 The chemical and physical stability, biocompatibility, ease of surface modifi- cation, low toxicity, straightforward synthesis, and low cost of iron oxide nanoparticles (IONPs) make them ideal for a variety of applications.12–15 Preparation methods and surface coating play a key role in determining the size distribution, morphology, magnetic properties, and surface chemistry of MNPs.16 Co-precipitation is the most widely used method for the synthesis of MNPs of controlled size and magnetic proper- ties. It is extensively used for biomedical applications of MNPs, because of the ease of preparation and the avoid- ance of harmful materials and procedures.16 Common problems in the preparation of magnetic nanoparticle dis- persions are agglomeration and oxidation, which may re- sult in the loss of dispersibility and magnetism. An addi- tional coating of the magnetic core helps to prevent particle agglomeration and aggregation.12 The coating method not only prevents agglomeration and oxidation of the particle, but also provides physical and chemical stability. The coat- ing provides an interface between the magnetic nanoparti- cle and the surrounding environment and offers the possi- bility for further functionalization. The properties of the coating may markedly differ from those of the nanoparti- cle core.15,17,18 Magnetic nanoparticles are commonly coated with organic (polymers or surfactants such as poly- ethylene glycol and dextran) or inorganic layers (gold, platinum, cobalt oxide, aluminium oxide, silica, activated carbon, etc.).19 The coating helps to obtain a specific affin- ity to target molecules, to increase dispersion stability, and to improve other physicochemical properties.9,20 The adsorption of heavy metal ions on MNPs com- bined with magnetic separation has been used extensively in water treatment and environmental clean-up.14,21 Func- tionalized magnetic nanoparticles act towards metal ions as a kind of “nano-sponges” and can easily be retrieved from solution with a magnet, thus they are excellent for the selective extraction of metal traces from wastewater or in- dustrial effluents. After the adsorbed ions are stripped, the nanoparticles can be reused, making this procedure a promising sustainable green technology.22,23 Core@shell MNPs composed of a maghemite core (γ-Fe2O3), and a functionalized silica coating have recently been applied for the adsorptive removal of Cu2+ and the rare earth ions Tb3+ and Dy3+.24,25 For the adsorption of Nd3+ from aqueous solutions, various magnetic nanoparticles have been synthesized. Ashour et al. synthesized magnetite nanoparticles func- tionalized with citric acid (CA@Fe3O4 NPs) or l-cysteine (Cys@Fe3O4 NPs) for the adsorption of La3+, Nd3+, Gd3+ and Y3+ from aqueous solution.26 Dupont et al. synthe- sized Fe3O4@SiO2(TMS-EDTA) nanoparticles for the ex- traction and separation of different rare-earth ions.27 Gal- houm et al. used Cysteine-functionalized chitosan magnetic nanoparticles for the sorption of La3+, Nd3+ and Yb3+ and hybrid chitosan magnetic nanoparticles func- tionalized by diethylenetriamine (DETA) for the recovery of Yb3+, Dy3+and Nd3+.28,29 Gok investigated batch adsorp- tion method as a green technology for removal and recov- ery of Nd and Sm using magnetic nano-hydroxyapatite adsorbent (MNHA).30 Li et al. prepared mesoporous mag- netic Fe3O4@mSiO2–DODGA nanomaterials for adsorp- tion and recycling of REEs. The surface of mesoporous Fe3O4 particles was modified with a diglycolamide li- gand.31 Liu et al. worked with magnetic bio-adsorbent Fe3O4-C18-chitosan-DETA (FCCD) composite to test the adsorption capacity of Dy3+, Nd3+, and Er3+.32 Miraoui et al. studied the sorption capacities of Nd3+ on magnetic na- noparticles grafted by poly(aminoethylene N-methyl 1-formic acid, 1-phosphonic acid) (PAEMFP).33 Molina et al. synthesized adsorbents based on functionalized mag- netite nanoparticles for the uptake of La3+, Pr3+ and Nd3+ from aqueous solutions.34 In this paper we extend the application of γ-Fe2O3- NH4OH@SiO2 nanoparticles functionalized with amino- propyl trimethly silane (APTMS) towards the adsorption of Nd3+ ions from dilute aqueous solutions. To the best of our knowledge, γ-Fe2O3-NH4OH@SiO2(APTMS) mag- netic nanoparticles have not been previously used for the removal of Nd3+ from aqueous solutions. The monitoring of the adsorption process of dis- solved heavy metal and rare-earth ions on MNPs requires the analytical detection of either the adsorbed ion concen- tration or the depletion of ions in solution. Aqueous solu- tions of Nd3+ appear strongly coloured and suggest the use of UV-Vis spectroscopy as a simple and reliable method for the determination of Nd3+ ion concentration in solu- tion. After separation of the MNPs with an external mag- net the remaining, Nd3+ ion concentration in the superna- tant is expected to reveal the REE removal efficiency at different adsorption time. Upon a re-dispersion of the MNPs, the adsorption process may continue thereby giv- ing fast and easy access to the characterization of the ki- netics and the optimization of the adsorption process. However, the polydisperse size distribution and the stabi- lization of the core@shell MNPs compete with the mag- netic force applied for their separation. We demonstrate that the additional separation processes of centrifugation and filtration are required for a complete removal of MNPs in order to obtain reliable information on the Nd3+ ion ad- sorption efficiency. These additional steps for MNP sepa- ration add complexity to the process and reduce the appar- ent benefits of REE recovery by the adsorption on MNPs. 2. Materials and Methods Iron(II) chloride tetrahydrate (FeCl2·4H2O), iron(I- II) chloride hexahydrate (FeCl3·6H2O), and tetraethyl or- 828 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... thosilicate (TEOS) were obtained from Merck. (3-amino- propyl)trimethoxysilane (APTMS, 97%) and 2-propanol ((CH3)2CHOH, ≥  99.8%) were obtained from Sigma Aldrich. Ethanol (C2H5OH) was obtained from Carlo Erba Reagents, ammonia solution (NH4OH, 25%) from Alka- loid AD, Skopje. Potassium chloride (KCl) and nitric acid (HNO3, ≥ 65%) were obtained from Kemika. All chemicals were used as received, without any further purification. Deionized water (dH2O) supplied by a water purification unit (MilliporeSigma, Burlington, USA) was used through- out the experiments. Neodymium(III) oxide (Nd2O3) ob- tained from Sigma Aldrich was used for the synthesis of neodymium(III) nitrate hexahydrate (Nd(NO3)3·6H2O). 2. 1. Synthesis of γ-Fe2O3-NH4OH@ SiO2(APTMS) For the preparation of the maghemite-silica core@ shell MNPs the protocol described by Kegl et al. was used.35 In brief, the γ-Fe2O3 magnetic nanoparticles were obtained by a co-precipitation method. To prepare 50 mL of 0.5 M Fe2+/Fe3+ solution in dH2O, FeCl2·4H2O and Fe- Cl3·6H2O were used in molar ratio 1:2. 25% ammonia solution (150 mL) was added to a round-bottomed flask and heated under reflux and constant stirring at 300 rpm. The temperature was maintained at 87 °C. Prepared 0.5 M Fe2+/Fe3+ solution (50 mL) was added instantaneously to the reaction mixture and kept for 1 h at 87 °C and pH 10.6. The obtained black coloured precipitate was then thor- oughly rinsed with dH2O and separated from the superna- tant using a permanent magnet. Rinsed γ-Fe2O3 particles were stabilized in 25 mL of 25% ammonia solution at 50 °C under constant stirring at 300 rpm for 24 h. The obtained particles were precipitated from the reaction mixture by a permanent magnet. The γ-Fe2O3-NH4OH particles were functionalized by SiO2 and APTMS. 2-propanol (66  mL), dH2O (15.42  mL), ammonia solution (1.7  mL, 25%), TEOS (0.324 mL, 99%) and APTMS (0.518 mL) were added to 4.93 mL aqueous dispersion of γ-Fe2O3-NH4OH. The re- action was carried out for 24 h in a closed vessel at room temperature and stirring at 500  rpm. The γ-Fe2O3- NH4OH@SiO2(APTMS) particles were precipitated from the reaction mixture by a permanent magnet and washed two times with ethanol and dH2O, respectively. 2. 2. Synthesis of Nd(NO3)3·6H2O Neodymium nitrate hexahydrate (Nd(NO3)3·6H2O) was synthesised from neodymium oxide (Nd2O3) powder. First, a small quantity of dH2O was added to cover the Nd2O3 powder, followed by HNO3 (≥  65%). Nd2O3 and HNO3 were used in molar ratio 1:2. The solution was heat- ed to 90 °C and mixed in a closed beaker to achieve a clear solution. When Nd2O3 was completely dissolved and the solution was clear, the liquid content of the mixture was evaporated at 110  °C. The collected light purple crystals were dried at room temperature. 2. 3. Adsorption Protocol Stock solutions of 0.05  M, 0.025  M and 0.01  M Nd(NO3)3·6H2O were prepared by dissolving an appropri- ate amount of obtained salt in dH2O. Adsorption experi- ments were conducted by mixing 5 mL of the stock solu- tion with 12  mg of magnetic nanoparticles (γ-Fe2O3- NH4OH@SiO2(APTMS)). Functionalized magnetic nano- particles were dispersed in the stock solution by placing the sample in an ultrasonic bath (Iskra PIO, Sonis 10) for 3 hours. After the reaction time, the adsorbent was separat- ed from the solution by an external magnet, centrifugation at 11000 rpm for 5 min (Eppendorf, centrifuge 5804 R), and filtration with 200 nm and/or 20 nm pore-size filters (Whatman, Anotop 25), to remove the remaining nanoad- sorbent. 2. 4. Characterization The thermal behaviour and stability of the magnetic nanoparticle samples was studied using a Mettler Toledo TGA/DSC1 thermogravimetric analyser in air and N2 at- mosphere, respectively, at a gas flow rate of 100 mL/min. The TGA curves with the weight pattern and heat flow were recorded as a function of temperature in the range of 25–600 °C with a heating rate of 10 °C/min, using alumina crucibles. The used MNPs were pre-dried for 24 hours at 80 °C. The presence of 6 mol of water in Nd(NO3)3·6H2O was confirmed using the same thermogravimetric analyser by direct heating in the range of 25–700 °C at a heating rate of 10 °C/min in N2, O2 and air atmosphere with 50 mL/ min flow rate. To confirm the formation of Nd2O3, X-ray diffrac- tion (XRD) analysis was conducted after TGA analysis us- ing an X’Pert PRO (PANalytical) X-ray diffractometer coupled with Cu Kα radiation with a wavelength of 0.15406 nm. The measurement was performed at room tempera- ture with a time step of 100 s in the angular range of 10° to 70° with a step size of 0.034°. Fully open (2.122 ) X’Celera- tor detector was used in the measuring protocol. The Brunauer, Emmet and Teller (BET) theory was used to determine the specific surface area of the nanopar- ticles by using the Micromeritics Tristar II 3020 Surface Area and Porosity system. The samples were degassed at 40°C for 24 hours prior to each measurement by using the Micromeritics FlowPrep 060 Gas Adsorption Sample Preparation Device. Specific surface area was determined in the relative pressure range of 0.05–0.3 in nitrogen gas and temperature of –195.8 °C. A Perkin Elmer Spectrum GX ATR-FTIR spectrom- eter was used to confirm the grafting of the organic ligands to the surface of the MNPs. Spectra were recorded over the 829Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... range of 4000 cm–1 to 400 cm–1 in transmission mode at a resolution of 4 cm–1. Samples were dried for 24 h at 80 °C, ground into a fine powder, placed on the ATR crystal, and pressed into a thick film. Transmission electron microscopy (TEM) images were obtained to analyse the morphology and size distri- bution of the nanoparticles. TEM analyses were performed with a JEOL 2010F model transmission electron micro- scope operating at 200 kV. The sample was prepared in a water solution, dropped onto a carbon-copper grid, and dried at room temperature. Digital Micrograph and Orig- inPro 2015 software were used for image analysis. Zeta potential and size of the MNPs were measured using a Malvern Zetasizer Nano ZS with automatically chosen settings. Size characterization of the samples was made by dynamic light scattering (DLS) measurements with a 4  mW He–Ne laser operating at a wavelength of 633 nm, and a detection angle of 173° (backward scatter- ing). A disposable cuvette was filled up to 1 cm with the particle dispersion. The zeta potential was measured using a combination of electrophoresis and laser Doppler veloci- metry. For all measurements, a voltage of 50 V was applied. The attenuator index and measurement position are auto- matically adjusted by the software. For zeta potential measurements a folded capillary cell was used. Nanoparti- cle titration was performed using the Malvern MPT-2 Au- totitrator in parallel with the Zetasizer Nano ZS. A titra- tion from neutral to high pH using 0.1  M NaOH was performed to determine the isoelectric point (IEP). A sub- sequent titration from high to low pH with 0.1 M HCl was performed to determine the reversibility of the zeta poten- tial. The concentration of MNPs in dH2O was 0.01% V/V. UV-Vis absorption spectra of Nd3+ solutions were recorded in the range of 200–800 nm using a Varian Cary 1 UV–Vis spectrophotometer and a quartz cuvette. For the preparation of the reference sample, magnetic nanoparti- cles (12 mg) were added to 1 mM KCl (5 mL). After 3 hours the MNPs were separated from the solution by an external magnet, centrifugation (11000 rpm, 5 min), and filtration with 200 nm and/or 20 nm pore-size filters. As samples, 5 mL of Nd(NO3)3·6H2O stock solution of different con- centrations (0.05, 0.025, 0.01 M, respectively) and MNPs (12 mg) were used. 3. Results and Discussion The results of TGA of MNPs in air (blue curve) and nitrogen (red curve) atmosphere are displayed in Fig. 1, where the mass loss in percentages of γ-Fe2O3-NH4OH@ SiO2(APTMS) NPs by heating up to 600 °C is presented. The TGA curves show that the mass loss in both atmos- pheres occurs in one major step. In the range from 30 °C to 180 °C absorbed alcohol and water molecules evaporate from the MNPs surface. The interval from 280 °C to 600 °C is associated with the thermal decomposition of amino groups (NH2) and the removal of the alkyl chains of silanes from the silica coating.11,35,36 Around 550°C the transition of maghemite to hematite occurs without any mass change.37 Figure 1: Mass loss of γ-Fe2O3-NH4OH@SiO2(APTMS) in N2 and air atmosphere The mass loss in air atmosphere corresponds to 6.7% of the sample weight. In N2 atmosphere the mass loss pre- sents 7.3% of the sample weight. The difference in mass loss between both atmospheres is 0.6%. Both curves show the same trend for the evolution of the mass loss with tem- perature. We cannot see any additional effect of oxygen on MNPs and therefore no significant sign of oxidation. The thermal decomposition of Nd(NO3)3·6H2O is a complex step-wise process, which starts with the simulta- neous condensation of 6 mol of the initial monomer Nd(NO3)3·6H2O into the complex [Nd(NO3)3·6H2O]6. The main volatile products of the thermal decomposition are water, nitric acid, the azeotrope of 68% HNO3 and 32% H2O, nitrogen dioxide and oxygen.38 The results of TGA in air (blue curve), nitrogen (red curve) and oxygen (green curve) atmosphere are displayed in Fig. 2, where the mass loss in percentages of Nd(NO3)3·6H2O by heating up to 700 °C is presented. Figure 2: Mass loss of Nd(NO3)3· 6H2O in N2, O2 and air atmos- phere 830 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... The thermal decomposition of samples resulted in a mass loss of 58.4% in N2, 58.1% in O2 and 65.2% in air at- mosphere. The formation of Nd2O3 as the final product of TGA was confirmed with XRD analysis. The theoretical mass loss for the thermal decomposi- tion of Nd(NO3)3·6H2O is 61.6%.38 There are no significant changes between oxidizing and inert atmospheres. All three curves have similar trends for the evolution of mass loss with temperature. The de- composition occurs in multiple steps, regardless of the at- mosphere. The stepwise mass loss in N2 atmosphere is ex- plained in accordance with the experimental results.38 The first mass loss up to 62 °C reflects the evaporation of water during melting of the hexahydrate. It constitutes 2%, that corresponds to 3 mol of water out of 6 mol avail- able at the beginning of the decomposition process, which are eliminated during evaporation. In the 62–358  °C range, the mass loss is 21.1%, representing the removal of 14 mol of H2O and 5 mol of HNO3. The next mass loss is 23.3% and it takes place in the range of 358–436 °C with the removal of 1 mol HNO3, 5 mol of H2O and 10 mol of NO2. In the range of 436–513 °C 6 mol of H2O and 2 mol of NO2 are removed, constituting the mass loss of 7.6%. The final mass loss equals 4.4% and takes place in the range of 513–660 °C where 5 mol of H2O and 1 mol of O2 are removed. The remaining mass becomes constant at around 660 °C. The sum of the partial losses yields 58.4% and the presence of 6 mol of water in Nd(NO3)3·6H2O was confirmed. The BET analysis gave specific surface area of 78.27 m2/g for γ-Fe2O3-NH4OH@SiO2(APTMS). BJH adsorp- tion average pore width for γ-Fe2O3-NH4OH@SiO2 (APTMS) is 14.07 nm and desorption average pore width is 13.85 nm with total pore volume of 0.32 cm3/g. Infrared spectra of synthesized nanoparticles γ-Fe2O3-NH4OH (black spectrum), γ-Fe2O3-NH4OH@ SiO2(APTMS) (blue spectrum) and Nd3+/ γ-Fe2O3- NH4OH@SiO2(APTMS) (green spectrum) are shown in Fig. 3. The absorption peak at 3380.21 cm–1, attributed to N-H and O-H bonds, is observed in all spectra and it is the broadest band. By comparing the spectra of the coated and the uncoated MNPs, it is observed that the band at 3380.21 cm−1 is less pronounced and shallower for the coated nanoparticles than for the γ-Fe2O3- NH4OH nanoparticles, indicating the presence of the SiO2 shell. The presence of Nd3+ on coated MNPs is vis- ible by an even reduced intensity of the peak at 3380.21  cm–1, which indicates a successful adsorption of Nd3+ ions on MNPs. The absorption band observed at 1627.39 cm−1 is found in all three samples and corre- sponds to the N-H bending vibration. The peak found at 1020.92 cm–1 corresponds to the asymmetric stretch- ing vibration of the Si-O-Si bond and confirms the si- lanol functional groups grafted on the surface of γ-Fe2O3-NH4OH particles. The peak observed at 542.16 cm–1 belongs to the stretching vibration of Fe-O, which confirms the presence of the magnetic core. When comparing uncoated and coated MNPs this peak shifts from 540.46 cm–1 to 542.16 cm–1. When Nd3+ is complexed with γ-Fe2O3-NH4OH@ SiO2(APTMS) nanoparticles, a new peak at 1307.39 cm–1 could be observed in the spectrum. Due to the adsorption of Nd3+ two peaks shift to a higher wavenumber. The peak at 542.16 cm–1, which corresponds to the stretching vibra- tion of Fe-O, is shifted to 548.99  cm–1. The peak at 1020.92  cm–1, which corresponds to the asymmetric stretching vibration of the Si-O-Si bond, is shifted to 1022.44 cm–1. The ATR-FTIR absorption spectra confirm the composition of the synthesised magnetic nanoparti- cles, the successful coating with SiO2(APTMS), and the adsorption of Nd3+. The TEM photos in Fig. 4 A-C display the morphol- ogy of the MNPs with increasing magnification. The magnetic nanoparticles have a diameter ranging between 2.5 nm and 22.5 nm, with an average particle size of 9.5 ± 1.9 nm (Fig. 4 D). This value is an estimate of size, an- alysed from 81 particles, as it is difficult to determine the real size of individual particles due to agglomerates and blurred boundaries between particles. Aggregation of parti- cles results in wider size variation. None of the crystalline particles are ideally spherical, as their surface ends with a crystal plane, which is not curved. If nanoparticles are formed in a way that they can achieve an equilibrium struc- ture, they will assume an octahedral form. But since the pre- cipitation is an instantaneous process, we usually get incom- plete shapes, which are somewhere in between a sphere and an octahedron. As shown in Fig. 4 C the particle is octahe- dral, but because of orientation and 2D projection of TEM images it looks like a hexagon. In Fig. 4 B and C a surface layer of amorphous SiO2 is visible surrounding the Fe2O3 core, which confirms the successful coating of the MNPs. Figure 3: ATR-FTIR spectra of γ-Fe2O3-NH4OH (top curve, black), γ-Fe2O3-NH4OH@SiO2(APTMS) (middle curve, blue) and Nd3+/ γ-Fe2O3-NH4OH@SiO2(APTMS) (bottom curve, green) 831Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... The measuring parameters and the specificity of each measurement technique often result in different sizes for the same sample. Dynamic light scattering (DLS) is com- pared with transmission electron microscopy (TEM) for the characterization of the size distribution of magnetic nanoparticles. TEM measures the geometric size of dry MNPs deposited on a support surface under ultrahigh vacuum conditions. The DLS technique measures the hy- drodynamic diameter that refers to how a particle diffuses within a liquid. Consequently, results obtained from DLS show a larger diameter than those from TEM. The signifi- cant difference in size between DLS and TEM results may be explained by an increased size in DLS due to the pres- ence of the dispersant and the formation of hydrate layers. The presence of bigger particles and aggregates enhances light scattering and can also contribute to larger size val- ues. Even though boundaries between particles in aggre- gates are not always obvious to recognize when analysing TEM images, the MNP size distribution observed by the TEM image-processing technique had a narrower range than that of DLS analysis. Fig. 6 shows the pH dependence of the particle size and the zeta potential for 0.01% V/V MNPs in dH2O. DLS reveals a hydrodynamic diameter of 250 nm at the current pH 7 of the MNP dispersion in dH2O. The TEM image in Fig. 4A already suggests larger aggregates of primary nan- oparticles, which maintain their assembly when dispersed in water. A pH titration was first performed from the na- tive pH 7 to pH 11 using 0.1 mol/L NaOH (measurement 1). Above pH 8, the particle size shows a sudden increase and approaches a steady diameter of 3 µm at pH 9. At higher pH we observe the onset of a trend towards smaller size, but the scatter of data disables a firm conclusion on a decreasing diameter. The direction of the pH change was then reversed by a titration towards the acidic range using 0.1 mol/L HCl. The diameter of 3 µm remains down to pH 8 followed by a sudden decrease and an approach of the initial diameter of 250 nm at pH 5. Below pH 3 we may assume again the onset of a trend towards a small growth of the particle aggregates. Although the changes in the hy- drodynamic diameter of the MNPs is obviously fully re- versible, we observe a hysteresis of the evolution of particle size with pH between the titrations to high and low pH, respectively. To understand the pH dependence of the particle size, the zeta potential of the MNPs was recorded in paral- Figure 4: TEM images of γ-Fe2O3-NH4OH@SiO2(APTMS) with increasing magnification (A: scale bar 100 nm, B: 20 nm, C: 5 nm) and size distri- bution of MNPs (D). 832 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... lel. The zeta potential is the key parameter that controls electrostatic interactions in particle dispersions, provides information about surface functionality and determines the dispersion stability. The zeta potential for particle dis- persions is calculated from the measurement of the elec- trophoretic mobility by electrophoretic light scattering (ELS). At pH 7 a positive zeta potential of ζ = +30 mV was observed, which indicates a significant positive charge density at the nanoparticle-water interface and confirm the presence of the amine functional groups of the outer- most APTMS coating of the core@shell maghemite nano- particles. The amine groups get protonated in water and assume a positive charge.39 When increasing the pH of the aqueous dispersion of the MNPs, the zeta potential de- creases and approaches the isoelectric point (IEP) at pH 9.8. Beyond the IEP the zeta potential assumes a negative sign indicating the charge reversal of the MNPs. The IEP 9.8 is indicative for the moderately basic character of the aminosilane.40 When continuing the pH titration from the alkaline to the acidic range, the zeta potential increases again and shows a charge reversal of the MNPs at a lower IEP 8.5. The electrokinetic charge density achieves a steady state indicated by the plateau value of the zeta potential of ζ = +35 mV below pH 5. We observe the same hysteresis for the pH dependence of the zeta potential and of the par- ticle size when repeating the titration from high to low pH. As a rule of thumb, a zeta potential below –25…– 30 mV or above +25…+30 mV describes a stable disper- sion where the aggregation of nanoparticles is suppressed by the electrostatic repulsion of particles with alike charge.41 This empirical observation is confirmed by the correlation of the hydrodynamic diameter and the zeta po- tential of MNPs shown in Fig. 5. As the initial zeta potential drops below ζ = +25 mV, the size of the MNPs starts to increase indicating the onset of the formation of larger aggregates. As the zeta potential approaches the IEP, the electrostatic repulsion between MNPs becomes weaker, and the average size of the particle aggregates, which remain suspended in dH2O, obtains its maximum diameter of 3 µm. Above the IEP 9.8, the nega- tive zeta potential steadily increases thereby introducing repulsive electrostatic forces between MNPs, which are now negatively charged. Since the zeta potential at pH 11 does not exceed the empirical threshold of ζ = ±25 mV, its effect on the disaggregation of particle assemblies remains rather small. In the opposite direction of the pH titration, the threshold value of ζ = +25 mV is observed at pH 6.5 where the hydrodynamic diameter returns to 250 nm, which is the average size of the MNP aggregates in dH2O. At very low pH, the significant volume of acid (0.1 mol/L HCl) added to decrease pH introduces a simultaneous increase in the ionic strength of the aqueous solution. Although the scatter of results in the range of pH 2.5–4 does not allow a conclusion on a decrease in the zeta potential, the electric double layer at the MNP-water interface gets suppressed at higher ionic strength, which again weakens the repulsive force between positively charged MNPs. The hysteresis observed when repeating the analysis of the pH dependence of the zeta potential and the shift of the IEP from pH 9.8 (for the first titration) to pH 8.5 (for the second titration) indicate a decrease in the average ba- sic strength of the functional coating of the MNPs. Obvi- ously, the stability of the silane coating on the silica shell of the MNPs at higher pH is limited. Figure 5: pH dependence of zeta potential and particle size for a dispersion of 0.01% V/V MNPs in dH2O 3. 1. Adsorption of Nd3+ Ions To elucidate the efficiency of Nd3+ ion adsorption on MNPs, an appropriate analytical method for quantifica- tion of the adsorption process was selected. The concen- tration of Nd3+ ions adsorbed on the MNP surface or the depletion of Nd3+ ion concentration in the aqueous solu- tion may be determined. There have been many analytical techniques used for the determination of the REEs in solid and solution samples; flame or graphite furnace atomic ab- sorption spectrometry, atomic absorption with chemical vapor generation, X-ray fluorescence spectrometry (XRF), inductively coupled plasma optical emission spectrometry (ICP-OES), inductively coupled plasma mass spectrome- try (ICP-MS), high-performance liquid chromatography (HPLC) and neutron activation analysis (NAA).41–43 Among the methods presented above for the characteriza- tion of various properties of the MNPs (composition, functional groups, size, charge), ATR FTIR and ELS may be considered to exhibit changes in the IR spectrum and in the zeta potential, respectively, of MNPs before and after adsorption of Nd3+ ions. As shown in Fig. 3, the ATR FTIR spectrum of the core@shell MNPs after adsorption of Nd3+ ions indicates the presence of neodymium by the additional peak at 1307 cm–1. The intensity of this peak compares with the peak indicating the N-H bending vibration (at 1627 cm–1) and is likely assigned to a vibration of the assumed Nd-NH 833Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... surface complex that shows responsible for the adsorption of Nd3+ ions on the positively charged γ-Fe2O3-NH4OH@ SiO2(APTMS) nanoparticles. Since the intensity of the peak at 1307 cm–1 is rather weak and a calibration protocol is thus not feasible to correlate peak intensity with Nd sur- face concentration, ATR FTIR is only applicable to qualita- tively confirm the presence of Nd. The zeta potential is occasionally applied to describe and to study the adsorption process of dissolved com- pounds such as surfactants, polymers (polyelectrolytes, polysaccharides), or proteins on material surfaces.44–46 It is therefore feasible to investigate the capability of ELS for monitoring the adsorption of Nd3+ ions on MNPs al- though the same positive sign of the charge reduces the sensitivity of zeta potential analysis for the characteriza- tion of adsorption processes. Table 1 shows the zeta potential of MNPs remaining in dispersion after an apparent separation by an external magnet followed by centrifugation and filtration with a 200 nm filter after adsorption of Nd3+ ions from aqueous solutions of Nd(NO3)3 with different bulk concentration (0.01, 0.025, 0.05 mol/L, respectively). The average zeta potential and the standard deviation were obtained from three repetitive measurements. A steady decrease in the positive zeta potential was found from ζ = +47.7 ± 4.2 mV after adsorption from a 0.01 mol/L Nd3+ to ζ = +23.4 ± 2.4 mV when the initial concentration of Nd3+ ions was 0.05 mol/L. In the same series, the conductivity of the corresponding solu- tions increased from 406 mS/m to 1366 mS/m. The de- pendence of the zeta potential on the ionic strength is therefore dominating the decrease in the magnitude of the positive zeta potential of remaining MNPs in dis- persion. Independent of the zeta potential and the ionic strength, the particle size remains at Dh = 80.5 ± 0.8 nm. This hydrodynamic diameter is significantly smaller than the size of Dh = 256 nm of the MNPs determined in the MNP stock solution in deionized water. The dif- ference in the particle size is explained by the removal of the larger fraction of MNPs by filtration with a 200 nm filter. The zeta potential of the bulk dispersion of MNPs in dH2O reads ζ = +28.2 ± 0.5 mV. This zeta po- tential is only 60% of the zeta potential determined in 0.01 mol/L Nd(NO3)3 (the lowest Nd3+ ion concentra- tion used in the series of adsorption experiments) al- though the conductivity of the MNP bulk dispersion (18.4 mS/m) is much lower than the corresponding conductivity (406 mS/m) of 0.01 mol/L Nd3+. If the di- minishingly small ionic strength of deionized water is considered, the Hückel approach for the calculation of the zeta potential can be used, which transfers ζ = +28.2 mV determined in the Smoluchowski limit to ζ = +42.3 mV and therefore closer to the zeta potential obtained in the 0.01 mol/L Nd3+ solution. However, the conduc- tivity of the bulk MNP dispersion of 18.4 mS/m is mul- tiple times higher than the conductivity expected for dH2O (theoretically 0.006 mS/m) indicating a signifi- cant contribution of the charged MNPs to the electric conductance of the aqueous dispersion. A conductivity of 18.4 mS/m corresponds to an ionic strength of ap- prox. 0.001 mol/L when considering a monovalent elec- trolyte such as NaCl or KCl. This ionic strength justifies the application of the Smoluchowski approach for the calculation of the zeta potential especially when consid- ering the average particle diameter of 256 nm. It can be concluded that either the influence of the particle size and/or the high concentration of MNPs in the bulk dis- persion affect the electrophoretic mobility and thus the zeta potential. In conclusion, the strong dependence of the zeta po- tential on the ionic strength does not qualify this method for the estimation of the surface concentration of adsorbed Nd3+ ions on the MNPs. Table 1: Zeta potential (ζ) and hydrodynamic diameter (Dh) of MNPs after adsorption of Nd3+ ions from aqueous solution with different bulk concentration of Nd(NO3)3. Nd3+ bulk Conductivity ζ Dh concentration (mS/m) (mV) (nm) (mol/L) 0 * 18.4 28.2 ± 0.5 256 0.01 406 47.7 ± 4.2 81.3 0.025 906 34.1 ± 3.4 80.3 0.05 1366 23.4 ± 2.4 79.8 * MNPs dispersed in dH2O, i.e., without separation For monitoring the depletion of Nd3+ ions in solu- tion upon adsorption on MNPs, the optical properties of aqueous solutions of this REE suggest the application of UV-Vis spectroscopy. Fig. 6 shows the UV-Vis spectrum of aqueous solutions of Nd(NO3)3 at different concentra- tion (0.01, 0.025, 0.05 mol/L, respectively) in the wave- length range of 540–780 nm. In this range we find two dis- tinct absorption peaks at 575 nm and 740 nm, which may be assigned to the electronic f-f transitions 4I9/2  2G(1)7/2 and 4G5/2 (575 nm, 17391 cm–1) and 4I9/2  4F7/2 (740 nm, 13514 cm–1).47 The absorbance recorded at these peaks was used to establish calibration curves for the estimation of the Nd3+ ion concentration from the measured absorb- ance after completing the adsorption process on MNPs. The corresponding calibration curves are shown in the in- set of Fig. 6. Coincidently at a given Nd3+ concentration, the peaks at 575 nm and 740 nm show almost the same absorbance. Fig. 6 also shows the UV-Vis spectra obtained after adsorption of Nd3+ on MNPs in solutions with the corre- sponding Nd3+ starting concentration (0.01, 0.025 and 0.05 mol/L, respectively) after separation of the MNPs by an external magnet, centrifugation and filtration using a 834 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... filter with 200 nm pores. We do not find any decrease in the absorbance for both f-f transition peaks, which sug- gests that adsorption of Nd3+ ions on γ-Fe2O3-NH4OH@ SiO2(APTMS) nanoparticles did not occur. However, up- on filtration of the supernatant obtained after centrifuga- tion of the MNP dispersion, that is remaining after apply- ing an external magnet, with a 20 nm filter, the absorbance decreases consistently for the peaks at both 575 nm and 740 nm. The evaluation of the depletion of Nd3+ from the stock solution using the UV-Vis calibration curves reveals a removal of 30% at pH 7 and after 3 h of interaction time. DLS and ELS measurements of the filtrate still show clear distributions of size and zeta potential with Dh = 51 nm and ζ = +17.8 ± 1.8 mV. We note that (i) small aggregates of primary MNPs pass the 20 nm filter and (ii) the filter shows a small rejection at the nominal threshold of 20 nm. The measured conductivity in the filtrate reads 1176 mS/m, which is 14% lower than the conductivity in the filtrate of the 200 nm filter. We may thus assume a larger zeta potential but observe the opposite. Obviously, the dif- ferent size fractions of MNPs exhibit different effective charge density represented by the difference in the zeta po- tential. The presence of MNPs in the filtrate of the 20 nm filter suggests a contribution of Nd3+ adsorbed on these MNPs to the UV-Vis absorption spectrum. The removal efficiency is therefore expected slightly higher than the es- timated 30%. From the change in the absorbance for MNPs filtered with a 20 nm filter after adsorption of 0.05  M Nd3+, we conclude that 0.035 M Nd3+ remain in solution. The ad- sorption efficiency after 3 hours is therefore 30%. 4. Conclusions Magnetic nanoparticles with a γ-Fe2O3 core, coated with SiO2 and functionalized with APTMS, were synthe- sized with the purpose of removing rare earth elements from aqueous solutions. The main reason for selecting Nd3+ as the source of REE is the fact that neodymium is one of the most important REE in terms of usage and ap- plicability. Furthermore, neodymium is among those RE- Es that show a lot of absorption peaks in the visible spec- trum, which makes UV-Vis absorption spectroscopy convenient for the determination of the Nd3+ concentra- tion in solution. The synthesized MNPs were character- ized in terms of size (TEM, DLS), charge (zeta potential; ELS), surface functionality (ATR-FTIR), and composition (TGA, XRD). The adsorption of Nd3+ on γ-Fe2O3- NH4OH@SiO2(APTMS) was performed in batch mode. The UV-Vis calibration curves revealed maximum adsorp- tion to be 30% at pH 7, with use of 12 mg of MNPs in 0.05 mol/L of neodymium solution. Based on the results, an obvious decrease in neodymium concentration becomes visible only after a reaction time of 3 hours, which makes this process uneconomical for large-scale use. The removal of γ-Fe2O3-NH4OH@SiO2(APTMS) from dispersion after Nd3+ adsorption requires multiple steps (removal with a magnet, centrifugation, and filtration) in order to measure the absorbance spectrum without a significant interfer- ence of MNPs remaining in the processed solution, which also requires time and adds additional steps to the process. Ban et al. reported the use of ultrafiltration to completely remove even the primary particles of MNPs.48 The slow Figure 6: Absorbance spectra of Nd3+ before and after adsorption on MNPs. The labels of spectra recorded after adsorption indicate the filter used for removing MNPs (pore size 20 nm, 200 nm). 835Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... adsorption of Nd3+ ions on γ-Fe2O3-NH4OH@SiO2 (APTMS) suggests a coating by different chelating com- pounds, such as polyethylenimine or negatively charged MNPs.34 Acknowledgements The authors acknowledge dr. Julija Volmajer Valh for her help with ATR-FTIR, dr. Sašo Gyergyek for TEM anal- ysis, Edi Kranjc for XRD analysis, and dr. Karl Gatterer of Graz University of Technology for donating Nd2O3. 5. References 1. D. A. Atwood: The rare earth elements: Fundamentals and Applications, John Wiley & Sons Ltd, 2012, pp. 2–27. 2. J. H. L. Voncken: The Rare Earth Elements: An Introduction, Springer, 2016, pp. 1–11. DOI:10.1007/978-3-319-26809-5 3. U.S. department of Energy, Critical Materials Strategy, https://www.energy.gov/sites/prod/files/DOE_CMS2011_FI- NAL_Full.pdf, (accessed: February 18, 2022) 4. V. Balaran, Geosci. Front., 2019, 10, 1285–1303. DOI:10.1016/j.gsf.2018.12.005 5. C. E. D. Cardoso, J. C. Almeida, C. B. Lopes, T. Trindade, C. Vale and E. Pereira, Nanomater., 2019, 9, 1–55. DOI:10.3390/nano9060814 6. K. Binnemans, P. T. Jones, B. Blanpain, T. Van Gerven and Y. Pontikes, J. Clean. Prod., 2015, 99, 17–38. DOI:10.1016/j.jclepro.2015.02.089 7. R. Ashour, R. Elsayed, A. F. Abdel-Magied, K. M. Forsberg, A. Uheida and M. Muhammed, Chem. Eng. J., 2017, 327, 286–296. DOI:10.1016/j.cej.2017.06.101 8. A. Negrea, A. Gabor, C. M. David, M. Ciopec, P. Negrea, N. Duteanu and A. Barbulescu, Sci. Rep., 2018, 8, 1–11. DOI:10.1038/s41598-017-18623-0 9. M. A. Agotegaray, V. L. Lassalle: Silica-coated Magnetic Nan- oparticles: An Insight into Targeted Drug Delivery and Toxi- cology, Springer, 2017, pp. 12–24. DOI:10.1007/978-3-319-50158-1 10. L. L. Vatta, R. D. Sanderson, K. R. Koch, Pure Appl. Chem., 2006, 78, 1793–1801. DOI:10.1351/pac200678091793 11. T. Kegl, A. Košak, A. Lobnik, I. Ban, Z. Novak and A. Kovač Kralj, J. Hazard. Mater., 2019, 386, 1–59. DOI:10.1016/j.jhazmat.2019.121632 12. S. Majidi, F. Zeinali Sehrig, S. Mussa Farkhani, M. Soleymani Goloujeh, A. Akbarzadeh, Artif. Cells Nanomed. Biotechnol., 2016, 44, 722–734. DOI:10.3109/21691401.2014.982802 13. Y. Bao, T. Wen, A. C. S. Samia, A. Khandhar, K. M. Krishnan, J. Mater. Sci., 2016, 51, 513–553. DOI:10.1007/s10853-015-9324-2 14. M. Sharma, P. Kalita, K. K. Senapati, A. Garg: Study on Mag- netic Materials for Removal of Water Pollutants, 2018, 61–78. DOI:10.5772/intechopen.75700 15. N. Zhu, H. Ji, P. Yu, J. Niu, M. U. Farooq, M. Waseem Akram, I. O. Udego, H. Li, X. Niu, Nanomater., 2018, 10, 1–27. 16. P. Xu, G. M. Zeng, D. L. Huang, C. L. Feng, S. Hu, M. H. Zhao, C. Lai, Z. Wei, C. Huang, G. Xin Xie and Z. Feng Liu, Sci. Total Environ., 2012, 424, 1–10. DOI:10.1016/j.scitotenv.2012.02.023 17. S. P. Gubin: Magnetic Nanoparticles, Wiley-VCH, 2009, 25– 58. DOI:10.1002/9783527627561 18. A. Jurgec, Funkcionalizirani magnetni nanodelci za odstran- jevanje težkih kovin iz odpadnih vod, https://dk.um.si/Izpis- Gradiva.php?id=74201&lang=slv. (accessed: February 20, 2022) 19. I. Ali, C. Peng, I. Naz and M. A. Amjed: Magnetic Nanostruc- tures, Springer, 2019, 161–179. 20. J. Kudr, Y. Haddad, L. Richtera, Z. Heger, M. Cernak, V. Adam, O. Zitka, Nanomater., 2017, 9, 1–28. 21. A. M. Gutierrez, T. D. Dziubla, J. Z. Hilt, Rev. Environ. Health, 2017, 32, 111–117. DOI:10.1515/reveh-2016-0063 22. O. Pereao, C. A. Bode-Aluko, O. Fatoba, K. Laatikainen, L. F. Petrik, Desalination Water Treat., 2018, 130, 71–86. DOI:10.5004/dwt.2018.22844 23. D. Dupont, W. Brullot, M. Bloemen, T. Verbiest and K. Bin- nemans, ACS Appl. Mater. Interfaces., 2014, 7, 4980–4988. DOI:10.1021/am406027y 24. O. Plohl, U. Ajdnik, S. Gyergyek, I. Ban, A. Vesel, T. Kraševac Glaser, L. Fras Zemljič, J. Environ. Chem. Eng., 2019, 7, 6777– 6789. DOI:10.1016/j.jece.2019.102913 25. O. Plohl, M. Finšgar, S. Gyergyek, U. Ajdnik, I. Ban and L. Fras Zemljič, Nanomater., 2019, 209, 1–20. DOI:10.3390/nano9020209 26. R. M. Ashour, R. El-sayed, A. F. Abdel-Magued, A. A. Ab- del-khalek, M. M. Ali, K. Forsberg, A. Uheida, M. Mu- hammed and J. Dutta, J. Chem. Eng., 2017, 327, 286–296. DOI:10.1016/j.cej.2017.06.101 27. D. Dupont, J. Luyten, M. Bloemen, T. Verbiest and K. Binne- mans, Ind. Eng. Chem. Res., 2014, 53, 15222–15229. DOI:10.1021/ie502546c 28. A. A. Galhoum, M. G. Mafhouz, S. T. Abdel-Rehem, N. A. Gomaa, A. A. Atia, T. Vincent and E. Guibal, Nanomater., 2015, 5, 154–179. DOI:10.3390/nano5010154 29. A. A. Galhoum, M. G. Mafhouz, S. T. Abdel-Rehem, N. A. Gomaa, A. A. Atia, T. Vincent and E. Guibal, Cellulose., 2015, 22, 2589–2605. DOI:10.1007/s10570-015-0677-0 30. C. Gok, J. Radioanal. Nucl. Chem., 2014, 301, 641–651. DOI:10.1007/s10967-014-3193-z 31. J. Li, A. Gong, F. Li, L. Qiu, W. Zhang, G. Gao, Y. Liu and J. Li, RSC Adv., 2018, 68, 39149–39161. 32. E. Liu, X. Zheng, X. Xu, F. Zhang, E. Liu, Y. Wang, C. Li and Y. Yan, NJC, 2017, 15, 7739–7750. DOI:10.1039/C7NJ02177A 33. A. Miraoui, M. A. Didi and D. Villemin, JRNC, 2016, 307, 963–971. DOI:10.1007/s10967-015-4267-2 34. L. Molina, J. Gaete, I. Alfaro, V. Ide, F. Valenzuela, J. Parada and C. Basualto, J. Mol. Liq., 2019, 275, 178–191. DOI:10.1016/j.molliq.2018.11.074 35. S. D. Topel, E. Polido Legaria, C. Tiseanu, . J. Rocha, J.-M. Nedelec, . V. G. Kessler and . G. A. Seisenbaeva, J. Nanoparti- cle Res., 2014, 2783. 36. V. Rocher, J. Manerova, M. Kinnear, D. J. Evans, Dalton Trans., 2014, 43, 2948–2952. DOI:10.1039/C3DT52386A 836 Acta Chim. Slov. 2022, 69, 826–836 Ambrož et al.: Assessment of the Capability of Magnetic Nanoparticles ... 37. P. Melnikov, I. V. Arkhangelsky, V. A. Nascimento, L. C. S. de Oliveira, G. A. Wandekoken, D. Muniz de Albuquerque, J. Therm. Anal. Calorim., 2019, 1117–1121. 38. J. G. Smith: Organic Chemistry, McGraw-Hill, 3rd edition, 2011, pp. 1075–1086. 39. P. Nygård, . K. Grundke, E. Mäder, C. Bellmann, J. Adhes. Sci. Technol., 2002, 16, 1781–1808. DOI:10.1163/156856102320396148 40. T. Luxbacher: The Zeta Potential for Solid Surface Analysis, Anton Paar GmbH, 2014, pp. 10–138. 41. B. Zawisza, K. Pytlakowska, B. Feist, M. Polowniak, A. Kita, R. Sitko, J. Anal. At. Spectrom., 2011, 26, 2373–2390. DOI:10.1039/c1ja10140d 42. A. Anggraeni, F. Arianto, A. Mutalib, U. Pratomo, H. H. Bah- ti: Mathematics, Science, and Computer Science Education (MSCEIS 2016): AIP Conf. Proc., 2017. 43. E. Zolfonoun, S. R. Yousefi, J. Braz. Chem. Soc., 2016, 27, 1–6. 44. Z. Adamczyk, M. Nattich, M. Wasilewska, M. Zaucha, Adv. Colloid Interface Sci., 2011, 168, 3–28. DOI:10.1016/j.cis.2011.04.002 45. J. Ethève and P. Déjardin, Langmuir, 2002, 18, 1777–1785. DOI:10.1021/la011224d 46. C. Payerl, M. Bračič, A. Zankel, W. J. Fischer, M. Kaschowitz, E. Fröhlich, R. Kargl, F. Stelzer, S. Spirk, Carbohydr. Polym., 2017, 164, 49–56. DOI:10.1016/j.carbpol.2017.01.088 47. P. Caro, D. R. Svoronos, E. Antic, M. Quarton, J. Chem. Phys., 1977, 66, 215–220. DOI:10.1063/1.433910 48. I. Ban, S. Markuš, S. Gyergyek, M. Drofenik, J. Korenak, C. Helix-Nielsen, I. Petrinić, J. Nanomater., 2019, 9, 1–17. DOI:10.3390/nano9091238 Except when otherwise noted, articles in this journal are published under the terms and conditions of the  Creative Commons Attribution 4.0 International License Povzetek Magnetni nanodelci (MNP) so zaradi svoje hitre sinteze, vsestranske funkcionalizacije in možnosti recikliranja z upora- bo magnetnega polja deležni vse večje pozornosti za različne aplikacije. Visoko razmerje med površino in prostornino disperzij MNP predstavlja zmožnost njihove uporabe kot adsorbenta za odstranjevanje ionov težkih kovin. Raziskali smo uporabnost MNP, sestavljenih iz maghemitnega jedra in obdanih s kremenčevo lupino, funkcionalizirano z amino- propilsilanom, γ-Fe2O3-NH4OH@SiO2 (APTMS), za odstranjevanje neodimovih ionov (Nd3+) iz vodne raztopine. MNP so bili karakterizirani glede na velikost, sestavo, površinsko funkcionalnost in naboj. Kljub obetavnim lastnostim MNP pa njihova odstranitev iz vodne disperzije z zunanjim magnetom ni zadostovala za zanesljivo kvantificiranje adsorpcije Nd3+ z UV-Vis spektroskopijo.