Synthesis and Adsorption Properties of Gadolinium-Imprinted Divinylbenzene-Based Copolymers.

Ion-imprinted divinylbenzene and methacrylic acid copolymers for rare-earth element adsorption with crosslink ratios 60 and 40% have been synthesized. Ion imprinting has been carried out via the trapping approach. Alizarin red S has been incorporated in the polymers as a nonvinylated ligand. The obtained materials were characterized via scanning electron microscopy and Fourier transform infrared spectroscopy. Synthesized polymers exist as gel-type nonporous spherical agglomerates from 2 to 6 μm with low surface areas. The sorption properties of the synthesized polymers with respect to lanthanides in static and dynamic modes have been studied. The most efficiently synthesized materials extract rare earth elements from solutions at pH 4-7. The maximum sorption capacity of the obtained polymers with regard to Gd is around 0.5 mmol/g. According to the obtained results, the imprinted polymer with a crosslink ratio 60% is characterized by the highest values of distribution coefficients for lanthanides. The total lanthanide breakthrough capacity for this polymer is 0.861 μmol/g. The adsorption column has been constructed using the imprinted polymer with a crosslink ratio 60%. The column extraction and preconcentration procedure for the trace Eu content determination in the strontium iodide recycled material by inductively coupled plasma-atomic emission spectrometry at a level of 1 × 10-6%(mass) has been proposed.

Introduction

From the middle of the 20th century, a new class of synthetic materials obtained by molecular imprinting has attracted the broad interest from scientists all over the world.[1,2] The inherent molecular recognition properties of molecular imprinted polymers (MIPs) are similar to natural receptors such as antibodies. However, unlike biological receptors, MIPs are easier and cheaper to prepare; with high physical, chemical, and thermal stability they are reusable and reproducible.[3] Similarly, the idea to research the phenomenon of ion imprinting by replacing the template molecule with a metal ion arose.[4] The most advanced application of ion-imprinted polymers (IIPs) is using them as sorbents for SPE.[5] IIPs can also be integrated into sensors[6] and employed as stationary phases in chromatographic columns.[7]

Previous investigations have shown that ion-imprinted ethylene glycol dimethacrylate and methacrylic acid (MAA) copolymers have large distribution coefficients but are not always characterized by a significant imprinting effect.[8] This issue may be attributed to the competitive interaction between metal ion and crosslinker moieties in the polymer matrix. According to this, a task to obtain sorption materials with the matrix that does not contain extra functional groups capable of nonselective metal ion binding appeared. For this purpose, divinylbenzene (DVB) has been chosen as a crosslinking agent because this compound evidently do not complexate with metal ions unlike ethylene glycol dimethacrylate.

In recent years, doped halides of alkali and alkaline earth metals have been intensely investigated as scintillators for radiation detection.[9] In most cases, the host crystals are doped with Ce3+, Pr3+, or Eu2+ ions because such materials exhibit high light yields and excellent energy resolution.[10] One of the currently most studied systems is SrI2:Eu single crystals.[11] In order to obtain stable reproducible crystals of high quality, it is necessary not only to eliminate impurities in raw materials,[12] but also to control the quantity of dopants at all stages of the production. An important concept for efficient production of halide-based scintillators is to recycle waste produced during mechanical processing and packaging into raw materials for the next batch. Typically, the dopant content is about 1%,[13] and there are no difficulties in its direct determination by analytical methods, but after purification, a demand for trace Eu content determination appears. Such task is nontrivial for most of the instrumental methods of analysis, and some preliminary sample preparation is required. Sorption preconcentration appears to be the easiest way to achieve the necessary limits of detection.

As has been noted, IIPs are excellent materials for such application. Our earlier studies represent the fact that imprinted polymers with the addition of Alizarin red S (ARS) can effectively remove small amounts of rare earths.[8] However, the use of IIPs for accurate determination of low concentrations of the target ion carries a certain risk of obtaining overestimated results. Solution contamination may occur due to the impossibility of 100% template removal from the polymer after synthesis, and therefore, its leakage from the polymer at the elution step is likely to happen. It can be avoided by using as a template, the ion that is very close in size and properties to the target ion. Thus, Gd-imprinted polymers may serve as the appropriate sorbents for Eu removal and preconcentration.

In summary, the objectives of this study were to develop novel ion-imprinted materials, to investigate their adsorption properties and ascertain the optimal sorption conditions, to develop a procedure of Eu column preconcentration in the SrI2 raw material prior to inductively coupled plasma–atomic emission spectrometry (ICP–AES) analysis.

Results and Discussion

Characterization Studies

Fourier Transform Infrared Spectroscopy

Figures and 2 depict the Fourier transform infrared (FT-IR) spectral analysis of the obtained polymers. The bands observed near 3050 and 3020 cm–1 in the polymers (Figures and 2) correspond to the C–H stretch in the aromatic ring and the bands near 2960, 2920, and 2850 cm–1—to aliphatic C–H stretch. It should be noted that the absorption bands of aliphatic C–H are more intense in the spectra of polymers than in the spectrum of DVB (Figure S1). On the other hand, the band around 3080 cm–1 which was assigned to =CH2 stretch is much more intense in the spectrum of DVB. A similar situation is observed for the band near 1630 cm–1 which indicates alkene C=C double bond stretching vibrations. Also in polymers’ spectra, appearance of the band near 1370 cm–1 can be observed. This band corresponds to symmetrical C–H bend in the methyl group which is obviously absent in DVB and styrene (Sty) (Figure S1 and S2). This proves high monomer conversion and confirms successful polymerization.

Figure 1

FT-IR spectra of polymers with crosslink ratio 60%: (a) 60-I, (b) 60-B.

Figure 2

FT-IR spectra of polymers with crosslink ratio 40%: (a) 40-I, (b) 40-B.

The bands specific to aromatic ring C=C stretch are observed in the range 1600–1400 cm–1 in all spectra. The band near 990 cm–1 corresponds to the aromatic ring C–H in-plane bend, while the group of bands in the range 900–700 cm–1 corresponds to out-of-plane vibrations of this bond. As can be seen, an additional band appears in this range in the spectra of polymers with a crosslink ratio 40%. This is due to the presence of monosubstituted aromatic rings (from Sty) in these polymers whereas only disubstituted aromatic rings are present in polymers with a crosslink ratio 60%.

Specific Surface Area by N2 Adsorption

The Brunauer–Emmett–Teller (BET) surface areas of obtained polymers presented in Table indicate gel-type resins formation.[14] Such materials typically have very low specific surface areas in the dry state because the polymer chains are in close molecular contact.

Table 1

Specific Surface Areas of the Synthesized Copolymers

polymer code	60-I	60-B	40-I	40-B
S, m2/g	20.6	4.4	17.0	4.1

Electron Microscopy

The respective scanning electron microscopy (SEM) micrographs are shown in Figure illustrating that the polymers are formed by spherical particles agglomerates with 2 ÷ 6 μm diameter. The particles have a smooth dense surface and are tightly fused with each other. No macropores are visible in the polymers’ structure. Such a morphology is common for gel-type polymers with a high crosslink ratio.

Figure 3

Micrographs of synthesized polymers: (a) 60-I, (b) 60-B, (c) 40-I, (d) 40-B.

When the crosslink ratio is high, the type of polymer primarily depends on the porogenic solvent amount and compatibility with polymer chains (if it is thermodynamically “good” or “bad”). If the porogen is “bad”, its phase separation occurs at the early stages of polymerization, when the polymer conversion is low and the system is saturated with unreacted monomers. So further growing polymer chains interconnect present particles and fill pores in them which leads to gel-type polymer formation.[14] Since the Hildebrand solubility parameters of dimethyl sulfoxide (DMSO) and Sty-co-DVB are 29.7 and 17.8 MPa1/2 respectively,[15] it can be assumed that gel-type polymer formation occurred due to DMSO being a “bad” solvent for this system.

These observations agree with the work of Santora et al.,[16] where polymers with a dramatically different morphology are formed depending on the porogen used.

Adsorption Studies

Effect of pH

As can be seen on Figure , the gadolinium adsorption onto obtained particles is strongly dependent on pH. The recovery increases with increasing pH and reaches the plateau after pH 4.0. It can be explained by the increasing of the negative charge density in response to the gradual surface functional groups deprotonation. Thus, in all subsequent experiments the pH was adjusted to 6.0.

Figure 4

Effect of the pH on Gd(III) binding. Gd(III) initial concentration: 10 mg/L; contact time: 1 h.

Adsorption Isotherms

The plots of q versus ce are depicted in Figure . To evaluate binding properties of the polymers, Freundlich and Langmuir isotherm models have been used. The values of Freundlich and Langmuir equation constants are shown in Table .

Figure 5

Adsorption isotherms of synthesized polymers. Gd(III) initial concentration: 0.064–1.0 mmol/L; pH 6.0; contact time: 1 h; t = 25 °C.

Table 2

Freundlich and Langmuir Equations Constants

	Freundlich isotherm			Langmuir isotherm
polymer	KF, mmol/g	n	R2	KL, L/mmol	qmax, mmol/g	R2
60-I	0.48	2.4	0.869	5.3	0.53	0.901
60-B	0.48	2.5	0.890	6.3	0.50	0.884
40-I	0.50	2.6	0.930	6.2	0.52	0.910
40-B	0.47	2.3	0.926	5.1	0.51	0.927

As can be seen, the correlation coefficients for two models do not differ much from each other. It is an often observed situation when different adsorption models result in an equally good fit to the same experimental data for imprinted polymers because of their inherent binding site heterogeneity.[17] However, the Freundlich equation better fits starting segments of isotherms. After analyzing Freundlich constants, it can be concluded that synthesized polymers have similar heterogeneity coefficients and show approximately identical affinity to the adsorbate.

According to the Giles classification,[18] the obtained isotherms are L-type. Isotherms of this form describe when activation energy for the removal of solute from the adsorbent surface does not depend on other adsorbate particle presence on that surface; that is, the interaction between them upon progressive saturation of the solid is insignificant.

Adsorption Kinetics

The influence of contact time on Gd(III) uptake by the polymer 60-I is shown in Figure . The data obtained revealed that the adsorption rate of Gd(III) onto a synthesized material is rapid and equilibrium is reached within approximately 20 min. Unfortunately, it is not possible to assess any kinetic mechanism reliably due to the lack of data in a short-time region.

Figure 6

Influence of the contact time on the adsorption of Gd(III) by polymer 60-I. Gd(III) initial concentration: 0.89 mmol/L; pH 6.0; contact time: 5–120 min.

Competitive Adsorption of Lanthanides

Distribution coefficient versus atomic number plots of lanthanides are shown on Figure . The observed nonmonotonous patterns indicate the so-called tetrad effect occurrence in rare earth adsorption by obtained polymers. The tetrad effect is the phenomenon when convex or concave curves consisting of La–Ce–Pr–Nd, (Pm)–Sm–Eu–Gd, Gd–Tb–Dy–Ho, and Er–Tm–Yb–Lu (tetrads) appear in physicochemical property dependance on the atomic number. It was first discovered in a system of solvent extraction.[19] The tetrad effect is also often observed in rare earth element geochemistry[20] and, of course, in adsorption studies.[21]

Figure 7

Distribution coefficient dependence on the lanthanide atomic number. Lanthanides initial concentration: 5 mg/L each; pH 6.0; contact time: 1 h.

The effect is explained in terms of Racah’s parameters of interelectronic repulsion changing during the formation of coordination bonds.[22] The tetrad effect confirms the chemical nature of lanthanide adsorption on the obtained materials due to the 4f-electron contribution for bonding and inner-shell complex formation between lanthanides and surface functional groups of polymers.

Also, it can be seen that for the pair of polymers with crosslink ratio 60% (60-I and 60-B), the imprinting effect is evident in contrast to the other pair with crosslink ratio 40% (40-I and 40-B) because the distribution coefficient pattern for imprinted polymer 60-I goes higher than for blank 60-B. It leads to the conclusion that despite the gel nature of the obtained resins which means their high swellability imprinted polymers with crosslink ratio 60%, they have the ability to retain imprinted cavities. This is understandable since water is a “bad” solvent for Sty–DVB copolymers, so they do not swell in working solutions and consequently do not change their structure after the contact with the solvent. But evidently, cavities in the polymer with a lower crosslink ratio are not rigid enough to provide a proper template recognition. Thus, only the material with a higher crosslink ratio fits the imprinted sorbent criteria.

The obtained sorbent is characterized by satisfactory sorption capacity and high distribution coefficients. Such properties suggest that complete extraction of lanthanides at small concentrations by this material can be easily achieved. It gives the grounds for applying polymer 60-I for trace europium preconcentration in dynamic conditions in order to control the quality of cleaning the recycled alkaline-earth metal halides which are to be used for scintillator production.

Dynamic Adsorption

First of all, the breakthrough capacities of lanthanides were determined. The values obtained are listed in Table .

Table 3

Breakthrough Capacities of Lanthanides

element	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Ho	Er	Tm	Yb	Lu
qD, 10–2 μmol/g	3.7	6.0	6.7	5.9	8.8	6.3	6.3	6.1	5.6	5.3	6.1	6.3	6.1	6.9

Thus, the total breakthrough capacity equals 0.861 μmol/g. This corresponds to 0.13 mg/g of europium which is sufficient for its extraction at trace levels.

Further, the lanthanide desorption from the column has been studied. The results are presented in Table .

Table 4

Lanthanide Desorption from the Column at Different Eluent Concentrations

element		La	Ce	Pr	Nd	Sm	Eu	Gd
r, %	c(HCl) = 0.24 mol/L	90	91	82	85	85	87	83
	c(HCl) = 1.2 mol/L	93	96	98	98	99	98	97

From the values above, it can be seen that the adsorbed ions are more efficiently eluted by more concentrated acid.

Trace Eu Determination Procedure in the Raw Material for SrI2 Single Crystal Production

Since SrI2 is doped with Eu2+, it is necessary to oxidize it prior to extraction on a synthesized polymer. Furthermore, the iodide ion must be removed from the solution because as shown by spike/recovery experiments (Table ), it causes a negative effect on Eu adsorption in the column—europium was not detected in the eluates. Such an effect may be explained by the formation of europium iodide complexes[23] that cannot dissociate due to the large excess of this anion which is why Eu cannot be bound by the sorbent. The spike/recovery assay has been carried out as follows: 100 mL volume of 100 g/L SrI2 solution spiked with 0.001 mg/L of europium (which corresponds to the Eu 1 × 10–6% content in SrI2) with pH 6 was passed through a column at a flow rate 0.5 mL/min. Then, 10 mL of deionized water was passed through the column at the same speed to wash the matrix solution. For europium elution, 5 mL of 1.2 mol/L HCl was allowed to pass through the column and the eluate was collected in a 5 mL volumetric flask.

Table 5

Spike/Recovery Assay Results for I– Medium

	Eu recovered, 10–6%
Eu spiked, 10–6%	x1	x2	x3
	<0.1	<0.1	<0.1
1.0	<0.1	<0.1	<0.1

Because standard reduction potentials of the couples NO3–/N2O4 and NO3–/NO are +0.80 and +0.96 V, respectively, while E°(I2/I–) = +0.54 V and E°(Eu3+/Eu2+) = −0.36 V,[24] one can conclude that HNO3 has a thermodynamic tendency to oxidize both I– and Eu2+. So the most convenient method for iodide removal and Eu oxidizing seems to be the reaction with HNO3. To test Eu removal from nitrate medium, the spike/recovery procedure was performed as previously described but the solution of SrI2 has been treated with hot nitric acid. The results of spike/recovery assay are shown in Table . As can be seen, the recovery rate of 95% was obtained.

Table 6

Spike/Recovery Assay Results for NO3– Medium

	Eu recovered, 10–6%
Eu spiked, 10–6%	x1	x2	x3	x̅	SD, 10–6%	, 10–6%
	<0.1	<0.1	<0.1
1.0	0.95	0.925	0.97	0.95	0.023	0.95 ± 0.06

Therefore, the following procedure for Eu determination in SrI2 at levels close to 1 × 10–6% was proposed. 10 g of SrI2 was transferred to a chemical beaker; then, 10 mL of deionized water and 8 mL of concentrated nitric acid were added. The mixture was carefully heated until iodine vapors are completely removed from the solution. Then, the sample was left to cool at room temperature, and another 75 mL of water was added. The acidity of the mixture was adjusted to pH 6 by the addition of NaOH solution. After this, the resulting solution was transferred to a 100 mL volumetric flask and then passed through the column at 0.5 mL/min flow rate. Then, the column was washed with 10 mL of deionized water, and finally, 5 mL of 1.2 mol/L HCl solution at the same rate was passed through the column. The eluate was collected in a 5 mL volumetric flask. The Eu concentration in the obtained solution was measured by ICP–AES.

The procedure developed allows improving the limit of detection (LOD) of Eu in SrI2 raw materials in comparison with conventional methods (Table ).

Table 7

Eu LODs in SrI2 Raw Materials for Different Established Methods

method of analysis	LOD, %
ICP–AES	1 × 10–5
column preconcentration prior to ICP–AES (proposed procedure)	1 × 10–7
stripping voltammetry	1 × 10–4
flame atomic absorption spectrometry	2 × 10–3

Moreover, the europium detection limit in the procedure can be easily reduced even more by scaling-up the column as well as by increasing the amount of strontium iodide solution that is passed through the column.

Conclusions

The method of DVB and MAA copolymerization with the addition of ARS in the presence of Gd3+ ions allows obtaining IIPs for the effective separation of lanthanides. The optimal acidity range for lanthanide sorption by synthesized materials is pH 4–7. Sorption isotherm starting segments are better described by the Freundlich equation, which indicates the heterogeneity of binding sites on the surface of polymers. The detected tetrad effect shows the chemical nature of the lanthanide sorption on the synthesized materials in the form of inner sphere complexes. The imprinting effect is achieved at a higher crosslink ratio which is consistent with the literary data. The total breakthrough capacity of lanthanides for the polymer 60-I is 0.861 μmol/g which corresponds to a europium breakthrough capacity of 0.13 mg/g. That value is sufficient for application of the synthesized material for trace europium column extraction. Adsorbed ions are better eluted from the column with more concentrated hydrochloric acid solution. Iodide ions adversely affect the adsorption of Eu in the column due to the formation of complexes, the dissociation of which is suppressed by the excess of I–. The nitric acid digestion step is needed for total iodide removal from SrI2 solution. The proposed procedure for Eu determination in the raw material of strontium iodide allows to reduce the LOD to 1 × 10–7% by ICP–AES.

Experimental Section

Materials and Reagents

MAA for synthesis, Merck KGaA; DVB for synthesis, Merck KGaA; Sty for synthesis, Merck KGaA; azobisisobutyronitrile (AIBN), UKRORGSYNTEZ Ltd., DMSO analytical grade; ARS reagent grade; Gd(NO3)3·5H2O reagent grade; NaOH Reag. Ph. Eur., Merck KGaA; hydrochloric acid Reag. Ph. Eur., Merck KGaA; and lanthanides solutions were prepared by dissolving their oxides, special purity grade, in nitric acid Reag. Ph. Eur., Merck KGaA.

Instrumentation and Equipment

The infrared spectra were recorded by the KBr pellet method using the FT-IR spectrometer Spectrum One, PerkinElmer, USA. Specific surface areas were determined by the BET method using the Kelvin 1042 sorptometer, Costech International, Italy. SEM (JSM-6390LV, JEOL, Japan) was used to study the surface morphologies of the polymers. The sample preparation has been carried out as follows: a small amount of polymer powder was dispersed in water; then, one drop of the suspension was placed onto a graphite substrate and dried. pH measurements were done using the STARTER 3100 pH-meter, OHAUS, USA. ICP–AES (iCAP 6300 Duo, Thermo Scientific, USA) was used to determine the concentrations of rare earths in solutions.

Synthesis of IIPs

Gd-imprinted polymers with different crosslink ratios were synthesized by the trapping approach. The crosslink ratios were calculated theoretically. The supplied DVB reagent contains about 60% (by weight) of the bifunctional monomer, and the other part of the mixture consists of ethylvinylbenzene (EVB) isomers. Therefore, the crosslink ratio of polymers obtained with only DVB reagent is 60% (by moles). The polymers were synthesized as follows: ARS (nonvinylated ligand), Gd(NO3)3·5H2O (ion template source), MAA (vinylated ligand), DVB reagent (crosslinker), Sty (for reducing the crosslink ratio), and AIBN (initiator) were dissolved in a small amount of DMSO (porogen). The reaction vessels were purged with argon for 30 min under ultrasonication and then sealed and placed in a water bath at 65 °C for 24 h. Upon completion of polymerization, the vessels were smashed, and the resulting polymers were ground in an agate mortar. Further, the powders were washed with a large amount of 0.1 M hydrochloric acid and distilled water by the decantation method to leach the template ion. After washing, the sorbents were dried in a vacuum oven for 8 h at 80 °C. Blank polymers were obtained by an analogous technique except that gadolinium nitrate was not added to the reaction mixture. They were also subjected to the same pretreatments as IIPs. The composition of the reaction mixtures is given in Table . The theoretical crosslink ratio of polymers was determined as the molar percentage of the crosslinker (DVB) in the mixture of monomers (DVB + EVB + Sty + MAA).

Table 8

Reaction Mixture Composition of the Synthesized Copolymers

polymer	Gd(NO3)3·5H2O, mmol	ARS, mmol	MAA, mmol	DVB/EVB, mmol	Sty, mmol	crosslink ratio, %
60-I	0.825	0.535	0.564	13.6/9.0		60
60-B		0.535	0.564	13.6/9.0		60
40-I	0.885	0.574	0.606	9.7/6.4	8.1	40
40-B		0.574	0.606	9.7/6.4	8.1	40

Adsorption Studies

The static sorption properties of polymers were studied on gadolinium ions because it served as the template during synthesis.

Effect of pH

The effect of pH on the Gd(III) extraction was studied by mixing 25 mg of the sorbents with 20 mL of the test solution with a Gd concentration of 10 mg/L in the pH range between 2.0 and 7.0. The solutions’ pH values were adjusted to desired values with NaOH or HCl solutions and monitored by a pH-meter. The obtained mixtures were then placed on an orbital shaker at 250 rpm for 1 h. Equilibrium gadolinium concentrations were determined after filtration by ICP–AES. Gadolinium recovery was calculated by the formulawhere c0 is Gd(III) initial concentration (mg/L); ce is Gd(III) equilibrium concentration in the solution (mg/L).

The adsorption properties of the IIPs were investigated by varying the initial Gd(III) concentration from 0.064 to 1.0 mmol/L at 25 °C and pH 6.0. For this, 20 mL of solution with defined Gd(III) concentration and adjusted pH 6 was added to 25 mg of the sorbent. Then, the mixtures were processed as previously described. The adsorption capacities were calculated by the formulawhere c0 is Gd(III) initial concentration (mmol/L); ce is Gd(III) equilibrium concentration in the solution (mmol/L); V is solution volume (L); and m is sorbent mass (g).

Freundlich isotherm equationwhere q is equilibrium adsorption capacity (mmol/g); KF is Freundlich constant (mmol/g); ce is Gd(III) equilibrium concentration in the solution (mmol/L); and n is heterogeneity coefficient.

Langmuir isotherm equationwhere q is equilibrium adsorption capacity (mmol/g); qmax is adsorptive monolayer capacity (mmol/g); KL is Langmuir adsorption equilibrium constant (L/mmol); and ce is Gd(III) equilibrium concentration in the solution (mmol/L).

Adsorption Kinetics

To evaluate the time needed for reaching adsorption equilibrium, the adsorption kinetics experiment was performed with the sorbent 60-I. The polymer portions of 25 mg were mixed on a shaker with 20 mL of 0.89 mmol/L Gd(III) solution with pH 6 at different time intervals ranging from 5 to 120 min. The adsorption capacities were calculated by the previous formula.

Competitive Adsorption

The experiment was carried out as follows: to 25 mg of polymer, 20 mL of lanthanide mixture solution with a concentration of 5 mg/L each and pH 6 was added. After shaking for 1 h and lanthanide equilibrium concentration determination, the distribution coefficients for the ions were calculated by the formulawhere c0 is lanthanide initial concentration (mg/L); ce is equilibrium lanthanide concentration in the solution (mg/L); V is solution volume, mL; m is sorbent mass, g.

To investigate the dynamic sorption properties, the fixed-bed column with an inner diameter of 5 mm and an adsorbent amount of 0.1 g was constructed. The column was connected to the peristaltic pump.

For the breakthrough capacity determination, a solution of lanthanide mixture with a concentration of 0.2 mg/L each and pH 6 was passed through the column with the flow rate 0.5 mL/min and small fractions of the solution coming out were collected. Then, the concentrations of lanthanides in these fractions were determined by ICP–AES. Breakthrough capacities were calculated by the formulawhere c0 is ion influent concentration, μmol/L; V is the volume of the solution passed through the column until the breakthrough point, L; and m is sorbent mass, g.

The breakthrough point comes when the ion effluent concentration to ion influent concentration ratio reaches 0.05.

For the elution studies, the column was loaded with the lanthanide mixture at the flow rate 0.5 mL/min. Then, they were eluted from the column with 3 mL of hydrochloric acid solution with the same speed. The lanthanide concentrations in eluates were measured by ICP–AES.