Microchemical Journal 113 (2014) 42–47

Contents lists available at ScienceDirect

Microchemical Journal

j ourna l homepage: www.e lsev ie r .com/ locate /mic roc

Polymeric ion-imprinted nanoparticles for mercury speciation insurface waters

Tanya Yordanova, Ivanka Dakova ⁎, Konstantin Balashev, Irina KaradjovaFaculty of Chemistry and Pharmacy, University of Sofia “St. Kliment Ohridski”, 1 James Bourchier Blvd., Sofia 1164, Bulgaria

⁎ Corresponding author. Tel.: +359 2 81 61 244; fax: +E-mail address: i.dakova@chem.uni-sofia.bg (I. Dakova

0026-265X/$ – see front matter © 2013 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.microc.2013.11.008

a b s t r a c t

a r t i c l e i n f o

Article history:Received 11 October 2013Received in revised form 8 November 2013Accepted 9 November 2013Available online 16 November 2013

Keywords:Mercury speciationIon-imprinted nanoparticlesSolid phase extractionCold vapor atomic absorption spectrometrySurface water

Selective sorption and enrichment of Hg(II) were achieved by polymeric ion-imprinted nanoparticles, pre-pared by copolymerization of methacrylic acid (functional monomer), trimethylolpropane trimethacrylate(cross-linking agent), 2,2′-azo-bis-isobutyronitrile (initiator) and Hg(II) complexes with chelating agent1-pyrrolidinedithiocarboxylic acid. The particles were characterized for their morphology, composition andstructure by using AFM, FTIR and elemental analysis. Extraction efficiency and selectivity of ion imprinted sorbenttoward Hg(II) were studied by batch procedures. Two distinct adsorption sites were identified on the surface ofthe polymeric nanoparticles through Scatchard analysis performed. Adsorption process was characterized byLangmuir isothermmodel. Solid phase extraction procedure was developed for Hg speciation and determinationin surface waters. The accuracy of the proposed speciation scheme was confirmed by the analysis of certifiedreferencematerial. Limits of quantification achieved (0.015 μg L−1 for Hg(II) and0.02 μg L−1 formethylmercury)and RSD values (7–12% for both species) completely satisfy requirements of European legislation for monitoringof surface waters.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Mercury and its compounds are considered as one of the most toxicand harmful environmental pollutants and according to the Europeanlegislation are identified as priority hazardous substances [1]. It is wellknown that toxicity and mobility of Hg strongly depend on its chemicalform.Mercury can naturally exists in several chemical formswith differ-ent properties but the main species to be identified and determined insurface waters are inorganic mercury (Hg(II)) and methylmercury(CH3Hg+). The latter is more hydrophobic, so it can easily crossbiomembranes and be accumulated by living cells. Generally, methyl-mercury is much more toxic than inorganic Hg and can be naturallyformed in the processes of biomethylation mediated by activities ofmicroorganisms in aquatic systems. Due to the significant differencesbetween toxicity, bioavailability and mobility of Hg(II) and CH3Hg+,speciation analysis is an actual research topic of modern analyticalchemistry.

In real unpolluted environmental samples total Hg content is usuallyvery low, for surface waters, typical concentrations are in the range ofng to pg L−1. Leopold et al. have summarized several recent reports,where levels of total Hg in uncontaminated natural waters are between0.03 and 90 ng L−1, while methylmercury ranges from 1 to 40% of totalHg concentration [2]. Due to their low concentrations and the influencesof coexisting substances in real water samples, development of analyti-cal procedures for reliable and accurate determination and speciation of

359 2 962 54 38.).

ights reserved.

Hg at ng L−1 levels is still an analytical challenge. A large number oftechniques for determination and speciation of Hg have been developedcommonly based on chromatographic [3–7] or non-chromatographicseparations followed by highly sensitive instrumental detection[8–12]. Nevertheless, sample pretreatment prior to measurement isoften demanded due to the extremely low Hg concentrations or matrixinterferences observed. Solid-phase extraction (SPE) is a widely usedanalytical procedure for preliminary separation and preconcentrationof the toxic elements which offers several advantages such as perfor-mance simplicity, low solvent consumption, minimization of samplevolumes, fast separation of phases without emulsion formation, highenrichment factor, good reproducibility and flexibility [13,14]. Relative-ly low selectivity of conventional sorbents used is the main disadvan-tage of this approach. During the last years ion-imprinted polymers(IIPs) have received much attention as selective sorbents [15] as theypermit simultaneous preconcentration and speciation of target ions.IIPs are synthetic materials with artificially generated recognition sitesable to specifically rebind a target chemical species. In brief, the synthe-sis of IIPs involves several stages: (i) complexation of a template species(metal ions or their complexes with specific ligands) with functionalmonomers, (ii) copolymerization of these monomers around thetemplate with the help of a cross-linking agent in the presence of aninitiator, and (iii) leaching out of the template which leaves behindbinding sites that are complementary in size and shape to the imprintedspecies. High selectivity of IIPs is explained by the polymer memoryeffect toward the metal ion interaction with a specific ligand, coordina-tion geometry, metal ion coordination number, ion charge, and size[15,16]. Numerous studies on IIPs and their application for selective

43T. Yordanova et al. / Microchemical Journal 113 (2014) 42–47

preconcentration and separation of metal ions have been reported[17–19] including Hg(II)-imprinted ones [10,20–24].

In the present study, spheric nanosized Hg(II)-IIP is synthesizedvia dispersion copolymerization using methacrylic acid (MAA) as afunctional monomer, trimethylolpropane trimethacrylate (TMPTMA)as a cross-linking agent, 1-pyrrolidinedithiocarboxylic acid (PDC) as anon-polymerizable chelating agent and Hg(II) as a template ion. Sizedistribution, morphology and composition of ion-imprinted nanoparti-cles are defined by AFM, FTIR and elemental analysis. The extractionefficiencies and selectivities of the imprinted with Hg(II)-PDC complex(P(PDC-Hg)) polymeric nanoparticles and non-imprinted (blank)polymer P(B) are studied in a batch system. Analytical scheme for Hgspeciation and enrichment in surface waters is developed based onthe high selectivity of polymeric nanoparticles (P(PDC-Hg)) towardHg(II). The versatility of proposed analytical method is confirmed bythe analysis of certified reference material BCR®-505 Estuarine Water.

2. Experimental

2.1. Materials

Methacrylic acid (MAA), trimethylolpropane trimethacrylate(TMPTMA), 2,2′-azobisisobutyronitrile (AIBN), 1-pyrrolidinedithio-carboxylic acid ammonium salt (PDC), thiourea (TU) (Merck, Darmstadt,Germany), and acetonitrile (ACN) (Labscan, Dublin, Ireland)were used toprepare the Hg(II) ion-imprinted and non-imprinted polymer sorbents.Stock standard solutions for Hgwere: inorganic Hg(II), stock standard so-lution for AAS, Trace CEPT™, 998 μg mL−1 in 2 mol L−1 HNO3, (Sigma-Aldrich, USA). Methylmercury stock solutions (100 mg L−1) wereprepared by dissolving the appropriate amount of methylmercury(I)chloride (highly toxic, handle with care), PESTANAL®, analytical stan-dard, Fluka (33368) in 5 mL of methanol and diluting to 100 mL withwater. The exact concentration of this solution was defined against stan-dards prepared from inorganic Hg(II). All stock solutions were stored indark glass bottles at 4 °C. Working standard solutions for calibrationwere freshly (daily) prepared by successive dilution and contained re-agents used for sample preparation or elution; e.g. 0.1 mol L−1 thioureain 0.1 mol L−1 HCl. Reductant solutions of sodium tetrahydroborate(III)NaBH4 (Merck) 0.04% (m/V) and 0.4% (m/V) in 0.1% (m/V) NaOH wereprepared fresh daily and was used without filtration. The stock standardsolutions for Cd(II), Cu(II), Fe(III), Pb(II), Mn(II), Ni(II) and Zn(II)(1000 μg mL−1) were Titrisol, Merck (Darmstadt, Germany), in 2%HNO3 (V/V).

All reagents were of analytical-reagent grade and all aqueous solu-tions were prepared in high-purity water (Millipore Corp., Milford,MA, USA).

For validation purposes a BCR®-505 EstuarineWater CRM stabilizedwith HNO3 at pH 1.6, with ‘Additional material information’ valueof 0.69 nmol kg−1 was used.

2.2. Apparatus

CVAAS measurements were carried out with a Perkin ElmerAAnalyst 400 atomic absorption spectrometer (Perkin Elmer, USA)equipped with a MHS 15 Mercury Hydride System and an externallyheated quartz tube atomizer at room temperature for inorganic Hg(II)and in air/acetylene flame for CH3Hg+. The optimal instrumentalparameters for Hg species determination include: pre-reaction purgetime 7 s and post-reaction purge time 60 s. Solutions of NaBH4 0.04%(m/V) and 0.4% (m/V) were used as reductants for inorganic Hg(II)and CH3Hg+ respectively.

The IR-spectra (4000–400 cm−1) in KBr were recorded on a Nikolet6700 FTIR spectrometer (Thermo Scientific, USA).

For AFM imaging NanoScope Multi Mode V system (Bruker Inc.,Germany) was used operating in tapping mode in air at room

temperature. Instrument setup, instrumental parameters and samplepreparation procedure are described in Supplementary materials(Section S2.2).

Elemental analysis (C, H, N, S) were performed using a Euro EACHNS-O elemental analyzer (EuroVector, Italy).

Amicroprocessor pHmeter (Hanna Instruments, Portugal)was usedfor pH measurements.

2.3. Synthesis of the Hg(II) ion-imprinted and non-imprinted sorbents

The synthesis of Hg(II) ion-imprinted and non-imprinted polymernanoparticles was adapted from the procedure reported by Dakovaet al. [25]. The Hg(II)-IIP was prepared via a cross-linking dispersioncopolymerization using MAA (1.8 mmol) as a functional monomer,TMPTMA (2.8 mmol) as a cross-linking agent, AIBN (52 mg) as an initi-ator, Hg(II) as a template ion (0.18 mmol), PDC as a non-polymerizablechelating agent (0.36 mmol) and acetonitrile (75 mL) as a solvent. Thesolution was saturated with dry nitrogen for 15 min and copolymeriza-tion was carried out at T = 60 °C for 24 h. The polymer particles ob-tained were recovered by centrifugation, washed with ACN to removeunreacted monomers and other ingredients. Mercury was removedfrom the sorbent by several, sequential elution steps using 4 mol L−1

HNO3 and 0.5 mol L−1 thiourea in 0.5 mol L−1 HCl as eluents. Thisprocedure was repeated until the Hg concentration (template ions)in the eluate solution is below the LOQ as measured by CVAAS. Non-imprinted polymer sorbent (called P(B)) was synthesized in the sameway as described above, in the absence of chelating agent and templateion. Reaction scheme for Hg(II)-IIP preparation is shown in Fig. S1(Supplementary material).

2.4. Hg(II) sorption/desorption studies

The effects of solution pH, concentration and volume of the eluent,sorption and elution time on the degree of sorption and degree of elu-tion of Hg(II) were evaluated by a batch procedure. Sorbent particles(50 mg) were added to an aqueous solution (10 mL) containing 5 ngHg. The pH of this solution was varied in the range 1–8, using NH4OHand HNO3 solutions. After pH adjustment the solution was stirredwith an electric shaker for 30 min and then centrifuged. In order to in-vestigate the degree of Hg(II) sorption the supernatant was removedand analyzed by CVAAS as an effluate. The sorbent was washed withdeionized water and Hg(II) was eluted from the sorbent particlesby 2 mL of various reagents for 30 min. The degree of Hg elution wasdetermined in the eluate solutions by CVAAS.

The kinetics of the Hg(II) sorption and desorptionwere investigatedin a batch system. Ten milliliters of Hg solution (5 ng) was treated with50 mgof Hg(II)-IIP particles at pH 7 for 5–40 min. The contents ofmetalions in the effluate and eluate solutions after sorption and elution,respectively, were determined by CVAAS.

The degree of sorption (DS %) of Hg(II) ions is defined as:

DS% ¼ Ai−Aeffð Þ=Ai½ �·100; ð1Þ

where Aeff (μg) is the cation amount in the effluate solution afterextraction by synthesized sorbents from a solution with a total cationamount Ai (μg).

The degree of elution (DE %) of sorbed Hg(II) ions from loadedsorbent is defined as:

DE% ¼ Ael= Ai−Aeffð Þ½ �·100; ð2Þ

where Ael (μg) is the amount of metal cations in solution after elutionprocess.

44 T. Yordanova et al. / Microchemical Journal 113 (2014) 42–47

The distribution ratio (D) is defined as

D ¼ Ai−Aeffð Þ=Aeff : ð3Þ

2.5. Capacity and adsorption isotherms

To measure the adsorption capacity (Q), 50 mg of dried sorbent(imprinted or blank)was added to 10 mL solutionwith various concen-trations of mercury ions (from 20 to 240 μmol L−1). Themixtures wereshaken for 30 min at room temperature and pH 7. The total mercuryions sorbed was measured by CVAAS after elution with mixture of0.1 mol L−1 thiourea in 0.1 mol L−1 HCl. Adsorption capacity wascalculated using the following equation:

Q ¼ C0−Ceð Þ·V½ �=m; ð4Þ

where: Q (μmol g−1) is the mass of Hg(II) ions adsorbed per unit massof the sorbent; V (L) is the volume of the solution, m (g) is the mass ofthe sorbent, C0 (μmol L−1) is the initial concentration of Hg(II) ions inthe solution; and Ce (μmol L−1) is equilibrium concentration of Hg(II)in solution after adsorption.

The Langmuir Eq. (5) was applied to the data to provide an insightinto the mechanism of the adsorption [26].

Ce=Qe ¼ Ce=Qmax þ 1=ðb·QmaxÞ; ð5Þ

where Ce (μmol L−1) is the equilibrium concentration of Hg ions in thesolution, Qe (μmol g−1) is the adsorption capacity of the adsorbedmercury ions onto the sorbents at equilibrium, Qmax (μmol g−1) isthe theoretical maximum adsorption capacity, and b (L μmol−1) is theLangmuir constant that relates to the affinity of binding sites.

2.6. Selectivity experiments

The selectivity of ion-imprinted and non-imprinted sorbents towardHg(II) was evaluated by batch procedure: 50 mg of sorbent was stirredfor 30 min with a 10 mL of synthetic metal mixture solution containing5 ng Hg(II) and CH3Hg+, 10 μg Pb(II), Fe(III) and Ni(II), and 2 μg Cd(II),Cu(II), Mn(II) and Zn(II) at a pH 7. The solutions were centrifuged,sorbent particles were washed twice with deionized water and elutionwas performed with 10 mL 0.1 mol L−1 thiourea in 0.1 mol L−1 HClfor 30 min. The concentrations of Hg(II) and CH3Hg+ were measuredby CVAAS and Cd(II), Cu(II), Fe(III), Mn(II), Ni(II), Pb(II), and Zn(II) inthe supernatant and eluate solutions were determined by flame AAS.

The selectivity coefficient (SHg/Me) is

SHg=Me¼ DHg=DMe ð6Þ

where DHg and DMe are the distribution ratios for Hg(II) and CH3Hg+,Cd(II), Cu(II), Fe(III), Pb(II), Mn(II), Ni(II) and Zn(II), respectively.

The relative selectivity coefficient (k′) is:

k′ ¼ SHg=Me imprintedð Þ=SHg=Me controlð Þ ð7Þ

where SHg/Me (imprinted) and SHg/Me (control) are selectivity coefficients ofthe imprinted and non-imprinted particles, respectively. The relativeselectivity coefficient is an estimate of the effect of imprinting onselectivity.

2.7. Analytical procedure for Hg(II) and CH3Hg+ determination in

surface waters

Inorganic mercury and methylmercury were determined in twosteps:

(1) Hg(II) determination: Water sample 50 mL (adjusted to pH 7with acetate buffer) was stirred with 50 mg P(PDC-Hg) for15 min. After centrifugation the supernatant was removed andpolymer washed twice with deionized water, finally Hg(II) waseluted with 2 mL 0.1 mol L−1 thiourea in 0.1 mol L−1 HCl.Hg(II) was measured by CVAAS under optimal instrumentalparameters using 0.5 mL of eluate solution.

(2) Total Hg determination: In parallel samewater (river, sea) sampleca. 50 mL was MW digested according to EPA (in order totransform CH3Hg+ into Hg(II)) and then prepared according tothe already described analytical protocol. In this case the sum(Hg(II) + CH3Hg+) was determined by CVAAS.

(3) CH3Hg+ determination: The content of CH3Hg+ in the analyzed

water sample was calculated by substraction of Hg(II) from thetotal Hg.

3. Results and discussion

Preparation of Hg(II)-IIP includes several steps (Fig. S1). At the firstone the prepolymerization complex is formed by non-covalent inter-actions between the functional monomer (MAA) and the templateHg(II) ion chelated by PDC. At the second step the copolymer networkis formed by the dispersion cross-linking copolymerization of theprepolymerization complex with the TMPTMA as a cross-linkingagent. Finally as a third step, Hg(II) ions were removed from copolymernetwork prepared leaving behind some specific binding sites with func-tional groups in a predetermined orientation and cavities with specialsize of templates.

3.1. Characterization studies

Atomic Force Microscopy (AFM) was used to investigate the mor-phology, size and shape of P(PDC-Hg) and P(B). The two- and three-dimensional high images of blank sorbent P(B) are shown in Fig. S2 Aand C (Supplementary material), while the same images of P(PDC-Hg)are presented in Fig. S3 A and C (Supplementarymaterial), respectively.

As can be seen from Fig. S2 polymeric particles of blank sorbent P(B)have close to spherical shape or in fewcases bean like shape. The sectionanalysis of the obtained AFM images and detailed measurement ofparticle sizes and shapes shows that diameters of P(B) particles areabout 23 ± 1 nm (see the histogram at Fig. S2 D).

The AFM images of imprinted particles P(PDC-Hg) (Fig. S3) showedthat they have more complex form (see the inset of Fig. S3) which canbe approximated to a irregular spherical shape. The section analysis ofthe obtained AFM images and detailed measurement of particle sizesand shapes shows that diameters of P(PDC-Hg) are about 73 ± 2 nm(see the histogram at Fig. S3 D), hence imprinted particles P(PDC-Hg)are more than 3 times “bigger” in size than P(B). More complicatedshapes as well as bigger sizes of imprinted nanoparticles might beexplainedwith the inclusion of template Hg complex into the copolymernetwork.

In the FTIR spectra of all prepared imprinted and controlsorbents the characteristic C_O stretching bonds which belong toester functional group in the structure of TMPTMA (seen between1730 and 1733 cm−1), and the strong and broad absorption bandsdue to the presence of the \OH stretching of carboxylic groups(seen at 3640–3150 cm−1) were observed. In the FTIR spectra ofP(PDC-Hg) there were no prominent features of the dithiocarbamategroups because they were obscured by the strong absorption peaks ofpolymer matrix around 1150 cm−1. However the vibration frequency

45T. Yordanova et al. / Microchemical Journal 113 (2014) 42–47

of C\N bond (in N\CS2 fragment) observed at 1465 cm−1 for blankpolymer P(B) was shifted to higher frequencies of 1475 cm−1 whichis an indication that a metal\ligand bond has been formed betweenHg and chelating groups in P(PDC-Hg).

3.2. Optimization of chemical parameters for quantitative sorptionand desorption of Hg(II) with imprinted P(PDC-Hg) and blankP(B) nanoparticles

The newly synthesized Hg(II)-IIPs (P(PDC-Hg) and blank (P(B)) sor-bents were tested for Hg(II) sorption from an aqueous solution at vari-ous pH (1–8) using the procedure described in Section 2.4. Quantitativesorption for Hg(II) was achieved for pH range from 5 to 7 with P(PDC-Hg) while the degree of sorption for blank polymer P(B) was around50% (Fig. 1). Evidently, the specific interaction between chelating ligandand Hg(II) does not take place in the case of P(B), thus explaining theobserved lower degree of sorption. The degree of sorption for bothstudied sorbents decreased at pH N 7, most probably related to thehydrolysis or precipitation of themercury hydroxide and its quick decom-position to oxide. Further the process of desorption e.g. eluent type, con-centration and volumewas studied. Quantitative elution of sorbed Hg(II)was achieved using 0.1 mol L−1 thiourea in 0.1 mol L−1 M HCl withminimum volume of 2 mL. The kinetics of both processes of sorptionand desorption of Hg(II) were investigated by batch procedure with50 mg of imprinted nanoparticles for 5–60 min. Quantitative sorptionwas reached for 20 min while for quantitative desorption 30 min has tobe used. In order to define preconcentration factor sample-to-eluentvolume ratio was varied between 10 and 100. Relatively low size ofsorbent particles restricts application of sample volumes higher than50 mL which yields preconcentration factor of 25.

3.3. Adsorption isotherms and Scatchard plots

The adsorption capacity of the polymeric nanoparticles is an impor-tant characteristic to determine the sorbent' amount required for quan-titative extraction of the analyte from sample solution and to evaluatethe binding properties of imprinted and non-imprinted particles. Inorder to determine the adsorption capacities of P(PDC-Hg) and P(B), ad-sorption experiments were conducted in the Hg(II) solutions with in-creasing initial concentrations (20–220 μmol L−1) under optimumconditions. Adsorption isotherms constructed (presented in Fig. S4aand b of the Supplementary materials) showed that the amount ofadsorbed Hg per unit mass of the sorbent increased with the initial con-centration of Hg(II). In order to reach the “saturation”, the initial Hg(II)concentrations are increased up to the plateau values, determining the

Fig. 1. pH-dependence of the degree of sorption (DS %) of Hg(II) ions with imprintedP(PDC-Hg) and blank P(B) polymers.

adsorption capacity values. The average maximum adsorption capaci-ties were 64.0 μmol g−1 and 21.9 μmol g−1 for P(PDC-Hg) and P(B),respectively. As expected the adsorption capacity of P(PDC-Hg) washigher versus non-imprinted polymer particles P(B), indicating thatthe cavities created after removal of template ions ensures higheraffinity of Hg(II) to the P(PDC-Hg) than to the P(B).

In order to study the binding properties of the P(PDC-Hg) and P(B),Scatchard analysis was performed by using the saturation binding data.The Scatchard equation is:

Qe=Ce ¼ Qmax–Qeð Þ=KD; ð8Þ

where Qe is the amount of mercury ions bound to sorbent at equilibri-um, Ce is the free mercury ion concentration at equilibrium, KD is theequilibrium dissociation constant of binding sites andQmax is the appar-ent maximum binding amount of mercury [27]. The values of KD andQmax could be calculated from the slope and intercept of the linearcurve plotted at Qe/Ce versus Qe. The shape of Scatchard plot is relatedto the type of the interaction of absorbate with the adsorbent. If astraight line is obtained by plotting Qe/Ce versus Qe, then the adsorbentpresents only one type of binding sites and if Scatchard plot shows adeviation from linearity, the adsorbent present more than one typeof binding sites.

Fig. 2. Scatchard plots for the adsorption of Hg(II) ions onto P(PDC-Hg) (a) and P(B) (b).

46 T. Yordanova et al. / Microchemical Journal 113 (2014) 42–47

Scatchard plot for P(PDC-Hg) and P(B) is presented in Fig. 2a and b.As can be seen from Fig. 2a the Scatchard plot for P(PDC-Hg) is not a sin-gle linear curve, but consists of two linear parts with different slopes,which indicated that there existed two kinds of binding sites of highand low affinities. The linear regression equations for the left (high af-finity binding sites) and right (low affinity binding sites) parts of thecurve in the figure were Qe/Ce = −1.309 Qe + 63.549 (R2 = 0.991)and Qe/Ce = −0.260 Qe + 18.096 (R2 = 0.966), respectively. Basedon the values of the slope and the intercept of the biphasic curve(Fig. 2a), the calculated values for KD were 0.764 and 3.843 μmol L−1

and for Qmax were 48.55 and 69.55 μmol g−1, respectively. It may beconcluded that the binding sites in the Hg(II)-imprinted sorbent couldbe classified into two distinct groups with different specific bindingproperties — thiocarboxylic groups in PDC (specific interaction) andcarboxylic group (non-specific interaction) in polymer matrix.

For P(B) one straight linewas observedwhich reveals only low affinitybinding sites on the surface of non-imprinted polymer particles (Fig. 2b).The linear regression equation was Qe/Ce = −0.075Qe + 1.629(R2 = 0.993). KD and Qmax were 13.351 μmol L−1 and 21.76 μmol g−1,respectively. The adsorption capacity of P(B) toward Hg(II) wasmuch lower than that of P(PDC-Hg). It hinted that P(B) did nothave the specific adsorption abilities of P(PDC-Hg). These results ofSchatchard analysis are typical for a pair of the imprinted and non-imprinted polymers obtained by the non-covalent approach [28].

The imprinted and non-imprinted sorbents were also characterizedthrough the Langmuir isotherm model (Eq. (5)). The calculated coeffi-cients of the Langmuir model along with regression coefficients (R2)are presented in Table S1 of the Supplementarymaterial. High R2 values(N0.97) for the sorption of Hg ions onto imprinted and non-imprintedsorbents show that the Langmuir equation gives a good fit to thesorption isotherm. The experimental values for sorption capacity arevery close to the calculated Qmax values. To predict the favorability of asorption system, the essential characteristics of the Langmuir equationcan be expressed in terms of a dimensionless factor, RL, which is definedas [29]:

RL ¼ 1= 1þ Cmbð Þ ð9Þ

where Cm means the maximal initial concentration of adsorbate. TheRL value indicates whether the type of the isotherm is favorable(0 b RL b 1), unfavorable (RL N 1), linear (RL = 1), or irreversible(RL = 0). The RL values for the adsorption of Hg(II) onto P(PDC-Hg)and P(B) were 0.01 and 0.12, respectively, indicating a favorableprocess.

3.4. Selectivity experiments

In order to prove the selectivity of P(PDC-Hg), a competitive adsorp-tion of Hg(II), CH3Hg+, Cd(II), Cu(II), Fe(III), Mn(II), Ni(II), Pb(II), andZn(II), from their model mixture was performed. The distribution(D) and selectivity (SHg/Me) coefficients of cations, achieved by usingimprinted nanoparticles P(PDC-Hg) and blank polymer P(B), are givenin Table S2. The distribution ratio of Hg(II) with P(PDC-Hg) is muchhigher than that of Cd(II), Co(II), Cu(II), Ni(II), Pb(II), Zn(II) andCH3Hg+. The greatest difference was found between distribution ratiosof Hg(II) and CH3Hg+ (99 and 0.03, respectively). The large size ofCH3Hg+ probably hinders the ion diffusion and further interactionwith functional groups resulted in very low extraction efficiency of

Table 1Analysis of BCR®-505 Estuarine Water for Hg(II) and CH3Hg+ (three parallel determinations).

Sample Found Hg(II), nmol kg−1, mean RSD, %

BCR®-505 0.62 9BCR®-505 spiked with CH3Hg+ 0.67 7BCR®-505 spiked with CH3Hg+ 0.70 9

P(PDC-Hg) toward CH3Hg+ and highest values for selectivity coeffi-cients (calculated by Eq. (6)). These results unambiguously confirmthe very good selectivity of P(PDC-Hg) toward the target (imprintedHg(II)) ions and its applicability for Hg speciation.

The values of the relative selectivity coefficient k′ (defined as a ratiobetween selectivity coefficients of imprinted and non-imprinted poly-meric nanoparticles) demonstrate hundreds to thousand times betterselectivity of P(PDC-Hg) in comparison with P(B) toward Hg(II) dueto the effect of imprinting (Table S2).

3.5. Stability and reuse properties of P(PDC-Hg)

Stability and reusability are important properties of the sorbents,which determine their analytical and economical efficiency. In orderto examine the reusability of the prepared P(PDC-Hg) sorbent, the sorp-tion/elution cycles were repeatedly performed by using the same nano-particles. Experiments performed showed around 5% lower sorptioncapacity of P(PDC-Hg) after 100 sorption/elution cycles. Although thesorption selectivity and extraction efficiency (degree of sorption anddegree of elution) of the sorbent do not changed significantly. Theresults showed that theHg(II)-imprinted nanoparticles could be repeat-edly used without any significant loss in their initial binding affinity.This is probably due to the good trapping of PDC ligand in the polymernetwork and its mechanical and chemical stability.

The stability of the P(PDC-Hg) was evaluated for a period of6 months. Experimental results demonstrated fair high stability:the sorption capacity of Hg(II)-imprinted nanoparticles and the degreeof sorption and elution of Hg(II) under optimal chemical conditionswere not changed after 6 months of storage.

3.6. Analytical application of P(PDC-Hg) and figures of merit

The prepared imprinted nanoparticles P(PDC-Hg) were tested forthe determination of Hg(II) and CH3Hg+ in surface waters. As a firststep recovery studies were performed by spiking water samples withHg(II) and CH3Hg+ and Hg determination under proposed speciationscheme (Section 2.7). Recoveries achieved varied in the range 92–96%for both Hg species in different types of waters e.g. river, sea andminer-al. Second the accuracy of analytical procedure for Hg(II) determinationwas confirmed by the analysis of certified reference material BCR®-505EstuarineWater for total Hg content. Sample amount of 50 mL was an-alyzed according to thefirst step of proposed speciation scheme. Resultsobtained of 0.66 ± 0.04 nmol kg−1 (Table 1) were in good agreementwith the (‘additional material information’) value of 0.69 nmol kg−1

Hg. In order to prove the accuracy for CH3Hg+ quantification, 30 mLsamples of BCR®-505 Estuarine Water were spiked with CH3Hg+ attwo concentration levels and analyzed according to the speciationscheme. Results obtained are summarized in Table 1. As can be seen re-coveries found for CH3Hg+ are between 96% and 104%which proves theversatility of analytical scheme for Hg speciation and applicabilityof imprinted nanoparticles P(PDC-Hg). The limit of quantificationachieved for Hg(II), calculated as 10 times standard deviation of theblank (50 mg P(PDC-Hg) eluted with 2 mL 0.1 mol L−1 thiourea in0.1 mol L−1 HCl), is 0.015 μg L−1 which completely satisfies the re-quirements of European Directive 2009/90/EC for total Hg determina-tion in surface waters [30]. The limit of quantification calculated forCH3Hg+ is 0.02 μg L−1 which is higher than environmentally relevantconcentrations of CH3Hg+ in unpolluted waters. The relative standard

Added CH3Hg+, μg L−1 Found CH3Hg+, μg L−1, mean RSD, %

– –

0.50 0.52 111.00 0.96 13

Table 2Comparative analysis of river water for total Hg.

Sample SPE by P(PDC-Hg), CVAAS, ng L−1 Direct ICP MS, ng L−1

Iskar (sampling site 1) 16.2 ± 1.2 17.5 ± 1.1Iskar (sampling site 2) 18.3 ± 1.5 19.3 ± 1.2

47T. Yordanova et al. / Microchemical Journal 113 (2014) 42–47

deviation varied in the range 6–9% for Hg(II) and 7–12% for CH3Hg+ forconcentration levels for both species of 0.03 to 1 μg L−1. Additionallycomparative analysis of river waters (river Iskar) from two samplingsites near Sofia was performed by proposed SPE procedure and directICP-MS measurement of total Hg. Results obtained are in reasonableagreement thus confirming validity of analytical method developed(Table 2).

4. Conclusions

Ion-imprinted polymeric nanoparticles P(PDC-Hg) were synthe-sized by dispersion polymerization in the presence of Hg(II) complexeswith chelating agent 1-pyrrolidinedithiocarboxylic acid as templatespecies. Scatchard analysis performed identifies two distinct adsorptionsites with different binding properties on the surface of P(PDC-Hg).As was expected, incorporation of PDC into the polymer matrix resultsin high selectivity of P(PDC-Hg) toward target Hg(II) species. Solidphase extraction procedure for Hg speciation and enrichment, basedon P(PDC-Hg) was developed. Limits of quantification and RSD valuesachieved permit speciation and determination of Hg in surface watersand could be used for routine monitoring.

Acknowledgment

This study was financially supported by the Bulgarian NationalScientific Foundation (Grant No. DFNI-T01/5 NANOSORBLAB).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.microc.2013.11.008.

References

[1] European Union Directive 2008/105/EC of the European Parliament andof the Council of 16 December 2008 on environmental quality standards in thefield of water policy, amending and subsequently repealing Council Directives82/176/EEC, 83/513/EEC, 84/156/EEC, 84/491/EEC, 86/280/EEC and amendingDirective 2000/60/EC of the European Parliament and of the Council, Off. J. Eur.Union L348 (2008) 84–97.

[2] K. Leopold, M. Foulkes, P.J. Worsfold, Methods for the determination and speciationof mercury in natural waters — a review, Anal. Chim. Acta. 663 (2010) 127–138.

[3] J.L. Gómez-Ariza, F. Lorenzo, T. Garcia-Barrera, Simultaneous determination ofmercury and arsenic species in natural freshwater by liquid chromatography withon-line UV irradiation, generation of hydrides and cold vapor and tandem atomicfluorescence detection, J. Chromatogr. A 1056 (2004) 139–144.

[4] C.M. Tseng, C.R. Hammerschmidt, W.F. Fitzgerald, Determination of methylmercuryin environmental matrixes by on-line flow injection and atomic fluorescencespectrometry, Anal. Chem. 76 (2004) 7131–7136.

[5] N. Demuth, K.G. Heumann, Validation of methylmercury determinations inaquatic systems by alkyl derivatization methods for GC analysis using ICP-IDMS,Anal. Chem. 73 (2001) 4020–4027.

[6] W.R.L. Cairns, M. Ranaldo, R. Hennebelle, C. Turetta, G. Capodaglio, C.F. Ferrari, A.Dommergue, P. Cescon, C. Barbante, Speciation analysis of mercury in seawater

from the lagoon of Venice by on-line pre-concentration HPLC–ICP-MS, Anal. Chim.Acta. 622 (2008) 62–69.

[7] Y.G. Yin, M. Chen, J.F. Peng, J.F. Liu, G.B. Jiang, Dithizone-functionalized solid phaseextraction-displacement elution-high performance liquid chromatography-Inductively coupled plasma mass spectrometry for mercury speciation in watersamples, Talanta 81 (2010) 1788–1792.

[8] J. Sanz, J.C. Raposo, J. Larreta, I. Martinez-Arkarazo, A. de Diego, J.M. Madariaga,On-line separation for the speciation of mercury in natural waters by flowinjection-cold vapour-atomic absorption spectrometry, J. Sep. Sci. 27 (2004)1202–1210.

[9] M.V. Balarama Krishna, D. Karunasagar, S.V. Rao, J. Arunachalam, Preconcentrationand speciation of inorganic and methyl mercury in waters using polyaniline andgold trap-CVAAS, Talanta 68 (2005) 329–335.

[10] I. Dakova, I. Karadjova, V. Georgieva, G. Georgiev, Ion-imprinted polymethacrylicmicrobeads as new sorbent for preconcentration and speciation of mercury, Talanta78 (2009) 523–529.

[11] E. Stanisz, J. Werner, H. Matusiewicz, Mercury species determination by task specificionic liquid-based ultrasound-assisted dispersive liquid–liquid microextractioncombined with cold vapour generation atomic absorption spectrometry,Microchem. J. 110 (2013) 28–35.

[12] Y. Gao, Z. Shi, Z. Long, P. Wu, C. Zheng, X. Hou, Determination and speciation ofmercury in environmental and biological samples by analytical atomic spectrometry,Microchem. J. 103 (2012) 1–14.

[13] V. Camel, Solid phase extraction of trace elements, Spectrochim. Acta B 58 (2003)1177–1233.

[14] D. Das, U. Gupta, A.K. Das, Recent developments in solid phase extraction in elemen-tal speciation of environmental samples with special reference to aqueous solutions,Trends Anal. Chem. 38 (2012) 163–171.

[15] T. Prasada Rao, R. Kala, S. Daniel, Metal ion-imprinted polymers—novel materialsfor selective recognition of inorganics, Anal. Chim. Acta. 578 (2006) 105–116.

[16] G. Wulff, Molecular imprinting in cross-linked materials with the aid of moleculartemplates — a way towards artificial antibodies, Angew. Chem. Int. Ed. 34 (1995)1812–1832.

[17] A.N. Pustam, S.D. Alexandratos, Engineering selectivity into polymer-supportedreagents for transition metal ion complex formation, React. Funct. Polym. 70(2010) 545–554.

[18] C. Branger, W. Meouche, A. Margaillan, Recent advances on ion-imprinted polymers,React. Funct. Polym. 73 (2013) 859–875.

[19] L.D. Mafu, T.A.M. Msagati, B.B. Mamba, Ion-imprinted polymers for environmentalmonitoring of inorganic pollutants: synthesis, characterization, and applications,Environ. Sci. Pollut. Res. 20 (2013) 790–802.

[20] Y. Liu, X. Chang, D. Yang, Y. Guo, S. Meng, Highly selective determination of inorgan-ic mercury(II) after preconcentration with Hg(II)-imprinted diazoaminobenzene–vinylpyridine copolymers, Anal. Chim. Acta. 538 (2005) 85–91.

[21] Y. Liu, Y. Zai, X. Chang, Y. Guo, S. Meng, F. Feng, Highly selective determination ofmethylmercury with methylmercury-imprinted polymers, Anal. Chim. Acta. 575(2006) 159–165.

[22] S. Büyüktiryaki, R. Say, A. Denizli, A. Ersöz, Mimicking receptor for methylmercurypreconcentration based on ion-imprinting, Talanta 71 (2007) 699–705.

[23] M. Firouzzare, Q. Wang, Synthesis and characterization of a high selectivemercury(II)-imprinted polymer using novel aminothiol monomer, Talanta 101(2012) 261–266.

[24] S. Xu, L. Chen, J. Li, Y. Guan, H. Lu, Novel Hg2+-imprinted polymers based onthymine–Hg2+–thymine interaction for highly selective preconcentration of Hg2+

in water samples, J. Hazard. Mater. 237–238 (2012) 347–354.[25] I. Dakova, I. Karadjova, I. Ivanov, V. Georgieva, B. Evtimova, G. Georgiev, Solid phase

selective separation and preconcentration of Cu(II) by Cu(II)-imprintedpolymethacrylic microbeads, Anal. Chim. Acta. 584 (2007) 196–203.

[26] G. Alberti, V. Amendola, M. Pesavento, R. Biesuz, Beyond the synthesis of novel solidphases: review onmodelling of sorption phenomena, Coord. Chem. Rev. 256 (2012)28–45.

[27] J. He, R. Lv, J. Zhu, K. Lu, Selective solid-phase extraction of dibutyl phthalate fromsoybean milk using molecular imprinted polymers, Anal. Chim. Acta. 661 (2010)215–221.

[28] P. Luliński, D. Maciejewska, Effect of functional monomers and porogens on mor-phology, structure and recognition properties of 2-(4-methoxyphenyl)ethylamineimprinted polymers, Mater. Sci. Eng. C 33 (2013) 1162–1169.

[29] Y. Liu, Z. Liu, Y. Wang, J. Dai, J. Gao, J. Xie, Y. Yan, A surface ion-imprintedmesoporous sorbent for separation and determination of Pb(II) ion by flame atomicabsorption spectrometry, Microchim. Acta 172 (2011) 309–317.

[30] European Union Commission Directive 2009/90/EC of 31 July 2009 laying down,pursuant to Directive 2000/60/EC of the European Parliament and of the Council,technical specifications for chemical analysis and monitoring of water status,Off. J. Eur. Union L201 (2009) 36–38.