Introduction

The release of various harmful heavy metal ions owing to industrial and agricultural processes currently represents one of the main causes of pollution.1 Mercury is one of the most hazardous elements for human health because of its relative solubility in water and living tissues2 and its tendency to bioaccumulate in the human body, causing weakness, neurological damage, chromosomal mutation and so on.3 The World Health Organization recommends a limit of 1 μg l−1 of mercury in drinking water,4 which requires a highly accurate, selective and sensitive method of measurement. Common methods used for mercury extraction from water samples include liquid–liquid extraction,5, 6 solid–liquid extraction,7 flotation8 and membrane filtration.9 The immobilization of organic ligands on the surface of an inorganic or organic solid support is usually achieved by modifying the surface with certain target functional groups that can enhance the affinity towards metal ions. Different chelating sorbents for preconcentration of mercury were obtained using dithiocarbamate resin,10 N-(2-pyridyl methyl) chitison,11 2,3-dimercapto propane-1-sulfonate12 and dithizone derivatives.13 According to these data, 1,5- diphenylcarbazone (DPC), a well-known classical chelating ligand for mercury extraction, can be considered a potential chelating agent available to enhance the absorption performance of solid supports. Using the experience of our research group in the field of molecularly imprinted polymers,14, 15, 16, 17, 18, 19, 20, 21, 22 imprinted technology was used to prepare a new class of sorbent materials for mercury extraction, which were prepared in the presence and absence of DPC, with the aim of evaluating the influence of this compound on the absorption behavior of polymers. The general procedure of ion-imprinted polymer (IIP) preparation consists of the formation of a ligand–metal complex followed by copolymerization in the presence of an excess of a cross-linking agent. After polymerization, the template ion is removed using a washing procedure, leaving within the polymer network three-dimensional recognition cavities with a predetermined orientation according to their stereochemical interaction with the template metal ion.23, 24 Only a few studies using mercury as the target ion are reported in literature. For its simplicity, the trapping approach using a non-polymerizable ligand was the most commonly used.25, 26, 27 Singh et al.3 prepared a Hg(II) imprinted polymer via the formation of a binary complex of mercury with 4-(2-thiazolylazo) resorcinol followed by thermal copolymerization with methacrylic acid and ethylene glycol dimethacrylate (EGDMA). Recently, a voltammetric sensor for mercury detection using a new porphyrin derivative as a non-polymerizable ligand was prepared.26 Diazoaminobenzene was also chosen as a complexing agent by Liu et al.25 to prepare Hg-imprinted copolymers in the presence of 4-vinylpyridine (4VP) and EGDMA.

In this work, three Hg(II)-imprinted polymers were prepared using different synthesis approaches to understand the influence of DPC, which was used as a non-polymerizable ligand, on the absorption performance of polymers. The choice to test this compound arises from the well-known high affinity of DPC for mercury ions, which are able to form a stable blue–violet Hg(II)–DPC complex. The use of DPC in IIP technology was recently demonstrated for lead(II) ion extraction.28 All polymers were characterized by Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM) dynamic light scattering (DLS) and electrophoretic light scattering (ELS). Comparative selectivity studies were performed using heavy metals typically found in polluted water such as Co(II), Pb(II) and Cu(II). Finally, to confirm the applicability of the prepared polymers to real-world water samples, drinking water was chosen and analyzed before and after incubation with the prepared polymer particles, and the extraction recovery was calculated.

Experimental procedure

Reagents

DPC, mercury(II) chloride (HgCl2), 4VP and EGDMA were supplied from Sigma-Aldrich (Steinheim, Germany). Hydrochloric acid (HCl) and α-α′-azoisobutyronitrile (AIBN) were purchased from Fluka (Steinheim, Germany). Analytical grade acetonitrile and ethanol were obtained from J.T. Baker (Deventer, Holland). Elemental standard solutions of Pb(II) and Hg(II) were prepared by appropriate dilution of 1000 mg l−1 stocks purchased from Fluka. Nitric Acid (67–69%) for trace metal analysis, and Cu(II) and Co(II) standard solutions (1000 mg l−1) were supplied from Romil-SpA (Prato, Italy). Buffer solutions were prepared with deionized water provided by a water purification system (Human Corporation, Seoul, Korea).

Instrumentation

FT-IR analysis was performed on a JASCO 660 plus infrared spectrometer (Gross-Umstadt, Germany). SEM observations were carried out on a JEOL JSM 6500F microscope (Peabody, MA, USA) equipped with a field emission source. DLS and ELS were conducted using a Malvern Zetasizer Nano ZS 90 (Worcestershire, UK) on diluted samples. Sonication was carried out using a Sonorex RK 102H ultrasonic water bath (Bandelin Electronic, Berlin, Germany, www.bandelin.com). Centrifugation was performed using a PK121 multispeed centrifuge (Thermo Electron Corporation, Waltham, MA, USA, www.thermoscientific.com). Batch rebinding experiments were carried out using a Cary 100 Scan UV–visible spectrophotometer (Varian, Palo Alto, CA, USA). For pH measurements, a pH meter Basic 20 (Crison, Alella, Barcelona, Spain, www.crisoninstruments.com) was used. Ion quantification was achieved by inductive coupled plasma atomic emission spectroscopy (ICP-AES) iCAP 6000 Series (Thermo Scientific, http://www.thermoscientific.com). UV–visible spectra were obtained with a Cary 100 Scan UV/vis spectrophotometer.

Synthesis of Hg(II)-imprinted polymers

To prepare a Hg(II)-imprinted polymer, bulk polymerization was first used to obtain IIP1. In brief, HgCl2(0.0468 mmol), 4VP (0.1872 mmol), EGDMA (0.936 mmol) and AIBN (0.037 mmol) were dissolved in 3 ml of acetonitrile and water (4/1, v/v). The polymerization was carried out at 65 °C for 24 h under magnetic stirring (400 r.p.m.). A second polymer, IIP2, was synthesized in a similar manner to the above procedure with a slight modification. In the first step, DPC was dissolved in 3 ml of a porogen solution containing 4VP. After HgCl2 addition, the resulting mixture was stirred for 5 h to form a stable complex between Hg(II) ions, 4VP and DPC. Finally, a third polymer, IIP3, was prepared by precipitation polymerization similar to IIP1 but with a higher volume of porogen (13 ml). All polymers were washed several times with ethanol to remove any unreacted materials and subsequently washed with 2 M HCl to extract Hg(II) ions until Hg(II) in solution was no longer detected by ICP-AES analysis. Finally, polymeric particles were washed with double distilled water to obtain a neutral pH. The resulting fine powders were dried under vacuum in a desiccator prior to absorption studies. The corresponding non-imprinted polymers were prepared using the same procedures but without the presence of the target ion.

Morphological studies

FT-IR analysis was performed using dry polymers dispersed in a matrix of KBr followed by compression at 10 tons to form pellets. DLS and ELS measurements were both carried out on diluted samples to establish the size and zeta potential of the polymer particles. The hydrodynamic diameter of the dispersed beads was determined at 25 °C by measuring the autocorrelation function at a 90° scattering angle. Three separate measurements were made to compute an average. The polydispersity index, a measure of the distribution of molecular mass in a given polymer sample, was evaluated for all prepared polymers and the results were compared. The surface morphology, size and shape of the polymer particles were examined by SEM.

Batch adsorption studies

The binding capacity of all polymers was evaluated by batch rebinding experiments by dissolving 5 mg of polymer particles in 1 ml of HgCl2 phosphate buffer solution (pH 8), spanning a concentration range from 0 to 600 mg l−1. The suspension was shaken for 18 h at room temperature. Then, the polymer was removed by filtration, and the resulting solution was analyzed by ICP-AES. The amount of analyte adsorbed on the polymer (mg g−1) was calculated by

where Ci and Ce represent the initial and equilibrium concentration (mg l−1), respectively, V is the volume of water solution (l) and m is the mass of polymer (g).

Scatchard analysis was performed using the following equation:

where Qe represents the equilibrium concentration of Hg(II) bound per gm polymer (mg g−1), Ka (M−1) is the association constant, and Bmax (μM g−1) is the apparent maximum number of binding sites. Therefore, Ka and Bmax of the polymer were determined from the slope and the intercept, respectively, by plotting Qe/Ce versus Qe.

A simple Langmuir absorption isotherm was also used to evaluate the maximum absorption capacity Qmax (μmol g−1) of the polymers:

where Qe and KL correspond to the amount of analyte ion adsorbed at equilibrium (μmol g−1) and the Langmuir constant (l μmol−1), respectively. Qmax is determined from the linear plot of 1/Qe against 1/Ce.

Comparative selectivity studies

Ion recognition capacity of Hg-IIP materials are well-reflected by Hg(II) selectivity in the presence of other competing ions. In particular, a solution (100 mg l−1) of Hg(II), Co(II), Cu(II) and Pb(II) was used for bath experiments. The following equation was employed to evaluate the selectivity of the different prepared polymers:

where Kd, Ci and Cf represent the distribution coefficient (l mg−1) of an ion on the polymer, the initial and the final concentration of solutions (mg l−1), respectively. The selectivity coefficient k of Hg(II) relative to the competing ion was calculated using the following equation:

To evaluate an imprinting effect, a relative selectivity coefficient k’ was defined as follows:

Removal of Hg(II) ions from drinking water

Batch experiments were conducted to explore the application of prepared polymers to real-world samples. In particular, the extraction capabilities of IIP2 and IIP3 in drinking water were evaluated. Fifty milligrams of dried polymer were suspended in 100 ml of drinking water containing 1 μg ml−1 of Hg(II) ions and stirred for 1 h. After batch experiments, unextracted and extracted concentrations of Hg(II) ions were determined.

Results and Discussion

Polymer preparation

In this work, different synthesis approaches for Hg(II)-imprinted polymer preparation were employed, and the influence of DPC, used as non-polymerizable ligand, on the absorption performance was evaluated. Bulk polymerization was first employed to prepare IIP1 and IIP2 in the absence and presence of DPC. This technique was chosen for its simplicity; however, particles irregular in size and shape were produced, and some interaction sites were destroyed during grinding, influencing the IIPs’ loading capacity.17 For this reason, a third polymer, IIP3, was prepared using precipitation polymerization, increasing the amount of porogen to obtain monodisperse spherical beads and to evaluate the contribution of morphological characteristics on the absorption performance compared with the addition of DPC. The presence of DPC in the prepolymerization step probably promoted the formation of a stable complex between 4VP, DPC and Hg(II) ions, influencing the subsequent synthesis process. 4VP was chosen as the functional monomer due to its high capacity to interact with mercury ions through its nitrogen atom.25, 26, 27, 28, 29, 30 The gentle conditions used in the washing steps to remove mercury ions enabled DPC trapping in the IIP2 matrix through π–π interactions with the molecules present in the polymeric structure, as stated in a recent paper.28 The trapping of DPC in a polymer mesh was further demonstrated through FT-IR analysis overlying the corresponding IR spectra of all polymers prepared. To highlight the substantial differences in each spectrum, an enlargement in the range of 1750–700 cm−1 was considered. The IR spectrum of IIP2 (Figure 1b) showed the presence of two additional peaks (1295.9, 1320.9 cm−1) that were attributed to the stretching of nitrogen bonded to the aromatic carbon of DPC, which is not found in IIP1 (Figure 1a, black circle). These peaks were slightly shifted compared with the DPC spectrum (1289, 1309 cm−1) due to the different chemical environments and the interactions established with the polymer matrix. Moreover, the IR spectra of leached and unleached IIP2 were also evaluated. A substantial difference between the leached (Figure 1b) and unleached (Figure 1c) IIP2 spectra was obtained. The peak shape of C=N (1636.3 cm−1, black circle) in the leached IIP2 spectrum (Figure 1b) was drastically different than in the unleached IIP2 spectrum (Figure 1c) emphasizing the possible interaction of mercury ions with nitrogen on 4VP. Moreover, the spectrum in (Figure 1c) showed a slight shift of the additional peaks (1280.5, 1322.0 cm−1, black circle) observed in IIP2, which is attributable to the C–N stretching of DPC. These results show the possible interaction of mercury ions both with the nitrogen bonded to the aromatic carbon of DPC and the nitrogen of the pyridine monomer. IIP1, IIP2 and IIP3 and the corresponding non-imprinted polymers were further characterized by DLS, ELS and SEM analysis. DLS measurements of IIP1, IIP2 and the corresponding non-imprinted polymers showed high polydispersity index values (IIP1=0.872; NIP1=0.790; IIP2=0.571; NIP2=0.504), suggesting the presence of many polydisperse aggregates, as confirmed from SEM images (Figures 2a and b). Indeed, irregularly shaped particles were obtained as expected using a bulk polymerization. By contrast, as observed in Figure 2c, SEM images of IIP3, prepared by precipitation polymerization, revealed the presence of monodisperse spherical beads with an average diameter of a few micrometers, as confirmed by DLS measurements (1183±200 nm). Moreover, the zeta-potential value described good stability of IIP3 particles suspended in water (39.6±5.7 mV).

Figure 1
figure 1

FT-IR spectra of IIP1 (a), IIP2 (b) and unleached IIP2 (c). The black circles underline the presence of DPC in IIP2 (b) compared with IIP1 (a) and their changes of peaks between unleached (b) and leached IIP2 (c). FT-IR, Fourier TransformInfrared Spectroscopy; IIP, ion-imprinted polymer.

Full size image
Figure 2
figure 2

SEM images (20000 × magnification) of IIP1 (a), IIP2 (b), IIP3 (c) and NIP3 (d). IIP, ion-imprinted polymer; SEM, scanning electron microscopy.

Full size image

Interestingly, a drastic reduction was observed in the average particle size between IIP3 and the corresponding non-imprinted polymer (Figure 2c). This variation, also noted between IIP1 and NIP1, was due to the nucleation stage of the reaction being influenced by the presence of the template.31 Indeed, when the number of nucleated particles increases, the average particle diameter decreases for a given degree of a monomer conversion. These data confirmed that DLS analysis was suitable only for monodisperse spherical beads with a lower polydispersity index value, whereas in other cases the SEM technique was considered the most reliable to obtain accurate morphological information.

Effect of pH value on the adsorption properties

Because it is well-known that Hg(II) ions can form insoluble complexes in alkaline aqueous solutions,32 prior to performing the binding experiments, the complete solubility of Hg(II) was verified at the concentrations used.

To evaluate the effect of pH on adsorption performance, 5 mg of polymeric particles were dissolved in 1 ml of HgCl2 buffer solution (100 mg l−1) with a pH of 3, 8 or 11. The results demonstrated an adsorption efficiency which was highly pH dependent for all polymers, with the highest adsorption capacity at pH 8. In particular, IIPs exhibited low affinities for Hg(II) ion extraction in highly acidic or alkaline conditions, as shown by the low adsorption performance obtained. Acidic solutions have a greater affinity for metal ions, thus, Hg(II) ions were distributed more in the acidic solution than on polymeric particles. At very alkaline conditions, mercury ions form complexes with hydroxide ions, producing soluble amphoteric hydroxides rather than being adsorbed on polymeric particles.33 As a result, low adsorption capacities for all polymers tested were recorded at both high and low pH values. For this reason, it was decided to use a phosphate buffer solution with pH 8 to study the adsorption performance of polymers.

Absorption studies

The binding behavior of all polymers was tested by batch rebinding experiments. The binding data were processed by Scatchard and Langmuir analyses to evaluate the binding characteristics of the prepared polymers. As seen in Figure 3, the absorption capacity increased with analyte concentration until a saturation point was reached for all the prepared polymers. IIP1 showed an experimental maximum adsorption capacity near 40 mg g−1, which was much lower than the maximum adsorption capacity of IIP2 (70 mg g−1) prepared in the presence of DPC. This behavior suggests that the complex formation between metal ions and DPC plays a substantial role in the adsorption of Hg(II) ions. Moreover, using precipitation polymerization for the preparation of IIP3, the adsorption capacity increased (53 mg g−1). This is due to the morphological characteristics of the sorbents that affect adsorption capacity significantly. Indeed, morphological differences observed between IIP1 and IIP3, which were due to the different polymerization techniques used, were the main reason for the higher absorption of IIP3 compared with IIP1. Most likely, the low aggregation of microbeads obtained (Figure 2c) left a greater number of binding sites available for Hg(II) binding, enhancing the absorption of Hg(II) ions. Even if IIP2 showed morphological characteristics similar to IIP1 with irregularly shaped particles (Figure 2b), the presence of DPC modified the chemical environment of IIP2, promoting the formation of ternary complexes with mercury ions and 4VP and improving its binding capacity beyond that of IIP3.

Figure 3
figure 3

Batch rebinding experiments for IIP1, IIP2 and IIP3. IIP, ion-imprinted polymer.

Full size image

To describe the Hg(II) ion distribution between the liquid and adsorbent phases and to understand the nature of binding sites, Scatchard and Langmuir models were used to fit the adsorption processes of all IIPs. According to the correlation coefficients (R2), the Scatchard model was more suitable than the Langmuir model for describing the binding behavior of polymers prepared by bulk polymerization (IIP1 and IIP2). Scatchard plots obtained for these polymers showed the presence of heterogeneous binding sites (Figure 4a) with a significant difference between the Ka values for the high affinity binding sites of IIP1 and IIP2, highlighting the important contribution of DPC in Hg(II) ion extraction (Table 1). In contrast to these findings, the correlation coefficients obtained for IIP3 revealed the presence of homogeneous binding sites, which were well-described by a Langmuir isotherm. These results can be justified due to the morphological characteristics of the polymers (Figure 4b). In Figure 5, the adsorbed Hg(II) ions on each IIP and the corresponding NIP, as well as the Imprinting Factor IF (ratio of the adsorbed Hg(II) amount by IIP and NIP) were reported. Notably, IIP2 and IIP3 showed better specificity for Hg(II) ions than IIP1. Therefore, both the trapping approach and the precipitation polymerization technique have enhanced the absorption performances of the polymers. To test the regeneration capacity of the polymer, after incubation of IIP2 with HgCl2 phosphate buffer solution at a pH of 8 (100 mg l−1), the polymer was centrifuged at 8000 r.p.m. for 15 min with a 2 M HCl solution until Hg(II) ions were no longer detected in the supernatant by ICP-AES. The regenerated polymer showed uptake efficiency comparable to that of the fresh polymer, even after three cycles.

Figure 4
figure 4

Scatchard plot of IIP1(▪) with equations y=1657.2x+0.0231 for high affinity sites (left line), y=1349.2x+0.4092 for low affinity sites (right line), and IIP2 (▴) with equations y=−12138x+4.4955 for high affinity sites (right line), y=9433.5x−0.0328 for low affinity sites (left line) (a); Langmuir isotherm of IIP3 (b) with equation y=0.562x−0.004. IIP, ion-imprinted polymer.

Full size image
Table 1 Scatchard and Langmuir isotherm constants for IIP1, IIP2 and IIP3
Full size table
Figure 5
figure 5

Comparison of IIPs absorption performance with the corresponding non-imprinted polymers and imprinting factor (IF) evaluation. [Hg(II)]=100 mg l−1. IIP, ion-imprinted polymer.

Full size image

Selectivity evaluation

Heavy metals typically found in polluted water, such as Co(II), Pb(II) and Cu(II) ions, were chosen as competitors to evaluate selectivity performance. As shown in Figure 6, the interference ions did not cause significant reduction of Hg(II) absorption capacity. As seen in Table 2, IIP2 prepared using DPC showed better Hg(II) retention than IIP1 and IIP3 and showed a discrete selectivity. However, IIP3 exhibited higher binding ability for Hg(II) ions than IIP1 and IIP2, as seen from selectivity coefficients, which are all >50.8. The relative selectivity coefficients of imprinted particles for Hg(II)/Cu(II), Hg(II)/Co(II) and Hg(II)/Pb(II) confirmed a clear selectivity effect; this was especially true for IIP3, which had values that were almost 93.3, 335.0 and 25.3 times greater than that of NIP3, respectively. These results suggest that with this polymer, Hg(II) ions can be selectively removed from an aqueous medium even in the presence of Cu(II), Co(II) and Pb(II). From these results, it can be hypothesized that the presence of aggregates in IIP1 and IIP2 can cause unspecific ion retention. Thus, IIP3 monodisperse spherical beads have morphological characteristics that suggest the presence of imprinted cavities and specific binding sites in a predetermined orientation free from aggregates that allows for better discrimination between the competitor ions, resulting in a high selectivity of IIP3 for Hg(II) ions. It can be concluded that morphological characteristics were more effective in obtaining highly selective polymers than the inclusion of the ligand.

Figure 6
figure 6

Selectivity studies of IIP1, IIP2 and IIP3 using a mixture solution of Pb(II), Cu(II), Co(II) and Hg(II). [M(II)]=100 mg l−1. IIP, ion-imprinted polymer.

Full size image
Table 2 Selectivity data on IIP1, IIP2 and IIP3
Full size table

Application to real samples

To verify the feasibility of the application of these polymers, the removal of ions from drinking water was tested. IIP2 and IIP3, which showed the best absorption performance, were incubated with drinking water samples spiked with 1 μg ml−1 of Hg(II) ions. Despite the high selectivity performance showed by IIP3, in drinking water, the ligand rather than the monodispersity of the polymer particles, had a more significant role in the removal of mercury ions. Thus, IIP2 demonstrated the best removal efficiency (78.8%, RSD 0.9%) compared with IIP3 (61.6%, RSD 1.0%), suggesting a good anti-interference ability in environmental water samples.

Conclusion

In this work, selective Hg(II) imprinted polymers were prepared using different synthesis approaches, and the influence of DPC, used as non-polymerizable ligand, on absorption performance was demonstrated. Absorption studies on IIP1 and IIP2 in the absence and presence of DPC confirmed a strong impact of this compound on the binding behavior of polymers against Hg(II) ions, with a significant difference between Ka values of high affinity binding sites of IIP1 (1.7±0.4 M−1) and IIP2 (12.1±0.5 M−1). Moreover, an increase of absorption capacity compared with IIP1 was also observed for IIP3, demonstrating that the morphological characteristics of the polymer influenced the presence of homogenous binding sites. Indeed, SEM images of IIP3 showed the presence of monodisperse spherical microbeads with a low propensity to aggregate, whereas the presence of some aggregates was observed for other polymers. Selectivity studies revealed that morphological characteristics of polymers were more effective than the ligand inclusion for the selectivity of Hg(II) imprinted polymers. However, batch experiments conducted using drinking water spiked with 1 μg ml−1of Hg(II) ions showed a greater incidence of ligand on mercury ion extraction with a removal efficiency near 80% for IIP2 after 1 h of treatment.