Iron(II)-(1,10-phenanthroline) complex imprinted membrane was prepared by ionic imprinting technology. In the first step, Fe(II) established a coordination linkage with 1,10-phenanthroline and functional monomer 2-vinylpyridine (2-VP). Next, the complex was copolymerized with ethylene glycol dimethacrylate (EGDMA) as a crosslinker in the presence of benzoyl peroxide (BPO) as an initiator. Potentiometric chemical sensors were designed by dispersing the iron(II)-imprinted polymer particles in 2-nitrophenyloctyl ether (o-NPOE) plasticizer and then embedded in poly vinyl chloride (PVC) matrix. The sensors showed a Nernstian response for [Fe(phen)3]2+ with limit of detection 3.15 ng mL−1 and a Nernstian slope of 35.7 mV per decade.

1. Introduction

Heavy metals constitute an important class of pollutants that degrade the environment due to their persistent nature and their industrial importance. Among these, iron is one of the most common metals in nature. Infiltration from rainwater in soils and underlying geologic formations dissolves iron, transporting it into groundwater. Large quantities of iron are also disposed in the environment as a result of anthropogenic activities. Iron and its compounds have widespread industrial applications, from constructional material for drinking-water pipes, to food colours, coagulants in water treatment, pigments in paints and plastics. In tap water, dissolved ferrous iron gives a disagreeable taste. An appropriate knowledge of iron levels in both environmental waters and water for human consumption is very desirable. The relevancy of this subject has conducted to numerous reported methods using UV/Vis spectrophotometry [112], chemiluminescence [1317], atomic absorption spectroscopy [1822], and voltammetry [23, 24]. The design and development of portable devices such as sensors rather than laboratory-based instruments in monitoring iron at trace levels in real samples is still of considerable interest.

Imprinted materials in sensors have attracted considerable attention during the past few years [2527]. Molecular imprinting mimics natural receptors with regard to their molecular/ionic recognition. Most investigations of molecular imprinting polymers have been carried out using polymerization in the presence of the template in order to incorporate specific template sites into the polymer. The development of synthetic membranes with molecular imprinting functionality is an important approach for future functional separation or purification materials [28].

Ionic imprinting is a process, in which a functional monomer is allowed to self-assemble around templated ions and subsequently is crosslinked as required. The selectivity of a specific target ion is obtained by providing the polymers with cavities, in which complexing ligands are arranged to match charge, coordination number, coordination geometry and size of that target ion. This process creates the ionic recognition site, which is a specific location for the target ion chemical functionality and spatial arrangement. Recent developments in molecular imprinting have been reviewed [29]. The high selectivity of ion-imprinted polymers (IIPs) arises from the memory effect of the polymer to the imprinted ions for example, from the specificity of interaction of ligand with the metal ions, the coordination geometry and coordination number of metal ions; the charge of the metal ions and to a large extent on the size of them [30]. In recent years, a lot of IIPs have been prepared for Pd2+ [31], Ni2+ [32, 33], Fe3+ [34], Cd2+ [35, 36], Zn2+ [37], Cu2+ [38], Dy [39], UO2 2+ [4042], Ca2+ [43], Er3+ [44], Th4+ [45], Hg2+[46], and Gd3+ [47, 48].

The integration of IIP materials in electrochemical sensors yields selective, sensitive, simple, and rapid analytical methods. Electrochemical transducers monitor the activity of a chemical species in a given matrix and rely on interactions between analyte and receptor. Host-guest interactions are coupled to a signal transduction mechanism that yields the useful information about the species involved. The binding between the analyte and receptor can be effectively monitored and measured by such electrochemical techniques.

In this study, an ion-imprinting polymer was introduced for selective extraction of iron complex from aqueous solution. In synthesis processes, 1,10-phenanthroline, 2-vinylpyridine (2-VP), ethyleneglycol dimethacrylate (EGDMA), and benzoyl peroxide (BPO) were used as Fe(II) complexing agent, monomer, crosslinker and initiator, respectively. The synthesized polymer was dispersed in 2-nitrophenyloctyl ether (NPOE) and embedded in polyvinylchloride (PVC) matrix, for the monitoring of traces of iron.

2. Experimental

2.1. Apparatus

All potential measurements between reference and indicator electrodes were measured by means of a Crison μpH 2002 decimilivoltammeter (±0.1 mV sensitivity). An Orion, 90-00-29, double-junction electrode was used as reference. The analytical output signal was transferred to a commutation point reconnected to one of six ways out, each with an electrical antenna connector for adaptation to electrode device. The selective electrodes had no internal reference solution and an epoxy-graphite matrix as conductive solid contact. All measurements were carried out with the electrochemical cell graphite contact ∣Fe selective membrane∣ test solution ∣∣ Na2SO4 salt bridge ∣∣ Ag/AgCl (3M KCl). The pH values were measured by a Crison CWL/S7 combined glass electrode connected to a decimilivoltammeter Crison, pH meter, GLP 22. All potential measurements were carried out under constant stirring, by a Crison micro ST 2038. Infrared spectra were collected in a Nicolet 6700 FTIR spectrometer. Atomic absorption spectrometric measurements of Fe(II) were made with Perkin-Elmer spectrometer (AAnalyst 200) using the recommended optimum conditions [49].

2.2. Reagents and Solutions

All chemicals were of analytical grade and deionized water (conductivity <0.1 μS cm−1) was employed throughout. 1,10 phenanthroline (phen), o-nitrophenyloctyl ether (o-NPOE), Bis(2-ethylhexyl)sebacate (BEHS) dibutyl phthalate (DBP), cyclodextrin (Fe°Cyc), tetrakis(4-chlorophenyl) borate, sodium tetraphenyl borate (TPB), PVC of high molecular weight, 4-vinyl pyridine (4-VP), and ethyleneglycoldimethacrylate (EGDMA) were purchased from Fluka. Tetrahydrofuran (THF), acetic acid, citric acid, benzoyl peroxide (BPO), and ammonium iron sulfate were obtained from Riedel-deHaën. The evaluation of the effect of pH and other interfering species required sodium hydroxide, hydrochloride acid, cadmium sulphate, potassium sulphate, sodium chloride (all purchased from Merck), nickel chloride, manganese sulfate, magnesium chloride (all from Chemika), barium chloride (Sigma), silver nitrate, and copper sulfate (from Riedel-deHaën).

A solution of 0.1 mol L−1 of 1,10-phenanthroline was prepared in 10% ethanol. A 0.1 mol L−1 stock solution of Fe2+ was prepared by dissolving 4.0 g of iron(II) ammonium sulfate in 3 mL of 0.5 mol L−1 H2SO4, diluted with water in 100 mL calibration flask. A 10−2 mol L−1 stock solution of tris-(1,10-phenanthroline) iron(I1) (ferroin) was prepared by mixing 10 mL of standard 10−1 mol L−1 iron(II) ammonium sulphate, 30 mL of 10−1 mol L−1 l,l0-phenanthroline and 10 mL acetate buffer (pH 4.7). The mixture was transferred into an l00 mL volumetric flask and made up to the mark with water. Standard ferroin solutions (10−3–10−8, mol L−1) were prepared by accurate dilution.

Solutions of the diverse ions used in the interference study were 0.01 mol L−1 in buffer. Hydroxylamine hydrochloride solution (10% w/v) was prepared by dissolving the reagent in water. Acetate buffer solutions of pH 3.5–6.6 were freshly prepared by mixing the appropriate amounts of 1 mol L−1 acetic acid and 1 mol L−1 sodium acetate solutions. All these solutions were stored in polyethylene containers.

2.3. Preparation of Ion Exchanger Sensor

It was carried out by mixing 50 mL of a 1.0 × 10−2 mol L−1 Fe(II) solution with 50 mL of a 1.0 × 10−2 mol L−1 sodium TPB solution or tetrakis(4-chlorophenyl)borate. Resulting solid was isolated by filtration, thoroughly washing with water, and kept in a dark flask inside a desiccator in order to prevent alterations caused by light and humidity.

2.4. Preparation of Metal Complex Imprinted Polymers

In a typical preparation of Fe2+ imprinted polymer (IIP), 1.0 mmo1 1,10-phenanthroline, and 1.0 mmo1 (NH4)2FeSO4·6H2O were weighed, placed into a 18 mm glass test tube and dissolved in 3 mL ethanol. To it 2.0 mmol 4-VP, 20 mmol EGDMA, and 70 mg BPO were added. The polymerization mixture were purged with nitrogen gas for 10 min, sealed, and then heated in a water bath at 60°C for 1 h. The resultant bulk polymer was ground and sieved, collected, and washed with ethanol/acetic acid (5 : 1) (V/V) overnight to remove the template. The reference polymer (NIP) was similarly prepared but without (NH4)2FeSO4·6H2O and 1,10-phenanthroline during the polymerization.

2.5. Constructions and Calibrations of Iron Membrane with Imprinted Material

The sensing membranes were prepared by mixing 200 mg of PVC powder and 7.0 mg of MIP and 2.0 mg of potassium tetrakis[2-chlorophenyl]borate [ISE I], 6.7 mg of MIP only [ISE II], 7.0 mg of NIP, and 2.0 mg of potassium tetrakis[2-chlorophenyl]borate [ISE III], 7.0 mg of NIP and 2.0 mg of potassium tetrakis[2-chlorophenyl]borate [ISE IV], and 2.0 mg of potassium tetrakis[2-chlorophenyl]borate [ISE V], with 400 mg of plasticizer o, NPOE and 200 mg of PVC. The mixture was stirred until the PVC was well moistened and dispersed in 3.0 mL THF. These membranes were placed in conductive supports of conventional or tubular shapes. Membranes were allowed to dry for 24 h and placed in a 1 × 10−3 mol L−1 [Fe(phen)3]2+ solution. The electrodes were stored in these conditions when not in use.

All potentiometric measurements were carried out at room temperature. The emf of each electrode was measured in buffer. Different concentrations were obtained by transferring different aliquots of 1.0 × 10−3 mol L−1 of [Fe(phen)3]2+ ions aqueous solutions to 100 mL beaker containing 50.0 mL of 10−2 mol L−1 acetate buffer of pH 4.7. Potential readings were recorded after stabilization to ±0.2 mV, and emf was plotted as a function of logarithm iron(II) concentration. Calibration graphs were used for subsequent determination of unknown iron(II) concentrations (see Table 1).

2.6. Constructions and Calibrations of Iron Membrane with Ion Exchange Material

Sensor solutions were prepared by dissolving an appropriate amount of sensor in about 62% BEHS or DBP. These were added of 32% of PVC formerly dissolved in 2 mL of THF. Composition of the resulting membranes is presented in Table 1. After application in the conventional support, each membrane was let dry for 24 h.

All electrodes were placed in a convenient support over a magnetic stirrer and immersed in 50.00 mL of IS or buffer solution. Suitable increments of a 2.00 × 10−2 mol L−1 Fe(II) standard solution were added to provide a series of Fe(II) concentrations ranging 4.10 × 10−6 mol L−1 to 3.33 × 10−3 mol L−1. The potential readings of the stirred Fe(II) solutions were measured at room temperature and recorded after stabilization to ±0.1 mV. A calibration plot was constructed connecting logarithm concentration with electromotive force.

3. Results and Discussion

Positioning of metal ions to match the arrangement of ligands on a substrate molecule through preorganization of the substrate with a metal-complexing monomer and its subsequent crosslinking polymerisation is an attractive means to prepare synthetic receptor materials. Furthermore, the strength and specificity of metal-ligand coordination may be enhanced by choosing the suitable functional groups for the IIP.

Despite the potential advantages offered by metal-coordination interaction, relatively few efforts have been taken to use this binding mode to design molecularly imprinted polymers. In this work, a noncovalent molecular imprinting method [50, 51] was used to synthesize a synthetic host for ferroin.

3.1. FTIR Spectra

Figure 1 presents the FTIR spectra of (a) IIP, (b) NIP, (c) FeSO4, and (d) 1,10-phenanthroline. It can be seen that the shape and position of most peaks in (a) and (b) spectra are similar. This was as expected, because both polymers have the same chemical functions. The slight difference between these was related to some ferroin complex sequestered within the imprinted matrix. This is confirmed by (d) spectra; the peak at 853.5 cm−1 should correspond to absorption band of C–H stretching vibration for H atoms adjacent in the pyridine ring in 1,10-phenanthroline compound. The characteristic ring frequencies in 1,10-phenanthrene were shown at approximately 1502 cm−1 the second appearing as a triplet with the centre component at 1585 cm−1 and the third band shifting to 1423 cm−1. The 1423 cm−1 band was most sensitive in this respect, but the other two showed a slightly discernable shift of the order of 10 to 25 cm−1.

3.2. Binding Studies

Adsorption isotherms yield important information concerning binding energies, modes of binding, and site distributions in the interaction of small molecules with adsorbent surfaces. In the liquid phase applications of imprinted materials, a molecule in solution interacts with binding sites in a solid adsorbent. The adsorption isotherms are then simply plots of the equilibrium concentrations of bound ligand (adsorbate) versus the concentration of free ligand. The isotherms can be fitted using various models. The most simple is the Langmuir type adsorption isotherm, where the adsorbent is assumed to contain only one type of site; adsorbate-adsorbate interactions are assumed not to occur, and the system is assumed ideal.

In order to investigate the binding performance of IIP, the equilibrium binding experiments were carried out by varying the concentrations [Fe(phen)3]2+ complex from 0.2 mmol L−1 to 2 mmol L−1 in the presence of a fixed amount (50 mg ) of IIP and the obtained data were plotted with the Scatchard analysis [52] to estimate the binding parameters of IIP where is the maximum apparent binding capacity and the equilibrium dissociation constant. As shown in Figure 2(a), the Scatchard plot is not linear, indicating that the binding sites in either IIP or NIP are heterogeneous in respect to the affinity for [Fe(phen)3]2+. Clearly, within the plot, there are two distinct sections and two straight lines can be obtained from the linear regression. This indicates that the binding sites in the IIP and NIP could be classified into two distinct groups with specific binding properties and of higher affinity binding sites can be calculated to be 152.4 μmol L−1 and 24.5 μmol g−1 dry polymer for IIP and NIP, respectively, from the slope and the intercept of the Scatchard plot (Figure 2(b)). Similarly, and of lower affinity binding sites were 3251.2 μmol L−1 and 129.1 μmol g−1, respectively.

3.3. ISEs Analytical Features
3.3.1. Imprinting Polymer

The dissolution of MIP within the selective membrane may conduct to alterations at the configuration of the imprinted shape. When the template is not extracted from the imprinted polymer, this change in configuration may be attenuated. Results obtained pointed out that template extraction was important, providing decreased limit of detection and higher sensitivity. These results suggest that configuration changes from MIP dissolution were not significant, even for NIP sensors. The synthesized IIP were incorporated into the PVC membrane and were tested as sensing materials in the proposed potentiometric sensor. The potential response obtained with the sensors prepared with [Fe(phen)3]2+ IIP membrane and blank membrane is given in Figure 3. As seen from the figure, the sensors exhibit linear potentiometric response to [Fe(phen)3]2+ ions with lower limit of linear range and ,  mol L−1, and detection limits of 4.45, 6.89, and 17.7 ng mL−1, for ISE’s [II], [IV], and [V], respectively. All sensors exhibit near-Nernstian slopes of (), and  mV decade−1, respectively. Addition of anionic additive potassium tetrakis[2-chlorophenyl]borate to either MIP [ISE I] or NIP [ISE III]-based membrane resulted in a Nernestian pattern with a slope () and () mV decade−1, lower limit of linear range of and  mol L−1, and lower detection limit of 3.15 and 5.12 ng mL−1, respectively. The composition and potentiometric response characteristics of the membrane sensors incorporating MIP and NIP as selective ion recognitions and with/without tetrakis[2-chlorophenyl]borate as anionic additive are shown in Table 1.

3.3.2. Ion Exchanger Sensor

Six membrane compositions were prepared by varying electroactive materials and solvent plasticizers (Table 1). Complexes of iron with phenanthroline, cyclodextrin, or tetrakis(4-chlorophenyl)borate served as ionophores and BEHS or DPB were used as plasticizers. All membranes were of plasticized PVC with 31-32 wt% PVC, 61–63 wt% plasticizing solvent and 3%-4% of anionic additive and 3 wt% of electroactive material.

Electrodes showed supra-Nernstian responses in buffer (Table 1). Slopes ranged 33.3 to 38.6 mV decade−1 and linear behavior was mainly observed from up to  mol L−1. Sensors of Fe°Cyc or Fe°PClPB were greatly dependent on the plasticizing solvent, showing better responses for DBP.

3.4. Effect of PH

The influence of pH on the potentiometric response of the proposed sensors was examined over a pH range of 3–10 for [Fe(phen)3]2+ standard solutions of  mol L−1. The pH of the solution was adjusted with either hydrochloric acid and/or sodium hydroxide solutions. The pH plot shows that the variation of solution pH over the range 3–6 has no significant effect on the potentiometric response for all membrane based sensors.

For the ion exchange membranes the potential versus pH profiles show that the electrodes do not respond to pH changes in the ranges 2.5–6 for sensors IX and XI and pH 4–12.5 for sensors VI, VII, VIII, and X. Only sensors IX and XI showed the effect of the severe precipitation of metals taking place in the alkaline range. All other sensors showed no effect from pH within this range. This is most evidently an abnormal behaviour for which no logical explanation may be found.

Because the solubility and ionisation of Fe(II) were both promoted by acidic solutions, studies under constant pH were carried out in the acidic range.

Evaluation of main operating features for all electrodes under constant pH was carried out in several buffer solutions prepared with different pHs within 2 and 6. Results showed that electrodes provided slopes ranging 67.0 to 202.2 mV decade-1 in citric acid pH 2.5, 33.3 to 40.6 mV decade-1 in citric acid pH 4, and 29.4 to 137.7 mV decade-1 in citric acid pH 6, over a wide concentration range. Best general analytical features for all electrodes were achieved with citrate buffer of pH 4.

3.5. Response Time

The dynamic response times of the sensors were examined by recording the potential readings at time intervals of 10 s over 2 min. The time required to reach 95% of equilibrium was ~20 s for electrodes [I, III, and V] with a recovery time ~30 and ~30 s for electrodes [II and IV] with a recovery time ~50 s. These results indicate that both sensors are amenable for use with automated systems.

3.6. Selectivity Studies

The selectivity of the chemical sensor is one of the most important potentiometric features. One component of the selective membrane exerting great influence upon this property is the electroactive material, as the mechanism of selectivity is mainly based on stereospecificity and electrostatic environment. It is dependent on how much fitting is present between locations of the lipophilicity sites in the two competing species in the bathing solution side and those present in the sensor [53]. The performance of the ferroin sensor in the presence of some cations was assessed by measuring the selectivity coefficient values using the separate solutions method [54]. The results obtained showed no effect for high concentrations (>1000-fold excess) of many common cations such as NH4+, Na+, K+, Ba2+, Ca2+, and Mg2+. These cations do not form complexes with 1,l0-phenanthroline reagent. Metal ions known to form insoluble metal phenanthroline chelates or metal halides (e.g., Pb2+, Hg2+, and Ag+) did not interfere. Metals which form water soluble charged complexes with phenanthroline such as Zn2+, Cu2+, Ni2+, and Cd2+ interfered seriously but interferences caused by these cations were completely circumvented by using suitable masking agents. The response behavior of the MIP and NIP membrane-based sensors towards these complexed cations were presented in Figure 4.

3.7. Analytical Application

In order to access the applicability of the iron selective electrodes, the potentiometric method was applied for the determination of iron in different materials of various natures by formation of ferroin followed by monitoring with the IIP-ferroin sensor. Iron contents (0.05–0.30 mg L−1) of different tap water collected from the laboratory taps and were added of buffer prior to analysis in order to ensure similar background as that of standard solutions were determined. The samples spiked with various standard iron concentrations display results agreed fairly well within ±1.5% with those obtained with the standard spectrophotometric method. Determination of Fe2+ in the presence of large quantities of Fe3+ has received considerable attention in corrosion and environmental studies. Several reports confirmed that the spectrophotometric ferroin method for determining Fe2+ in the presence of excess Fe3+ is not reliable and the recovery of Fe2+ is always high [55].

In the present work, mixtures of Fe2+ and Fe3+ covering the concentration ratios of 1 : 1 to 1 : 50 (Fe2+ : Fe3+) were prepared and treated with l, l0-phenanthroline at pH 4.7. After a reaction time of 1 min, EDTA was added to mask Fe3+, and [Fe(phen)3]2+ was potentiometrically measured with the ferroin sensor. In a second run, the total iron [Fe2+ and Fe3+] was assessed by a prior reduction with hydroxylamine and formation and measurement of ferroin. Results with average recoveries of 98.4% (S.D. = 1.5%) and 95.7% (S.D. = 1.4%) were obtained () for iron(II) and iron(III), respectively.

The proposed sensor was also used to determine iron in pharmaceutical preparations both in batch and in flow conditions. The results obtained with some polyvitamin/mineral tablets and capsules showed an average recovery of 95.5% of the nominal value and a mean standard deviation of ±0.9% (Table 2). These results were checked and confirmed by an independent method from the pharmacopoeia. The results obtained showed a close agreement between the data of the spectrophotometry and ferroin sensor, thus confirming the accuracy of the proposed method.

4. Conclusions

Proposed Fe potentiometric detectors are simple, of low cost, and easy to manipulate. The electrodes based on dispersion of mimic receptor for ferroin in PVC matrix and plasticized in -NPOE might be useful detectors for analysis of environmental waters. They display high selectivity and wide dynamic response range. In general, the overall procedure is precise, accurate, and inexpensive regarding reagent consumption and equipment involved, especially compared to other methods previously reported (Table 3). Considering its routine application, a main advantage arises from composition and quantity of emitted effluents, with small concern in terms of environmental issues. Aside from dilution with buffer, no sample pretreatment or separation steps are required.