f-Block element separation mediated by carboxylated Fe₃O₄ nanoparticles as robust adsorbents in acidic systems

Abstract

Citric acid functionalized magnetic nanoparticles (CA-MNP) were synthesized and studied for sorption of two representative trivalent lanthanide and actinide ions (Eu³⁺ and Am³⁺). The material was characterized using various analytical techniques such as FTIR, SEM, TEM and zeta potential measurement etc. which indicated successful synthesis of magnetic nanoparticles and citric acid coating on its surface. The uptake study indicated efficient sorption of Eu³⁺ (76.7%) and Am³⁺ (60.9%) at optimum concentration of CA-MNP) 2.5 mg mL^− 1. Sorption of metal ion onto the substrate was confirmed by EDXRF results. The sorption process indicated better kinetics for Eu³⁺ compared to previously reported studies and both the radionuclides were found to be following pseudo-second order rate kinetics. The rate constants were found to be 4.76 × 10^− 9 mg g^− 1 min^− 1 and 2.45 × 10^− 7 mg g^− 1 min^− 1 for Am³⁺ and Eu³⁺, respectively. Thermodynamic study indicated ΔG values as − 15.8 kJ mol^− 1 and − 17.9 kJ mol^− 1 for Am³⁺ and Eu³⁺, respectively indicating the spontaneity of the sorption processes. Stripping with different reagents viz. 0.1 M EDTA, 0.1 M oxalic acid, 0.1 M Na₂CO₃, and 0.1 M HNO₃ indicated best results with 0.1 M EDTA. >99% of loaded Eu³⁺ is stripped in two stripping cycles while stripping of Am³⁺ requires more number of cycles to quantitatively strip the loaded radionuclide.

Introduction

Origin of radionuclides in environment is primarily from natural sources such as primordial radionuclides (U, Th and their daughter products) and due to anthropogenic activities, such as fallout from nuclear weapon testing and nuclear accidents¹. Elevated concentrations of radionuclides in environmental matrices may be found in U/Th mineralized regions (mostly natural radioactive series elements) as well as due to dispersal of radioactive elements in nuclear incidents (fission/activation products from nuclear reactors and radioisotopes from the radioactive sources) near the incident site. One of the important applications of nuclear energy is production of electrical power. It is anticipated that the total energy-generating capacity of nuclear power for global demand may increase by as much as 82%. Presently the strategy adapted for most of the countries to manage the nuclear waste generated in nuclear power reactors is the interim storage of waste vitrified into the glass matrix followed by burial in deep geological repositories at suitable locations. Despite of stringent protocols for storage and monitoring of the stored vitrified nuclear waste, leaching and subsequent dispersion of radionuclides such as fission and activation products comprising of lanthanides and actinides into environment is possible under accidental scenario and natural calamities. The released radionuclide may contaminate various environmental matrices such as soil, ground water, surface water etc. and ultimately ends up accumulating in living organisms including humans via food chain. This may result in exposure of general public to the radiation doses significantly higher than the limits regulated by authorities². Amongst the various environmental matrices, groundwater is of major concern as it is primary source of drinking water and irrigation. Contaminants from ground water have a large probability to end up in food chain resulting in harmful effects to the public. Considering the facts stated, it is important to develop analytical methods for quick identification as well as quantification of radionuclides released into groundwater along with proper remediation strategies.

In recent years lots of emphasis is given to the treatment of contaminated water with sorption of radionuclides onto various substrates such as natural clays³. Also, ion exchange resins or membranes has been reported as an efficient method for decontamination^4,5. Various studies have suggested the use of carbon nanomaterials such as carbon nanotubes (CNTs) and graphene oxide (GO) for the efficient uptake of radionuclides from contaminated water^6,7,8. Magnetic nanoparticles such as Fe₃O₄ coated with complexing agents provide an efficient method for pre-concentration of radioactive element from aqueous medium. Their high surface area and small size (high surface area to volume ratio), better adsorption kinetics, simple mode of operation, easy recovery by magnetic separation are some of the advantages over other materials used for pre-concentration/sorption of radioelements⁹. In view of these advantages, Fe₃O₄ is considered as potential sorbent for radionuclides from aqueous medium^10,11. Further, coating with environmentally benign complexing agents gives added advantage from waste management point of view^12,13. Citric acid (CA) is a common complexing agent which can interact with the lanthanides and actinide ions and influence the fate of these radioelements in aqueous medium¹⁴. Application of CA as decontamination agent in nuclear facilities is well known. Though different studies were carried out on interaction of CA with radionuclides involving its complexation behavior, influence of CA on sorption of radionuclides on clay minerals etc^15,16,17, only a few studies reported uptake of radionuclides by CA and its composite materials^18,19. Considering the above facts, a detailed study on uptake of trivalent lanthanide and actinide on Citric acid-functionalized magnetic nanoparticles (CA-MNP) was thought as an important contribution to the present knowledge. In the present study, CA-MNP were synthesized and characterized by various analytical techniques such as UV-Vis spectroscopy, Fourier-transform infrared spectroscopy (FTIR) analysis, Zeta potential (ZP) measurements and X-ray diffraction (XRD) analysis. Additionally, Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) analysis was carried out for the examination of morphology and mean particle size of the prepared CA-MNP. The uptake of the representative lanthanide and actinide (Eu³⁺ and Am³⁺) from their aqueous solutions by the prepared CA-MNP was studied using the batch extraction method. The effects of various parameters, such as equilibration time and pH, were also studied. The sorption kinetics and adsorption isotherms for these ions were also studied in detail to understand the mechanism of the sorption of metal ions onto the prepared material.

Result and discussion

Characterization of CA-MNP

The absorption spectra of as-synthesized CA-MNP are shown in Fig. 1a the hump at 378 nm confirms the presence of the carboxylate. In contrast, the uplifting of the absorption spectra from the range of 600–450 nm confirms the formation of magnetite nanoparticles.spectra from the range of 600 –450 nm confirms the formation of magnetite nanoparticles (Fig. 1a). An inset shows the separation of Fe₃O₄ nanoparticles after applying the magnetic field, confirming the magnetic nature of these nanoparticles. The surface coating of carboxylate from the carboxylic acid is examined using the FTIR measurements. The characteristic band of Fe–O stretching revealing the spinel crystal structure of the nanoparticles is observed at 539 cm^− 1 in as-synthesized MNPs. This confirms the formation of Fe₃O₄ NPs. The less intense peaks at 1631 cm^− 1 and 2111.49 cm^− 1 are due to the O–H stretching due to any moisture remaining in the MNPs²⁰. The functional groups of the carboxylate-coated MNP were confirmed using the FT-IR spectrum (Fig. 1b). It is observed that the Fe–O stretching band has slightly shifted to 564 cm^− 1 due to the coating of citric acid, which strengthens the Fe–O bond. The aliphatic C–H str is observed at 1242 cm^− 1. The C=O str originating from the free carboxylic bond is observed at 1589 cm^− 1 and the C=O str originating from the coordinated carboxylic bond is observed at 1365 cm^− 121. Carboxylate groups of CA from complexes with Fe atoms on the surface of Fe₃O₄ rendering partial single bond character to the C=O this results in weakening of C=O which shifts the stretching frequency to lower vibrational values. The broad band of O–H str is observed at 2990 cm^− 1 due to the presence of water on the surface of Fe₃O₄ which confirms that CA-MNP are hydrophilic in the nature. These additional peaks confirm the coating of citric acid on the NPs.The crystalline structure of the synthesized MNP was confirmed using the XRD Fig. 1c represents the XRD pattern recorded for citric acid functionalized MNP. The presence of sharp and intense peaks confirmed the formation of highly crystalline nanoparticles. The peaks present at 2θ = 30.1°, 35.5°, 43.3°, 53.4°, 57.3° and 62.7° corresponding to (220), (311), (400), (422), (511), and (440) reflections of magnetite, respectively, can be clearly seen, indicating the presence of crystalline spinel structured magnetite-Fe₃O₄ phase of iron oxide²². The zeta potential of the as-synthesized MNPs was recorded as + 25.22 mV. In contrast, carboxylated MNP shows the surface charge of − 14.67mV (Fig. 1d).

Fig. 1

Physicochemical characterization of magnetite nanoparticles (MNP) and Carboxylate functionalized magnetite Nanoparticles (CA-MNP) (a) UV–Vis spectra of MNP and synthesized CA-MNP. (b) XRD Spectra MNP. (c) FTIR spectra of MNP and of carboxylate functionalized MNP and possible mechanism. (d) Zeta potential measurements MNP and CA-MNP.

The zeta-potential measurements show that adsorption of CA onto the surface of Fe₃O₄ nanoparticles resulted in a highly negative surface charge. The negative value of zeta potential for the CA-MNP further confirmed the presence of negatively charged carboxylate groups on the surface of MNP nanoparticles, allowing them to have good stability^23,24. Specifically, some of the carboxylate groups from citric acid coordinate to iron cations on MNP surface to form robust coating as shown in the Fig. 1b inset. While uncoordinated carboxylate groups extend into water medium conferring to higher degree of stability of MNP in water.

Fig. 2

Electron microscopy of MNP and CA-MNP (a). FESEM of CA-MNP with inset of MNP only (b). TEM image of CA-MNP.

The dimensions and morphology of the synthesized nanoparticles were studied by FESEM and TEM. As shown in Fig. 2a,b. The synthesized MNPs are spherical in nature and are monodispersed. The agglomeration has been observed in the uncoated MNP because of the magnetization involved as shown in Fig. 21 inset. Meanwhile, the carboxylate-coated MNPs are uniformly dispersed and separated from each other, as shown in Fig. 2a. confirming the typical characteristics of citric acid and Carboxylate coating. The citric acid-functionalized NPs were further characterized using TEM and the average size was 219 ± 3 nm with spherical morphology (Fig. 2b). From all the physicochemical characterization, it was confirmed the synthesized carboxylate-coated Fe₃O₄ shows good stability and spherical morphology making them suitable for environmental remediation applications.

Batch sorption studies

Effect of amount of CA-MNP taken for equilibration

To optimize the amount of CA-MNP to be used for extraction, batch sorption study was carried out by varying the amount of CA-MNP taken for equilibration with the feed solution (solutions at pH 4 spiked with Eu³⁺/Am³⁺ tracer). Figure 3a depicts the effect of amount of CA-MNP taken on the extraction efficiency of Am³⁺ and Eu³⁺. Initially, the experiment carried out with 0.5 mg of CA-MNP indicated 34.5% uptake of Am³⁺ and 60.3% uptake of Eu³⁺. Higher uptake of Eu³⁺compared to Am³⁺ is previously reported and attributed to preferential binding of Eu³⁺ with oxygen donor ligands via hard-hard acid base interaction compared to Am³⁺²⁵. Considering the appreciably higher uptake of Eu³⁺ at 0.5 mg of CA-MNP, uptake of Eu³⁺ was studied at lower amount of CA-MNP (0.1 mg and 0.25 mg). The observation indicated 13.9% and 23.07% uptake of Eu³⁺. The uptake of Eu³⁺ was observed to attain a maximum value when 2.5 mg of CA-MNP was taken for equilibration. For higher amount of CA-MNP taken, no significant increase in uptake efficiency was observed for Eu³⁺. Uptake of Am³⁺ was observed to be considerably lower compared Eu³⁺ for all experiments. For this reason, the extraction of Am³⁺ was studied at higher CA-MNP amount (3.0 mg and 3.5 mg). The observations indicated only a slight increase in the uptake efficiency compared to that at 2.5 mg CA-MNP (Fig. 3a). Considering the above observations, further studies were carried out with 2.5 mg of CA-MNP which corresponds to adsorption of 60.9% and 76.7% of Am³⁺ and Eu³⁺ respectively.

Fig. 3

(a) Effect of amount of CA-MNP taken for equilibration; (b) Effect of pH on extraction of Am³⁺ and Eu³⁺.

Effect of aqueous phase pH

The extraction of Am³⁺ and Eu³⁺ from aqueous feed solution was observed at varying initial pH of feed solution ranging from 2 to 6. The observations are presented in Fig. 3b. Uptake of Eu³⁺ as well as Am³⁺ was observed to be increasing with increase in initial pH of the feed solution. Lower uptake at low pH values can be attributed to low dissociation or protonation of citric acid at lower pH resulting in lower interaction with metal ion while at higher pH values, complete dissociation of the citric acid takes place resulting in more stable interaction with the metal ion which in turn increases the uptake. However, Am³⁺ and Eu³⁺ are susceptible to hydrolysis beyond pH = 4. In previous studies, decrease in uptake of Am³⁺ and Eu³⁺ on functionalized carbon nanotube composite was reported beyond pH 4 due to hydrolysis resulting in adsorption of hydrated metal ion on the walls of equilibration tube²⁶. Similar observations were also reported by Bardovskyi et al. for adsorption of Gd³⁺ onto diglycolamic acid functionalized silica surfaces²⁷. Gujar et al. also reported decrease in uptake of Am³⁺ and Eu³⁺ by the GO-PES beads beyond pH 4²⁸. In view of the above reported observations, further studies were carried out at pH = 4 to avoid loss of metal ion due to hydrolysis, if any. For trivalent americium, more than 90% exists as Am³⁺ upto pH 5, followed by a drastic reduction in the relative concentration. Almost negligible amount of americium exists as Am³⁺ in a feed of pH ~ 8. In nitrate based medium there would be a steady concentration of ~ 8% for Am(NO₃)²⁺ upto a feed of pH ~ 6 followed a gradual reduction in relative composition. From pH ~ 4.5, there was a steady increase of Am(OH)²⁺ following a Gaussian distribution pattern with peak in the range of pH 7-7.5. Similar Gaussian distribution in relative concentration of Am(OH)²⁺ was started from pH ~ 6 with a maxima at pH ~ 8. In view of this, it can be inferred that, there was a steady concentration of Am³⁺ ~92% in the feed acidity of pH ~ 2–5, where the extraction of Am³⁺ was performed. Almost ~ 8% of Am(NO₃)²⁺ was also present alongwith majority of Am³⁺. In the feed acidity range of pH ~ 5–6, Am(OH)²⁺ started building up. Therefore, the sequestration processes were purely dominated by Am³⁺ or Eu³⁺.

Effect of equilibration time

The kinetics of uptake of metal ions on the CA-MNP was studied by equilibrating the CA-MNP with the feed solution at pH = 4 spiked with Eu³⁺ /Am³⁺ tracer for varying time intervals. The observations for effect of equilibration time on uptake of Eu³⁺ and Am³⁺ is presented in Fig. 4a. From the figure, it can be observed that the extraction of Eu³⁺ by the CA-MNP is much faster as compared to the Am³⁺. Extraction of Eu³⁺ attains the equilibrium value within 15 min. while extraction of Am³⁺ requires about 120 min to attain the equilibrium. A study reported by Mengxue Li et al. reported sorption of Eu on nano magnetite attaining equilibrium in more than 8 h²⁹. Another study by Ping Li et al.³⁰, indicates sorption of Eu³⁺ on nano-iron oxide attains equilibrium in about 10 h. A much faster kinetics is observed in present study for uptake of trivalent metal ions. Considering the results, the faster uptake of metal ions by CA-MNP in present study can be attributed to binding of metal with carboxylate ions present on the surface of CA-MNP. The sharp contrast in the present study with the earlier reported results also indicates advantage of surface functionalization of the magnetic nanoparticles.

To get more insights into the sorption kinetics, the data was fitted into various sorption models such as Lagergren (pseudo first-order), intra-particle diffusion and pseudo-second-order kinetics models. The Lagergren’s first-order kinetics is represented by the following equation^31,32,33.

$$\log (q_{e} - q_{t} ) = \log q_{e} - \frac{{kt}}{{2.303}}$$

(1)

where, q_e and q_t are amount of radionuclide adsorbed by the CA-MNP at equilibrium and at time ‘t’, respectively while k is the adsorption rate constant and t is the time in minutes. The regression coefficient for Lagergren’s first-order kinetics was observed to be 0.466 and 0.975 for Eu³⁺ and Am³⁺ (Fig. 4b). The values indicate poor fitting of the kinetics data into the Lagergren first-order kinetics model. Based on this observation, the data was fitted into the Webber Morris intraparticle diffusion model represented by the following equations^34,35

$$q_{t} = kt^{{0.5}} + c$$

(2)

where, q_t is the amount of radionuclide adsorbed onto the CA-MNP at the time ‘t’. The linear regression coefficients for Am³⁺and Eu³⁺ were found to be 0.281 and 0.855, respectively which indicated poorer fitting and hence, it was concluded that, the sorption of Am³⁺ and Eu³⁺, did not follow the Webber Morris Intra-particle diffusion model. However, the positive intercept indicates the multiple elementary processes were responsible for the overall sorption processes. Further, the kinetic data was fitted into the pseudo-second-order kinetics model presented by Eq. (3)^36,37.

$$\frac{t}{{q_{t} }} = \frac{1}{{kq_{e}^{2} }} + \frac{t}{{q_{e} }}$$

(3)

where, k is the pseudo-second-order rate constant. The plot of t/q_t v/s t gave straight lines with linear regression coefficient 0.986 and 0.999 for Am³⁺ and Eu³⁺, respectively. For both the radionuclides, the linear regression coefficients were found to be close to unity indicating the sorption of Am³⁺ and Eu³⁺ follows to the pseudo-second-order kinetic. The rate constants were evaluated as 4.76 × 10^− 9 mg g^− 1 min^− 1 and 2.45 × 10^− 7 mg g^− 1 min^− 1 for Am³⁺ and Eu³⁺, respectively.

Figure 4b–d are depicting the linear regression analyses for the kinetics data by different kinetics models.

Fig. 4

(a) Variation of extraction of Am³⁺ and Eu³⁺ by CA-MNP with time and The linear regression analyses for different kinetics models (b) Lagergren or pseudo-first order; (c) pseudo-second-order kinetics model; (d) Intra-particle diffusion model.

Sorption thermodynamics

Understanding the thermodynamics of the sorption processes is very important aspect and needs to be looked into to get insights on the uptake process and factors driving it. The value of the thermodynamic parameters such as Gibbs free energy (ΔG°), enthalpy (ΔH°) and entropy (ΔS°) can be calculated by the following three equations^38,39,

$$\ln K_{D} = \frac{{\Delta S^{ \circ } }}{R} - \frac{{\Delta H^{ \circ } }}{{RT}}$$

(4)

$$\Delta G = \Delta H - T\Delta S$$

(5)

$$\Delta G^{ \circ } = - RT\ln K_{D}$$

(6)

where, K_D can be considered as conditional equilibrium constant, R is universal gas constant (8.314 J mol^− 1 K^− 1) and T is the temperature in K. The plots of lnK_D vs. 1/T are presented in Fig. 5. The K_D values were observed to be increasing with increase in temperature indicating endothermic nature of sorption process in present investigation. ΔH values were calculated from the slopes of above mentioned plots were found to be 22.9 kJ mol^− 1 and 22.7 kJ mol^− 1 for Eu(III) and Am(III) ion uptakes, respectively. The positive enthalpy change further indicates endothermic nature of sorption process. The ΔS_Eu was evaluated from the intercept of the above mentioned plot and found to be 0.14 kJ mol^− 1 k^− 1, while that for Am, i.e. ΔS_Am was 0.13 kJ mol^− 1 k^− 1. Consequently, the ΔG values were estimated using Eq. (5) as − 15.8 kJ mol^− 1 and − 17.9 kJ mol^− 1 for Am³⁺ and Eu³⁺, respectively. The negative ΔG values revealed the spontaneity of the sorption processes. Further, the positive values of entropy change for both the metal ions indicate the process is entropy driven. Also, higher uptake of Eu³⁺ is supported by lower value of ΔG compared to that of Am³⁺.

Fig. 5

The Vant Hoff plots for Am³⁺ and Eu³⁺ sorption onto the CA-MNP.

Sorption isotherms

To get an insight into the sorption process, the uptake study was carried out by varying the relative amount of the CA-MNP and the metal ion (Am³⁺ and Eu³⁺). The experimental data was modeled using various isotherm models such as Langmuir, Freundlich and D-R isotherms. The observations are presented in Fig. 6a–f.

The linear form of the Langmuir sorption isotherm is given by the Eq. (7)^40,41. The isotherm postulates monolayer sorption of sorbate at specific sites on the sorbent surface -.

$$\frac{{C_{{eq}} }}{q} = \frac{1}{{bq_{{\max }} }} + \frac{{C_{{eq}} }}{{q_{{\max }} }}$$

(7)

where, C_eq is the equilibrium concentration of metal ion in the aqueous phase, q is the amount of metal ion taken up by the CA-MNP while q_max and b are the maximum amount of metal ion uptake at saturation and the sorption coefficient (L mg^− 1), respectively. The plot of C_eq/q vs. C_eq (Fig. 6a and d for Eu³⁺ and Am³⁺ respectively) showed poor fitting for the Langmuir model which indicated the sorption of the radionuclides onto the CA-MNP does not follow the monolayer sorption. Further, the data were fitted to the Freundlich isotherm for which the linear form is given as follows^42,43.

$$\log \left( {\frac{x}{m}} \right) = \log K_{f} + \frac{1}{n}\log C_{e}$$

(8)

Where, C_eq is the metal concentration in the solution at equilibrium while x/m represents the amount of metal ion sorbed per unit mass of the CA-MNP. K_f is a constant dependent on the sorption capacity while n is related to the metal ion sorption intensity. The results are presented in Fig. 6b,e for Eu³⁺ and Am³⁺respectivelywhich yielded a regression coefficient of 0.953 for Am³⁺ and 0.798 for Eu³⁺ indicating better fit of the data to the Freundlich adsorption isotherm compared to Langmuir model. The observation indicates possibility of multilayer sorption of the metal ions. The values of K_f(16.14 mg g^− 1 for Am³⁺ and 41.30 mg g^− 1 for Eu³⁺) and n (0.677 for Am³⁺ and 0.786 for Eu³⁺) calculated from the linear fitting of the data to the Freundlich model. The maximum sorption capacity was also determined by equilibrating the 2.5 mg mL^− 1 CA-MNP with 1500 mgL^− 1 Eu³⁺ solution spiked with ^152–154Eu tracer overnight (batch sorption study). The overnight sorption data indicated equilibrium value of 37.6 mg g^− 1. The values observed graphically from Freundlich model was in close agreement with the result observed for maximum sorption observed by batch study. The data were also fitted in Dubinin–Radushkevich (D–R) isotherm equation as given below^44,45.

$$\ln q = \ln X_{m} - \beta \varepsilon ^{2}$$

(9)

where X_m is the maximum sorption capacity, b is related to the mean sorption energy, and ε is the Polanyi potential, expressed by the following equation,

$$\varepsilon = RT\ln \left( {1 + \frac{1}{{C_{{eq}} }}} \right)$$

(10)

where, R is the gas constant and T is the absolute temperature. The D-R parameters were then calculated from the plot of ln q vs. ε². The β in the D–R equation is related to the mean free energy (kJ mol^− 1) of sorption (E) empirically by the following equation.

$$E = \frac{1}{{\sqrt { - 2\beta } }}$$

(11)

The sorption energy was calculated as 11.18 kJ mol^− 1 for Eu³⁺ and 9.13 kJ mol^− 1 for Am³⁺, indicating the uptake of both the radionuclides occurs through chemisorption.

Fig. 6

The linear regression analysis for (a) Langmuir, (b) Freundlich, and (c) D–R isotherm model for Eu^3+, and (d) Langmuir, (e) Freundlich, and (f) D–R isotherm model for Am³⁺.

Radiation stability

Since the CA-MNP material is being used to sorb Am and Eu, it is susceptible to damage due to radiation exposure. Such exposure may result in significant changes in physical and chemical properties of surface functional groups resulting in alteration of uptake capacity of the material with respect to studied radionuclides. To assess the stability of the CA-MNP beads, the beads were exposed to γ radiation using a ⁶⁰Co source (with a strength of 5 kGy/h), where Gy is the absorbed dose defined as the absorption of 1 J of radiation per kg of matter. A batch extraction study was performed with the exposed material and the results are summarized in Fig. 7. It can be seen from Fig. 7 that irradiation at 300 kGy results in significant decrease in uptake of both the radionuclides. Further irradiation to 500 kGy results in increased uptake of Am marginally while uptake of Eu³⁺ decreases. Interestingly, uptake of Eu³⁺ as well as Am³⁺ by irradiated CA-MNP to 1000 kGy dose was found to be higher compared to CA-MNP irradiated at lower radiation doses. Radiolytic degradation of citric acid results in formation of mono and di-carboxylic acids⁴⁶. Further, due to degradation of citric acid coating, the iron oxide surface itself becomes available for sorption of radionuclides. Iron oxides/hydroxides have been reported as scavengers of the metal ions including trivalent lanthanides and actinides from aqueous medium^47,48. Hence higher uptake of Am³⁺ and Eu³⁺ by CA-MNP irradiated at 1000 kGy compared to CA-MNP irradiated at lower doses may be attributed to formation of degradation products of citric acid as well as availability of bare Fe₃O₄ surface for uptake of the radionuclides.

Fig. 7

Effect of radiation exposure on the extraction of Am³⁺and Eu³⁺ by the CA-MNP (Gy is the absorbed dose).

From the Table S-1 it can be observed that the concentration of Fe in the aqueous phase after equilibration with the CA-MNP is < 0.1 µg mL^− 1. Also, it can be seen from the table that the pH of the aqueous phase before and after equilibration was not altered significantly up to pH 4. The observations indicate that the CA-MNP are stable under experimental conditions.

Back extraction studies (stripping)

Stripping of loaded metal ions from the sorbent medium is an important aspect of pre-concentration studies. Efficient stripping/recovery of metal ion is required in order to reuse the sorbent resulting in reduction of waste generation, safe handling of process waste and is also important from economic considerations. A study was carried out to efficiently strip the adsorbed radionuclides (Am and Eu) from the CA-MNP. Various reagents, viz. 0.1 M EDTA (ethylenediaminetetraacetic acid; Merck, purity 99%), 0.1 M oxalic acid (SD Fine Chem.; purity 99%), 0.1 M Na₂CO₃ (SDFine Chem.; purity 99%), and 0.1 M HNO₃ were used as stripping reagents. These observations are shown in Fig. 8.

Fig. 8

Stripping of Am³⁺and Eu³⁺ from the CA-MNP (concentration of reagent = 0.1 M).

The stripping efficiencies of the reagents for stripping of loaded radionuclides was observed to be 10.32, 98.45, 84.56 and 59.19% for Na₂CO₃, EDTA, Oxalic acid and 0.1 M HNO₃ respectively for Eu³⁺ and 3.41, 41.97, 29.05 and 46.85 respectively for Am³⁺. The stripping of Eu³⁺ was observed to be higher compared to the Am³⁺. It can be seen from the Fig. 8 that > 99% of loaded Eu³⁺ is stripped in two stripping cycles while stripping of Am³⁺ requires more number of cycles to quantitatively strip the loaded radionuclide. Amongst all the reagent studied, 0.1 M EDTA is found to be the most suitable reagent for stripping. Similar results were observed for stripping of Am and Eu from the loaded multi walled carbon nanotubes where 0.1 M EDTA was found the most suitable stripping agent^26,49. The relative stability of the complex formed during the stripping of metal ions largely depends on structure and denticity of the ligands. Carbonate ligand coordinates in mono as well as bidentate fashion. Similarly, oxalic acid binds with Am³⁺/Eu³⁺ in bidentate manner forming a stable five membered chelate complex. However, EDTA can bind metal ions resulting in six membered chelate rings forming a very stable structure (four from the carboxylic groups and two from the amine group). The formation of a highly stable metal − chelate complex is responsible for the strong aqueous-phase complexation.

Evidencing of metal ion sorption on the substrate

Figure 9 shows the Energy-dispersive X-ray fluorescence (EDXRF) spectra of the pristine and Eu-loaded CA-MNP. The spectra of Eu-loaded CA-MNP show a peak at 5.73 keV, corresponding to Eu. Furthermore, a weak intensity peak at 6.1 keV was also observed in the Eu-loaded CA-MNP. The observations indicated the successful sorption of metal ions onto the substrate.

Fig. 9

Energy-dispersive X-ray fluorescence (EDXRF) spectra of pristine and Eu-loaded CA-MNP.

Experimental

Materials and methods

Synthesis of citric acid-functionalized magnetite nanoparticles (CA-MNP)

Ferric chloride (FeCl₃), Ferrous chloride (FeCl₂), Sodium hydroxide (NaOH), and Citric Acid (C₆H₈O₇) were purchased from LobaChemie India. All chemicals are of analytical grade and used as received. In a typical synthesis, 4 mM of FeCl₃ and 2.5 mM of FeCl₂ were dissolved in 80 ml of milli-Q water in the bottom flask, and the temperature was increased to 70 °C in refluxing condition under a nitrogenate atmosphere with constant stirring of 1000 rpm. The stirring and temperature were maintained at 70 °C for 1 h. Then 20 ml of 1 M NaOH was added dropwise to the reaction mixture and kept at the same temperature for 1 h. Then, 5 ml of an aqueous solution of citric acid(1mM) was added to the above reaction mixture, and the reaction temperature was slowly raised to 80 °C with continuous stirring. The black-colored precipitate was obtained by cooling the reaction mixture to room temperature, which was thoroughly washed with water. During this step, a permanent magnet was used to separate pure samples from the supernatant and unreacted material.

Physico-chemical characterization of the citric acid-functionalized -magnetite nanoparticles

Various analytical techniques were employed in the physicochemical characterization of Citric acid functionalized nanoparticles (CA-MNP). UV–Vis spectroscopy, conducted using a Jasco V-750 Spectrophotometer, confirmed the formation and assessed the optical properties of CA-MNP. The absorption spectra were acquired within the 190–900 nm range with a 1 nm resolution. Fourier-transform infrared spectroscopy (FTIR) analysis, performed with a Shimadzu IR Affinity, aimed to determine the presence of functional groups on the surface of CA-MNP responsible for reduction and stabilization during synthesis. Zeta potential (ZP) measurements, utilizing Dynamic Light Scattering (DLS), provided insights into the surface charge distribution of the synthesized nanoparticles. X-ray diffraction (XRD) analysis, using a Rigaku Miniflex, investigated the crystalline structure of MNP. Additionally, Scanning Electron Microscopy was performed on Nova NanoSEM NEPEP303, and Transmission Electron Microscopy (TEM) analysis, carried out on a TECNAI G2 SPIRIT BIOTWIN, allowed for the examination of morphology and mean particle size.

Radiotracers

²⁴¹Am (60 keV) and ^152–154Eu (121. 8 keV, 244.7 keV and 344.3 keV) tracers were obtained from Laboratory stock and used after checking its radiochemical purity by gamma spectrometry using a NaI (Tl) detector (Para Electronics) coupled to a multi-channel analyzer (ECIL, Mumbai).

Stability of the CA-MNP

Batch studies were carried out to evaluate the stability of the CA-MNP under the experimental conditions. For this purpose, the CA-MNP (2.5 mg mL^− 1) was equilibrated with the feed solution devoid of metal ion at different pH from pH 1–4. The stability of the CA-MNP was evaluated by estimation of Fe leached out in the aqueous phase using ICP-OES (HORIBA: ULTIMA-2). Also, pH of the aqueous phase before and after equilibration with CA-MNP was determined. The observations are presented in supporting information.

Batch sorption studies

Uptake of Eu³⁺ and Am³⁺ ions onto Citric acid-functionalized -magnetite nanoparticles (CA-MNP) was studied by batch sorption method. CA-MNP were suspended in water (10 mg mL^− 1) and a known amount of the suspension was taken for equilibration with the feed solution containing the radiotracer at a particular pH in leak-tight stoppered Pyrex tubes. Experiments were carried out in a thermostat (R-square, Scientific automation Company, Thane) maintaining the temperature at 25 ± 0.1 °C. After equilibration, the equilibration tubes were kept in contact with a strong magnet to separate the magnetite Nanoparticles from the aqueous phase after which a suitable aliquot was taken for the estimation of the metal ion. Leftover radioactivity in the solution was estimated by gamma-ray spectrometry. The uptake efficiency of CA-MNP was determined using the concentration of metal ion (Eu³⁺/Am³⁺) into the solution before and after equilibration by Eq. (12).

$$\% \,Extration = 100 - \left( {\frac{C}{{C_{0} }} \times 100} \right),$$

(12)

where, C is the concentration of metal ion after equilibration; C₀ is the concentration of metal ion before equilibration. The reliability of the method was tested by duplicate measurements and the standard deviation between the measurement was found to be < 5%. Batch sorption experiments in the same manner were carried out separately to study the effect of amount of CA-MNP taken for equilibration, pH of the solution and effect of temperature on the uptake of metal ion onto the substrate and sorption kinetics.

Evidencing of metal ion sorption on the substrate

To confirm the sorption of metal ion onto the CA-MNP was the CA-MNP was equilibrated with the 0.1 M Eu³⁺ solution (Eu(NO₃)₃dissolved in water) at pH 4 followed by separation of the solid and aqueous phase by magnetic separation. The supernatant was discarded and the residue was washed with pH 4 solution to remove any unadsorbed metal ion in the equilibration tube. The EDXRF spectrum of the loaded CA-MNP was performed on Jordan Valley EX3600M spectrometer, Israelequipped with Rh X-ray source and Si(Li) detector for evidencing the metal ion sorption on the substrate.

Back extraction studies (stripping)

A study was carried out to strip the adsorbed radionuclides (Am and Eu) from the CA-MNP. Different reagents, viz. 0.1 M EDTA (ethylenediaminetetraacetic acid; Merck, purity 99%), 0.1 M oxalic acid (SD Fine Chem.; purity 99%), 0.1 M Na₂CO₃ (SDFine Chem.; purity 99%), and 0.1 M HNO₃ were used as stripping reagents. For stripping study, the CA-MNP was first loaded with the radionuclide (Am/Eu) as described in the batch sorption studies at pH 4 followed by separation of the solid and aqueous phase by magnetic separation. The supernatant was discarded and the residue was washed with pH 4 solution to remove any unadsorbed metal ion in the equilibration tube. Complete removal of unadsorbed metal ion was confirmed by gamma ray spectrometry indicating insignificant counts in the washing. The residue is then equilibrated with the stripping solution followed by separation of phases by magnetic separation and evaluation of stripped radioactivity by gamma spectrometry.

Selectivity study

Selectivity of the CA-MNP towards uptake of desired metal ion is studied by equilibrating the CA-MNP (2.5 mg mL^− 1) with the solution containing mixture of metal ions (Na⁺, K⁺, Ca²⁺, Mg²⁺, Mn²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺ and Ce³⁺) along with Eu at pH 4. After equilibration, the concentration of metal ions was determined by ICP-OES (HORIBA: ULTIMA-2). The detailed observations are presented in supporting information. Selectivity of the CA-MNP towards uptake of desired metal ion is studied by equilibrating the CA-MNP (2.5 mg mL^− 1) with the solution containing mixture of metal ions (Na⁺, K⁺, Ca²⁺, Mg²⁺, Mn²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺ and Ce³⁺) along with Eu at pH 4. After equilibration, the concentration of metal ions was determined by ICP-OES (HORIBA: ULTIMA-2). The observations are presented in Fig. S-1. It was found that the uptake of other metal ions (Mg, Mn, Co, Ni, Cu, Zn and Ce) by CA-MNP was comparable to Eu. However, even in the presence of the competing ions, more than 50% Eu present initially in the solution was taken up by the CA-MNP. In view of presence of multiple carboxylic groups present along with the iron oxide base, poor selectivity of the material is anticipated. Functionalization of sorbent with more selective ligands may result in enhanced selectivity towards the desired radionuclide. The present study was intended to indicate the feasibility of using the functionalized magnetic nanoparticles for uptake of radionuclides from aqueous medium. Findings of the present study will be helpful for us in the development of magnetic nanomaterials with better selectivity.

Conclusions

Citric acid-functionalized magnetic nanoparticles (CA-MNP) were prepared and characterized using various analytical technique. Formation of Fe₃O₄ nanoparticles was confirmed by FTIR indicating characteristic band of Fe-O reobserved at 539 cm^− 1. Similarly coating of citric acid over the magnetic nanoparticles was confirmed by FTIR indicating shift in Fe-O stretching band from 539 to 564 cm^− 1, the C = O str originating from the free carboxylic bond at 1589 cm^− 1 and the C = O str originating from the coordinated carboxylic bond at 1365 cm^− 1 peaks. Further, the negative value of zeta potential for the CA-MNPalso confirmed the presence of negatively charged carboxylate groups on the surface of MNP nanoparticles.

Application of CA-MNP for uptake of trivalent lanthanide (Eu³⁺) and actinide (Am³⁺) was studied in detail. 2.5 mg mL^− 1 was found as optimum concentration of CA-MNP for sorption of radionuclides studied corresponding to adsorption of 60.9% and 76.7% of Am³⁺ and Eu³⁺ respectively at pH 4. While previous studies indicated higher time for equilibrium sorption of Eu³⁺, CA-MNP attains equilibrium sorption of Eu³⁺ within 15 min. However, extraction of Am³⁺ requires about 120 min. to attain the equilibrium. Further, modeling of kinetic data indicated pseudo-second-order kinetics of sorption for both the radionuclides. Thermodynamic study indicated ΔG values as -15.8 kJ mol^− 1 and − 17.9 kJ mol^− 1 for Am³⁺ and Eu³⁺, respectively. The negative ΔG values revealed the spontaneity of the sorption processes. Further, the positive values of entropy change for both the metal ions indicate the process is entropy driven. The sorption of trivalent metal ion was confirmed by appearance of Eu³⁺ characteristic peak at 5.7 keV in EDXRF spectrum as well as change in the particle size observed by DLS data. Irradiation with gamma radiation indicated significant loss of efficiency starting from 300 kGy which increases with radiation dose. Interestingly, the uptake efficiency increases after radiation dose of 1000 kGy. This may be attributed to formation of degradation products of citric acid and availability of bare iron oxide particles due to degradation of citric acid. Stripping with different reagents viz. 0.1 M EDTA, 0.1 M oxalic acid, 0.1 M Na₂CO₃, and 0.1 M HNO₃ indicated best results with 0.1 M EDTA. >99% of loaded Eu³⁺ is stripped in two stripping cycles while stripping of Am³⁺ requires more number of cycles to strip the loaded radionuclide quantitatively. The study indicated potential use of Fe₃O₄ magnetic nanodots for efficient sorption/preconcentration of trivalent f-block elements from aqueous solutions.