A ditopic calix[4]pyrrole amide derivative: highlighting the importance of fundamental studies and the use of NaPh 4 B as additive in the design and applications of mercury( II ) ion selective electrodes

The use of calixpyrroles as cation monitoring systems is very limited. The importance of fundamental studies for providing the basis for the selection of an appropriate ionophore for use in the design of ion selective electrodes cannot be overemphasised. Thus a new ditopic receptor, namely meso -tetramethyl- tetrakis-(4 N , N -diethylacetamide) phenoxymethyl]calix[4]pyrrole (CPA), bearing cation and anion recognition sites has been synthesized and characterized using NMR measurements in acetonitrile. This new receptor interacts with Ca( II ), Sr( II ), Ba( II ), Pb( II ) and Cd( II ) perchlorate salts and Hg( II ) (nitrate as counter-ion). Among the cations tested, 1 : 1: complexes were found between CPA and the above mentioned cations except Hg( II ) for which a 1 : 2 (ligand : metal cation) complex is found. These fundamental studies provided the basis for the design of a Hg( II ) ion selective electrode (ISE). A detailed investigation on possible interactions between the analyte and the individual components of the membrane in solution as well as in membranes prepared in the absence of the receptor, revealed that NaPh 4 B, an additive frequently used in the preparation of mercury ion selective membranes, interacts with Hg( II ). The implications of these ﬁ ndings on previously reported (Hg( II ) ISE) are discussed. The optimum working conditions of the new electrode on its potentiometric response were established. Selectivity coe ﬃ cients of Hg( II ) relative to other cations were calculated. The advantages of this electrode relative to others reported in the literature are discussed. Analytical applications include its use as an indicator electrode as well as for the quantitative determination of this cation in an amalgam.


Introduction
Monitoring water quality is a process that requires sensors activated with materials capable of selectively recognising pollutants.The drive for a mercury(II) monitoring system is based on the following facts: (i) both inorganic as well as organic mercury species are strongly toxic and their presence in the food chain is undesirable.They are likely to be transferred to the global food chain where their concentration increases over time, a process known as bio-magnication, which has been observed for mercury species that result from common practices in industrial processes.(ii) Mercury can be released into the environment from different sources such as chloroalkali, incineration of coal, dental, medical, gold mining, button cell battery manufacturing, other industrial waste and even natural sources (e.g.volcanoes) This means that mercury can be present in different forms as an environmental pollutant, dominantly in the form of inorganic Hg(II), with atmospheric deposition being the dominant source of mercury over most of the landscape.2][3] Therefore, even though methyl-mercury is recognised as the most toxic form of mercury, it is mainly the product of the transformation of Hg(II).So in order to control or reduce the amount of methyl-mercury entering the food chain, it is critical to monitor the amount of Hg(II) in water.
Calixpyrroles or meso-substituted octamethyl porphyrinogens are generally crystalline stable materials which were rst synthesized by Baeyer 4,5 in the 19th century by the condensation reaction involving pyrrole and acetone in the presence of HCl.This work was followed by Brown et al. 6 who proceeded with the production of modied calix [4]pyrroles, while Floriani et al. 7,8 investigated the metallation of such macrocycles.3][14][15][16][17][18][19] With the exception of the work by Floriani 7,8 on the complexation of iron with deprotonated meso octamethyl calix [4]pyrrole, which dates back to the early nineties, no other work can be tracked in the literature concerning the interaction of calix [4]pyrrole and its derivatives with cations.We have previously reported the thermodynamics of complexation of sulphur containing heterocalixpyrroles with the Hg(II) cation. 13,16More recently an ion selective electrode based on one of these ligands has been reported based on research carried out at Surrey. 20ithin this context we have now synthesised and characterised meso-tetramethyl-tetrakis-[(4N,N diethyl acetamide) phenoxymethyl]calix [4]pyrrole, CPA and its complexation properties with a variety of cations, mainly alkali, alkaline-earth and heavy metal cations.Fundamental studies were carried out to gain information on the sites of ligand-ion interaction through 1 H NMR measurements, while the composition of the complexes as well as qualitative information on the stability and therefore the selectivity or hosting ability of the receptor for the cation were obtained from conductance data.The outcome of this investigation was used to explore the capability of the receptor for the design of a Hg(II) ion selective electrode.Optimal working conditions for the potentiometric response of the electrode in terms of its membrane composition and the pH of the aqueous solution were investigated together with the characteristics of this electrode regarding linear working range, detection limit, lifetime and interferences.The advantages of this new electrode relative to those based on ionophores (such as crown ethers derivatives, [21][22][23][24][25][26] calixarene derivatives, [27][28][29] cryptand 22, 30 tetra-aza macrocycles [31][32][33] as well as other compounds [34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53] ) are discussed.
Synthesis of meso-tetramethyl-tetrakis-[(4N,N diethylacetamide)phenoxymethylcalix [4]pyrrole, (CPA) Meso-tetramethyl-tetrakis-(4-hydroxyphenyl)calix [4]pyrrole (1.7 g, 2.26 mmol), potassium carbonate (2.5 g, 18.1 mmol) and 18-crown-6 (0.125 g) were vigorously stirred and reuxed for 1 h in freshly reuxed MeCN (150 ml) under a nitrogen atmosphere.Then N,N-diethyl chloroacetamide (2.46 ml, 17.9 mmol) was added drop-wise and the reaction was reuxed overnight.The reaction was monitored by TLC using DCM/MeOH (9 : 1 mixture) as the developing solvent system.Aer cooling down the reaction, the solvent was removed under vacuum.The solid obtained was re-dissolved in DCM and extracted several times with water to remove the excess of potassium carbonate.The organic phase was separated and dried over anhydrous MgSO 4 , then ltered through a gravimetric funnel using a lter paper.The DCM was evaporated in a rota-evaporator and the oily product was recrystallized in MeCN to give white needle crystals (70% yield).These were dried under vacuum at 80 C.

NMR measurements of the CPA-ion complexes
For these measurements, the macrocycle was dissolved in the deuterated solvent.This technique was used to characterise the ligand, provide information about the ion-ligand interaction and establish the sites of coordination of the ligand upon complexation with the ion. 1 H NMR, 13 C NMR, 13 C DEPT NMR, HSQCDEPT NMR and COSY NMR were recorded.To assess whether the ion complexes with the ligand or not, a known concentration of the ion salt ($8 Â 10 À3 mol dm À3 ) was added to the NMR tube containing a 10 times lower concentration of the ligand, to ensure that the whole ligand has complexed with the appropriate anion or cation salt.
All NMR measurements were recorded at 298 K using a Bruker AC-500 E pulsed Fourier transform NMR spectrometer.Typical operating conditions for 1 H measurements involved a "pulse" or ip angle of 30 , spectral frequency (SF) of 500.25  MHz, delay time of 1.60 s, acquisition time (AQ) of 1.819 s, and line broadening of 0.55 Hz.Solutions of the samples of interest were prepared in the appropriate deuterated solvent, and then placed in 5 mm NMR tubes using TMS as the internal reference.For the 13 C NMR measurements, the operating conditions were as follows: spectral frequency of 75.469MHz, spectral width of 301 ppm, delay time of 0.279 s, acquisition time of 0.721 s and line-broadening of 1.4 Hz with TMS as internal reference.The same NMR equipment was used to establish possible analyteadditive and analyte-plasticizer interactions.

Conductance measurements
For these measurements, a Wayne-Kerr Autobalance Universal Bridge (type B642) was used. 54The conductivity cell constant was determined by the method of Jones and Bradshaw. 55onductometric titrations of the metal-cation salt (as perchlorates salts, except mercury where nitrate was the counter-ion used), with the ligand (CPA) were carried out in MeCN at 298.15 K. To perform these measurements, the conductometric cell was lled with accurately weighed amounts of the cation salt solution prepared in the appropriate solvent (25 ml) and le under stirring to reach thermal equilibrium.Then, a solution of the ligand prepared in the same solvent was stepwise added into the vessel.Aer each addition, stable conductivity readings were taken and molar conductance values were calculated.Molar conductances, L m (Scm 2 mol À1 ) against the ligandcation concentration ratio [L]/[M n+ ], were used to determine the stoichiometry of the complex.

Preparation of the CPA based mercury(II) selective electrode
The Hg(II) selective membrane electrode based on the CPA ionophore was made using a procedure similar to that previously described in the literature. 34,56The membrane electrode with different compositions of CPA (1 and 2%) were prepared by dissolving different amounts of PVC, using DOS or o-NPOE as plasticizers.Anion additives used were NaPh 4 B, or OA, in freshly distilled THF (3 cm 3 ).Membranes without CPA were also prepared to assess possible interactions between the individual components of the membrane and the analyte.The composition ratios were changed within the range most commonly encountered in the literature (1-7% ionophore, 28-33% PVC, 60-69% plasticizer and 0.03-2% ionic additive). 57The resulting mixture was poured to a glass dish (d ¼ 30 mm).The solvent was evaporated off at room temperature.Aer 2 d, a resulting homogeneous PVC membrane (about 0.2 mm thickness) was cut in circles with a diameter of $4 mm and mounted into a PVC tube with a PVC paste (PVC dissolved in THF solvent as adhesive).An internal Ag/AgCl reference electrode was placed in the PVC tube which was lled with a solution of Hg(NO 3 ) 2 (1.0 Â 10 À2 mol dm À3 ) and KCl, (0.1 mol dm À3 ).The membrane electrochemical cell was set up as follows: Ag/AgCl| internal solution (Hg 2+ in KCl) |PVC membrane.The electrode was conditioned for 2-3 d with the same solution of Hg(NO 3 ) 2 (1.0 Â 10 À2 mol dm À3 ) used as the lling internal solution.When the electrode was not in use, it was stored in Hg (NO 3 ) 2 .
Scanning electron microscopic experiments.Micrographs of membrane surfaces were taken by S-3200 N microscope (Oxford instrument) equipped with an X-ray detector for EDAX analysis, a Back Scatter Robinson Detector for imaging, and a Secondary electron detector.
Construction of the Hg(II)-ISE electrochemical cell.A Ag/AgCl with a exible connector (MF-2052, RE-5B) lled with a NaCl aqueous solution (3 mol dm À3 ) was used as an external reference electrode while the silver wire coated with a thin layer of silver chloride was used as the internal electrode.The reference electrode was used in conjunction with the CPA based electrode.
Preparation of standard solutions of the mercury(II) salt.A stock solution of the mercury ion salt (1 Â 10 À2 mol dm À3 ) was freshly prepared by dissolving an accurate amount (0.171 g) of the salt (as nitrate) in deionized water (50 cm 3 ) while working solutions in the range 5.0 Â 10 À8 to 5 Â 10 À2 mol dm À3 were prepared by appropriate dilution of the stock solution with deionized water.
Potential measurements of Hg(II)-ISEs based on CPA.The performance of the proposed CPA based Hg(II)-ISE was investigated by carrying out calibration curves.This procedure was followed in order to determine the composition which gives the best potential response.Different volumes of standard solution of the Hg(II) ion salt were successively added to a vessel containing deionized water (75 cm 3 ) to cover the concentration range from 6.67 Â 10 À8 to 4.32 Â 10 À2 mol dm À3 .The additions of standard solutions were carried out from the low to the high concentration.The potential measurements were made in magnetically stirred solutions and stable potential readings were taken aer a period of time aer each addition.The ion activities of the metal salt solutions were calculated from the ion concentrations according to Debye-Huckel procedure. 24,58Cell potential values E(mV) were plotted as a function of the log 10 of the mercury ionic activity.The response characteristics of the ISE were evaluated from the calibration graph.
Effect of soaking time on the response of Hg(II)-ISE.The effect of soaking time on the performance of the CPA based Hg(II)-ISE was examined.The electrode was soaked in aqueous solution containing Hg(NO 3 ) 2 (1 Â 10 À2 mol dm À3 ) the same as the test solution for different periods of time ranging from 6 to 48 h.The slope values obtained by plotting E(mV) as a function of log 10 [Hg(II)] were also plotted as a function of soaking time.Thereaer, the optimum soaking time for the Hg(II)-ISE was determined.
Effect of pH of the aqueous solution of the Hg(II) salt on the electrode response.The effect of pH of the test solution on the response of the Hg(II)-ISE was examined.The potential was measured in the 2.0-12.0pH range at two concentrations of Hg(II) solutions (1.0 Â 10 À2 and 1.0 Â 10 À3 mol dm À3 ).These experiments were achieved through following the variation of potential of the proposed electrode with the change in the solution pH for both concentrations used; the pH value was adjusted by the addition of a very small volume of HNO 3 or NaOH (each 0.1-1.0mol dm À3 ).The potential readings were plotted against the corresponding pH values.The optimum pH values obtained at both concentrations of Hg(II) solutions (1.0 Â 10 À2 and 1.0 Â 10 À3 mol dm À3 ) were used for further studies.
Determination of selectivity coefficients of ions relative to Hg(II).

In this work, potentiometric selectivity coefficients (K pot
Hg,M ) of Hg(II) relative to other cations were determined by using two IUPAC recommended methods, namely, the separate solution method (SSM) 59 and the matched potential method (MPM). 60hese two methods were carried out under the following experimental conditions: for the SSM method, the concentration of the Hg(NO 3 ) 2 (primary ion) was 2.23 Â 10 À4 mol dm À3 ; the same concentration was used for the interfering ions.For the MPM method, the concentration of Hg(NO 3 ) 2 was 1.0 Â 10 À2 mol dm À3 while 1.0 Â 10 À1 mol dm À3 was the concentration of interfering ions.
Assessment of the dynamic response time of the Hg(II)-ISE.Measurements of the dynamic response time of the electrode were conducted at different Hg(NO 3 ) 2 concentrations (1 Â 10 À7 , 1 Â 10 À6 , 1 Â 10 À5 , 1 Â 10 À4 , 1 Â 10 À3 and 1 Â 10 À2 mol dm À3 ).Potentials resulting from the addition of each concentration of the primary ion (Hg(II)) solution (50 cm 3 ) were plotted against dynamic response times of the electrode.The reversibility of the electrode was investigated by successive immersions of the Hg(II) ISE at two different concentrations of the Hg(II) salt solution from the lower to the higher concentration and vice versa.Finally, the potential values against response times were plotted.
Determination of the lifetime of the Hg(II)-ISE.The lifetime of the Hg(II)-ISE was monitored over a period of time by measuring the potential values of the calibration solutions of Hg(NO 3 ) 2 .Calibration curves were obtained aer each examined time in the 6.67 Â 10 À8 to 4.32 Â 10 À2 mol dm À3 concentration range.The slope values obtained by plotting E(mV) as a function of log [Hg(II)] were also plotted against the period of time tested.
Analytical applications of the CPA based Hg(II)-ISE Determination of Hg(II) ion concentration with EDTA in aqueous solution.The Hg(II)-ISEs (1%) with o-NPOE (68%) as plasticizer and OA as ionic additive (1%) were used as an indicator electrode in the potentiometric titration of the mercury(II) ion salt with EDTA at the standard temperature.The aqueous Hg(II) nitrate solution (60 cm 3 , 1.0 Â 10 À3 mol dm À3 ) was titrated with aliquots of EDTA (0.5 cm 3 , 5.0 Â 10 À3 mol dm À3 ).The amount of mercury(II) ion was accurately determined from the end point of the resulting titration by plotting E(mV) as a function of log [Hg(II)].
Analysis of mercury(II) in a real sample.The proposed Hg(II)-ISE as indicator electrode was tested for the potentiometric determination of mercury(II) in a dental amalgam lling.Thus an accurate weight of the dental amalgam alloy (43.1% Ag, 26.1% Cu and 30.8%Sn) (0.6 g) was dissolved in a minimum volume (20 ml) of HNO 3 (60%).The solution was intensely heated to near dryness.The acid treatment and the evaporation process of the solution were repeated at least three times.The residue obtained was dissolved in a small amount of water and ltered.The clear solution was quantitatively transferred into a volumetric ask (50 cm 3 ) and lled to the mark with distilled water aer adjusting the pH value of the solution to 5.6.
Thereaer, the standard addition method was applied where small increments (0.1 cm 3 ) of the standard solution of Hg(NO 3 ) 2 (3.0 Â 10 À2 mol dm À3 ) were added to the dental amalgam solution (50 cm 3 ) at the standard temperature.The changes in potential readings were recorded aer each addition of the standard solution.The Hg(II) concentration in the dental amalgam solution was calculated.The results obtained with the CPA based electrode were compared with data obtained by atomic absorption spectrometry (AAS).

Results & discussion
For the structural characterization of CPA, NMR experiments were carried out.In a rst stage 1D 1 H NMR proton and 13 C NMR were run in CD 3 CN (Chemical shis shown in the Experimental section).However the ID 1 H NMR was not sufficient to attribute the protons of the different units, which become non-equivalent because of the loss of symmetry aer substitution, therefore we proceeded with 13 C NMR-HBBD, 13 C NMR-DEPT and HMQC 2D NMR experiments to establish a correlation of 1 H and 13 C signals.Comparative analysis of 13 C NMR-HBBD and 13 C NMR-DEPT spectra given in ESI † help to identify the number of signals attributed to ve quaternary carbons (one sp 3 was deduced, which was found to be in accord with the elemental analysis.However, the complete assignments, especially in the overlapped regions ( 1 H at 3.35 ppm) had to be made from the 2D data.The 1 H- 13 C correlation (ESI †) was used for the assignment of hydrogens to carbons in a straight forward manner.The set of protons at 3.35 ppm showed correlation with two carbons C-11 (40.56 ppm) and C-11 0 (39.27 ppm), this indicates clearly that this set of protons integrates the two methylene of the acetamide arm group (2CH 2 -N).In addition the correlation of C-6 and C-7 to H-4 and H-5 benzene ring protons made the assignment of the pyrrole protons easy.
Having characterised the receptor, 1 H NMR measurements were carried out to assess cation-ligand interactions.These were carried out in CD 3 CN given that in this dipolar aprotic solvent ion salts are predominantly dissociated.As far as the alkaline-earth metal cations are concerned, downeld chemical shi changes relative to the free ligand (ppm) were observed by the addition of Mg(II) (H-5, 014), Ca(II) (H-4, 0.15, H-5, 0.16), Sr(II) (H-4, 0.10, H-5, 0.15) and Ba(II) (H-5, 0.13) salts to CPA in CD 3 CN.Therefore, it can be concluded that these cations being of the hard Lewis acid type may interact with CPA through the phenolic and the carbonyl oxygens of the amide moiety (hard Lewis base). 61As far as the heavy metal cations are concerned, the most signicant downeld chemical shi changes for Pb(II) are observed in H-5 (0.21 ppm) and H-6 (0.20 ppm) while for Cd(II) the signal corresponding to H-6 disappeared.The addition of Hg(II) to the NMR tube containing a solution of CPA in CD 3 CN led to signicant downeld chemical shi changes in the methylene protons (H-6, 0.16 ppm) between the phenolic oxygen and the carbonyl group of the acetamide arm group and the one next to the nitrogen atom of the acetamide nitrogen (H-7, 0.14 ppm, H-7 0 , 0.16 ppm) as well as in the methyl terminal protons of acetamide arm (H-8 0 , 0.13 ppm).Upeld chemical shi changes observed in the NH proton of the pyrrole units were found to be less signicant and may result from the conformational changes of CPA upon complexation.These results suggest that the interaction of Hg(II) with CPA takes place through the amide functional group.We therefore conclude that Pb(II), Cd(II) and Hg(II) interact with the donor atoms of the acetamide functional groups.
Among the anions tested (halides, nitrate, hydrogen sulphate), dihydrogen phosphate, triuoromethane sulphonate, thiocyanide strong interaction was found only with the uoride ion where the most remarkable downeld chemical shi change was found for H-1 (4.52 ppm) as a result of the interaction of this anion with the NH functional groups of the pyrrole units.Preliminary studies in different dipolar aprotic solvents revealed a change in the composition of the anion complex in moving from one solvent to another.Therefore 1 H NMR titrations were carried out in three solvents: CD 3 CN, d 6 -DMSO and d 7 -DMF at 298 K. Volumes of TBAF were added to the solution of CPA.The 1 H NMR spectrum (CD 3 CN, 500 MHz, 298 K) of CPA receptor showed one broad singlet at 7.98 ppm (integrating 4H) which was assigned to the four chemically equivalent NH protons' resonance.Addition of 0.5 equivalent of tetrabutylammonium uoride (TBAF) to a solution of the receptor leads to an equilibrium favoring the cone-like conformation via the formation of the 1 : 1 uoride-CPA complex.The conformational change of the receptor was easily observed in the 1 H NMR spectrum, where two different signals were observed for the NH protons, one at 7.98 ppm, corresponding to the free receptor, and the other at 12.00 ppm, corresponding to the uoride complex.This nding was taken as evidence that the exchange between the complex and the free receptor is slow on the 1 H NMR timescale.
Aer the addition of 1.0 equivalent of TBAF, only the signals at 12 and 12.5 ppm, corresponding to the CPAF À complex were observed in the 1 H NMR spectrum under these conditions, as expected for a highly stable complex.
Surprisingly, the pyrrole NH resonance was found to split into two peaks in the presence of uoride.It was considered that a likely cause for this splitting might be due to coupling between the NH protons and the 19 F nucleus of the bound uoride anion.The coupling constant was found to be 49.5 Hz as was previously reported in the literature. 62The stoichiometry of complexation was determined from plots of the integrals of the most signicant downeld chemical shi changes in the protons of the anion complex (H-1, H-2, H-3, H-6) as shown in Fig. 1(a).
The results obtained from the 1 H NMR titration of CPA with the uoride anion salt in d 6 -DMSO and d 7 -DMF (Fig. 1(b) and (c)) clearly indicate that in the former solvent, CPA interacts with the anion with the formation of a 2 : 1 (CPA : F À ) complex in the latter solvent the composition of the complex is 1 : 1. 1 H NMR data for the free ligand in these solvents show that DMSO interacts with calixpyrrole through hydrogen bond formation between the basic oxygen atoms of the solvent and the NH moiety of the pyrrole ring.This was reected in the signicant downeld chemical shi changes observed in H-1 in moving from CD 3 CN to d 6 -DMSO (1.47 ppm) while insignicant upeld chemical shi changes were observed for the remaining protons.To a lesser extent d 7 -DMF also interacts with CPA.A downeld shi in H-1 was observed (1.31 ppm) with additional ones in H-4 (0.10 ppm) and H-6 (0.11 ppm).Sessler and coworkers 63 reported the X-ray structure of the parent calix [4]  pyrrole in a 1,2 alternate conformation in which two molecules of d 7 -DMF are hydrogen bonded to two adjacent pyrrole units.However these authors found that in solution (benzene) only 1 : 1 complexes are formed.The same situation was found in DMSO.Apparent stability constants in benzene for the parent calix [4]pyrrole with d 6 -DMSO (log K s ¼ 1.21) and with d 7 DMF (log K s ¼ 1.05) were reported.As far as CPA is concerned, a possible explanation of the different compositions of the uoride complexes in these solvents must be attributed to their different molecular geometries.Thus DMSO is characterized by a pyramidal geometry (sp 3 hybridization with a lone pair of electrons) while DMF has a planar structure (sp 2 hybridization).It is also known that DMSO can be associated with the formation of head to tail dimers through dipole-dipole interactions. 64,65As a result the binding sites of the receptor to enter interaction with the anion are reduced.Therefore two calixpyrrole units rather than one are required to enhance the number of binding sites required to enter complexation with the anion in d 6 -DMSO.
In an attempt to test the ability of this receptor to interact simultaneously with Hg(II) and F À , the binding properties of CPA with Hg(II) were examined by 1 H NMR titration experiments in CD 3 CN in the absence and the presence of F À (Fig. 2).
From the results of this experiment the following conclusions can be drawn.As stated above, the addition of 1 equivalent of TBAF to CPA in CD 3 CN caused a remarkable downeld chemical shi change of the NH protons and upeld shis of the remaining protons (H-2; H-3; H-4,5; H-6, H-7,7 0 and H-8,8 0 ) of this receptor.Further addition of TBAF does not make any signicant change in the chemical shis of the protons, suggesting the formation of a 1 : 1 (CPA : F À ) complex.In the rst three additions of the Hg(II) salt solution to the uoride complex a substantial upeld chemical shi change of the NH protons of the complex ($3.6 ppm) was observed.Inspection of Fig. 2(a) shows the formation of a 2 : 1 (Hg(II) : CPA) complex.The integrals of the NH protons of the 1 : 1 uoride complex with the Hg(II) salt is shown in Fig. 2(b).The break observed in this gure at the Hg(II) : CPA ratio of 0.5 clearly indicates that two units of the uoride complex interacts with only one metal cation.These may be the result of an exchange between the two TBA + cations with the bivalent mercury ion.It is reasonable to assume that the TBA + cations are large hydrophobic ions of low charge density relative to Hg(II) and therefore these are likely to be easily displaced by the latter cation.In addition, ion-ion interactions are stronger than ion dipole interactions.Consequently Hg(II) will be more attracted by the negatively charged uoride complex than by the donor atoms of the amide functionality of the receptor.

Conductometric titrations of ions with CPA in MeCN
Conductometric titrations of all metal cation salts (perchlorate as the counter-ion) investigated, except Hg(II) (as nitrate), led to a dened change in the curvature at the 1 : 1 (ligand : metal ion) stoichiometry, suggesting the formation of complexes of moderate stability.As far as Hg(II) is concerned.The conductometric titration curve (Fig. 3), which is a plot of the molar conductance, L m against the ligand : metal cation ratio breaks at 0.5 indicating that two metal cations interact with one unit of receptor.A decrease in conductance due to the formation of the 1 : 2 complex (eqn (1)) is observed   In the presence of an excess of the ligand, it seems that one of the mercury cations in the complex is transported to the free ligand with the formation of a 1 : 1 complex (eqn (2)) as shown by the increase in conductance.
As far as the F À anion is concerned, the conductometric titration curve consisted of a combination of two linear segments with a well-dened change in the curvature at the 1 : 1 (CPA : F À ) molar ratio, suggesting the formation of a relatively strong complex.
Based on the evidence obtained from the fundamental studies described above, it is concluded that CPA has a higher hosting ability for Hg(II) relative to other ions investigated, therefore a Hg(II) selective liquid membrane electrode based on this receptor was designed.Before discussing the characteristics of the ISE we assessed whether or not interactions are found between the analyte and the individual components of the PVC based membrane (additive and plasticizer) in the absence of the receptor.

Selection of additive and plasticizer for the preparation of the Hg(II) ion selective membrane
Two additives (sodium tetraphenylboron, NaPh 4 B and oleic acid, OA) were tested to nd out whether or not these interact with the analyte and interfere in the selective behaviour of the ionophore for the analyte.Thus 1 H NMR data obtained by the addition of Hg(II) (nitrate as counter-ion) shown in Table 1.
These results clearly indicate that there is a strong interaction between the cation and this additive.This is reected in the signicant downeld chemical shi changes observed in all the aromatic protons.In an attempt to observe the behaviour of NaPh 4 B, this salt was introduced in a PVC based membrane in the absence of the receptor.Potentiometric measurements carried out demonstrated that the membrane responds to Hg(II) showing a Nernstian behaviour (slope 29.1 AE 0.6 mV/pHg(II) in the 3.25 Â 10 À5 to 2.5 Â 10 À3 mol dm À3 concentration range with a response time of about 10 s).This was further corroborated by Scanning Electron Microscopy (SEM) carried out on the membrane with the aim of identifying the locations of the Hg(II) cation on the PVC membrane in the absence of the receptor and to understand the changes to the membrane morphology.Fig. 4(a1) and (a2) represent the SEM micrographs of the membrane surface before and aer it has been used in the Hg(II) ISE.
Fig. 4(a1) shows the typical clear and brous PVC membrane surface without exhibiting any dense areas or defects.The Fig. 4(a2) shows the Hg(II) as white dense patches on the membrane which have been detected in the EDAX spectrum in Fig. 4(b2), where the new peaks characteristic of Hg (compared to Fig. 4(b1) are noted and this indicates the binding of Hg(II) to the Ph 4 B À anion sites).A change in the morphology of the membrane as a function of the ageing time was also observed.It is therefore concluded that a PVC membrane containing NaPh 4 B is able to detect Hg(II) in the absence of the CPA receptor.These ndings as demonstrated below have implications on the working characteristics of the electrode in the presence of the receptor and open serious questions about whether or not the electrode responses that have been previously reported with this additive were due to the receptor or to the presence of NaPh 4 B particularly when fundamental studies have not been carried out.No interaction was found between the mercury(II) cation and oleic acid by 1 H NMR measurements.In addition, blank experiments in the absence of the receptor were carried out.It was found that when the membrane components were 68% o-NPOE, 30% PVC, and 1% OA, the electrode membrane displayed no response towards Hg(II) and therefore this was the additive selected for the preparation of the membrane containing the CPA receptor in the below experiments.In the same way, we tested two plasticizers, DOS and o-NPOE.Two PVC membranes were produced using OA as additive and either DOS or o-NPOE.No interaction was detected with the DOS plasticizer, in the concentration range 1.3 Â 10 À3 to 1.5 Â 10 À2 mol dm À3 , and with O-NOPE in the concentration range of 7.1 Â 10 À4 to 1.34 Â 10 À3 mol dm À3 .

Effect of membrane composition on the potentiometric response characteristics of the CPA based Hg(II)-ISE
The characteristics of the electrode such as its Nernstian behaviour, liner range (L-R), detection limit (D-L) and response time were investigated using ve electrode membranes (E1-E5, composition electrode in Table 2) of varying compositions: two different plasticizers (o-NPOE or DOS) (E1 and E2) in the absence of additive (E3), in the presence of an increased amount of the receptor (E4).Also included data obtained in the presence of the receptor when NaPh 4 B was used an additive (E5).Calibration curves for the ve electrodes tested are shown in Fig. 5.
The nature of plasticizers used (o-NPOE and DOS) (E1 and E2) effects the characteristics of the Hg(II)-ISE.The results showed that the latter (E2) had a better response when compared with the former (E1).The linear working response was within a wide range of Hg(II) concentration from 1.81 Â 10 À7 to 1.78 Â 10 À3 mol dm À3 with a detection limit of 6.0 Â 10 À8 mol dm À3 and response time in about 10 s.In addition, the slope obtained for this electrode (E2) was found to be 30 AE 1 mV dec À1 , which is close to the Nernstian response (29.5 mV dec À1 at 298 K).The difference in the response characteristics between E1 and E2 may be attributed to the dielectric constant of the plasticizers used which in turn may have an effect on the membrane permittivity and the mobility of the ionophore molecules. 24Accordingly, o-NPOE was chosen to be the optimum plasticizer for further studies with the CPA based electrode.Clearly in the absence of additive (E3) the response of the electrode in the presence of the receptor is indeed very weak.
Comparison of the potentiometric response characteristics of the membrane electrode E4 with respect to E2 revealed that the best performance was observed for the latter.The increase of the amount of ionophore led to a lower electrode response.This may be also ascribed to a decrease in the solubility of the ionophore in the membrane.As previously stated 58 the ionophore should be soluble in the membrane matrix and should be sufficiently lipophilic to avoid leaching from the membrane into the sample solution. 24As far as E5 is concerned the data in Table 2 show that the presence of NaPh 4 B decreases the detection limit of the electrode by a factor higher than 10 which may be due to a possible interaction between this salt and the receptor.It is well established that the parent calix [4]pyrrole does not complex the tetraphenyl borate anion, but the possibility exists that its derivative containing aromatic group interacts with this anion through pi-pi interactions.From the above mentioned results, the membrane electrode E2 was chosen for further investigation under different experimental conditions.
The surface analysis of the PVC membrane casted with THF containing CPA, OA as additive and o-NPOE as plasticizer was characterised by SEM coupled with EDAX.Fig. 6 As can be seen from Fig. 8, the pH effect on the response of the electrode was investigated under the same conditions stated above over a 1.5-12.0pH range of aqueous solutions of Hg(NO 3 ) 2 (1.01 Â 10 À3 and 1.01 Â 10 À2 mol dm À3 ).The results show that the electrode potential remains constant in the 4.3-8.5 pH range using 1.01 Â 10 À3 mol dm À3 Hg(II) solution.At pH lower than 4.30, the increase of the acid nature of the solution gives rise to an increase in the electrode potential.At pH higher than 8.5, the potential values were found to decrease sharply due to the formation of hydroxyl-complexes of Hg(II) in solution.By altering the concentration of the mercury salt to 1.01 Â 0 À2 mol dm À3 , a similar prole was observed.Therefore for further studies, an aqueous solution of Hg(NO 3 ) 2 at pH 5.6 was used.

Effect of the concentration of the internal solution on the response of Hg(II)-ISE based on CPA
The concentration of the internal reference solution has a remarkable inuence on the electrode response (as shown in Fig. 9).The optimum concentration of Hg(NO 3 ) 2 was found to be 1 Â 10 À2 mol dm À3 .A slope value of 29 AE 1 mV dec À1 close to the theoretical Nernstian slope was found.However when the concentration was decreased to 1 Â 10 À3 mol dm À3 Hg(NO 3 ) 2 the electrolyte was unable to transfer the charge from the detected ion in the ISE membrane to the measuring system and as a result the electrode response decreased.Therefore, the concentration of 1 Â 10 À2 mol dm À3 Hg(NO 3 ) 2 was used for further studies reported in this paper.2.
the theoretical Nernstian slope (29.5 mV dec À1 ).Aer longer periods of soaking, the Hg(II)-ISE did not show further improvements in response.Therefore, it can be seen that the conditioning for the proposed (Hg(II)-ISE) is a necessary step to increase the number of active sites of the ionophore used.

Response time and reversibility for Hg(II)-ISE based on CPA
The response time for this ISE (E2) was tested with Hg(NO 3 ) 2 solutions by changing from lower to higher concentrations within a range of 1 Â 10 À7 to 1 Â 10 À2 mol dm À3 .The results (Fig. 11) showed that the response time of the CPA based Hg(II)-ISE was 10 s.The reversibility of the electrode was examined by alternatively dipping it in aqueous solutions of Hg(NO 3 ) 2 at two different concentrations (1 Â 10 À4 and 1 Â 10 À3 mol dm À3 ).Results are shown in Fig. 12.It can be observed that the proposed electrode exhibited a satisfactory reversibility although the time required for attaining a stable potential (or equilibrium state) by changing from a low to a high concentration was longer than when changing from a higher to a lower concentration.Therefore, the response time was found to be about of 8 s when the concentration of Hg(NO 3 ) 2 solution was changed from 1.0 Â 10 À3 to 1.0 Â 10 À4 mol dm À3 , but the response time increased to 13 s on the switch from 1.0 Â 10 À4 to 1.0 Â 10 À3 mol dm À3 .

Lifetime of the CPA based Hg(II)-ISE
The lifetime of the ISE was monitored with time as shown in Fig. 13.It was found that this electrode exhibited a good stability in the response up to a period of 68 d.Indeed the mean dri of slope observed over this period was found to be less than 0.02 mV dec À1 .Therefore, it can be concluded that the lifetime of the CPA based Hg(II)-ISE is 68 d.

Potentiometric selectivity coefficients of the CPA based Hg(II) ion selective electrode
The potentiometric selectivity coefficients of the CPA based Hg(II)-ISE were determined for a variety of interfering metal ions such as Na(I), K(I), Ag(I), Mg(II), Ca(II), Ba(II), Sr(II), Mn(II), Co(II), Ni(II), Cu(II), Zn(II), Cd(II), Pb(II), Fe(III) and Al(III) using two methods: the Separate Solution Method (SSM) 59 and the Matched Potential Method (MPM). 60The values of selectivity coefficients obtained under optimum experimental conditions are given in Table 3.Although the trend observed is the same for data derived from these two methods, the values differ given that different assumptions are made in these methods.
It can be seen from the log K plot Hg 2þ ;M nþ values (by both methods) that the CPA based Hg(II) selective electrode is more selective towards mercury over all of the interfering ions    However, the log K plot Hg 2þ ;Ag þ values for the Ag + interfering ion obtained from the SSM and MPM methods were À0.46 and À0.99 respectively which does not seem to be a strong interference.The reliability of the electrode as an analytical tool was tested and this is now discussed.

Analytical applications of the CPA based Hg(II)-ISE
The CPA based Hg(II) ISE was used as (i) an indicator electrode for the potentiometric titration of Hg(II) with EDTA and (ii) the determination of mercury in an amalgam sample.Thus Fig. 14 shows the potentiometric titration curve resulting from the titration of an aqueous solution of Hg (NO 3 ) 2 (1 Â 10 À3 mol dm À3 ) with EDTA (0.005 mol dm À3 ) using the optimized electrode.A sharp end point was obtained.
Table 4 shows the results obtained for the determination of mercury(II) in a lling dental amalgam together with those obtained by atomic absorption spectrometry.An excellent agreement is found between the two methods.Furthermore, the species such as Cu(II), Zn(II) and Sn(II) present in the amalgam samples have no interfering effect.Therefore, it was concluded that the CPA based Hg(II)-selective electrode could be successfully employed for the quantitative determination of Hg in a real dental amalgam sample.

Advantages of the CPA based ISE relative to existing ionophore based Hg(II) electrodes
A great deal of the published material involves the use of the NaPh 4 B (or related salts) as an additive, [20][21][22][23][24][25]28,29,31,[34][35][36][37][38][39]41,43,44,[46][47][48][49][50][51][52][66][67][68][69][70][71][72][73][74][75] which as shown in this paper interacts with Hg(II) and this has implications on the characteristics reported for these electrodes. In the absence of fundamental studi, the response of the electrode may result from the response of the additive rather than the ionophore (this may be occurring in the prior studies mentioned).Other ISEs for the detection of Hg(II) used plasticisers other than o-NPOE or DOS.40,42,45,53,76 Given that a plasticizer may have an effect on the potential response of the electrode a comparison between the research carried out with   ISE in which the same additive and plasticiser have been used is that reported by Somer and co-workers 77 using tridodecyl methyl ammonium iodide as the ionophore.However the linear range for this electrode (5 Â 10 À5 to 1 Â 10 À1 mol dm À3 ) with a DL of 1 Â 10 À5 mol dm À3 is less sensitive than the one reported in this paper.

Final remarks
Despite the efforts being made worldwide on the design of Hg(II) ISEs by the use of cyclic and acyclic ligands a comprehensive review at the state of the art in this particular area of research leads to the conclusion that an important drawback is the lack of a link between fundamental studies and their applications.This is clearly demonstrated in this paper where we show that the currently (and almost universally used) tetraphenyl borate additive salts (used in the design of Hg ISEs incorporating PVC membranes) act by themselves as selective detectors for this cation.It may well be that in some cases the additive and receptor can have a cooperative effect (offering an improvement in the characteristics of the electrode) but this need to be   investigated.Quite clearly this is not the case in the system investigated in this paper.
While calix [4]pyrroles and derivatives have been extensively used for the design of ISEs and chemically modied electrodes, most efforts have been directed to anions rather than cations. 78owever the possibility of altering the selectivity through structural modication (as demonstrated in this paper) opens the way for further developments on the use of these receptors for the design of cation selective electrodes.Further studies are being carried out to assess this electrode for monitoring uoride in water as well as its selectivity in the presence of mercury(II) uoride.
and four sp 2 , including a carbonyl function at 166.01 ppm and another one integrating two carbons [C 1 -C 4 and C]O]), three methine [(CH) 3 all sp 2 all integrating two carbons each], three methylene one oxygenated and two link to a nitrogen atom [(CH 2 ) 2 N(CH 2 -O)] and three methyl [(CH 3 ) 3 ].The assignment of the N-H functionality was made directly from the 1 H NMR. Consequently the formula [C1-(C]) 4 -(CO)-

1 H
NMR measurements of CPA with metal cations and anions in CD 3 CN at 298 K Hardly any chemical shi changes (Dd ppm) were observed in the 1 H NMR spectrum of CPA by the addition of alkali metal cations (Li(I), Na(I), K(I), Rb(I) and Cs(I)) as perchlorates in CD 3 CN at 298 K.The addition of Ag(I) salt to CPA also did not lead to signicant chemical shi changes in any of the protons.

1 H
NMR titration of CPA with the uoride anion in deuterated dipolar aprotic solvents In an effort to study the behavior of the receptor in solution, 1 H NMR studies of CPA were carried out in CD 3 CN, d 6 -DMSO and d 7 -DMF in the absence and presence of the uoride anion.

Fig. 1
Fig. 1 (a) 1 H NMR titration of CPA with the fluoride anion in CD 3 CN at 298 K. (b) 1 H NMR titration of CPA with the fluoride anion in d 6 -DMSO at 298 K. (c) 1 H NMR titration of CPA with the fluoride anion in d 7 -DMF at 298 K.

Fig. 2
Fig. 2 Shifts and integrals of NH protons of the CPA/F (1 : 1) complex titrated with Hg(II) in CD 3 CN at 298 K.

Fig. 4 (
Fig. 4 (a1 & a2) PVC membrane (a1) prepared before being used in the electrode and (a2) after it has been used for several times in the ion selective electrode; (b1 & b2) EDAX spectra of the membranes showing the elemental analysis.
(a) and (b) and 7(a) and (b) show SEM-EDAX of the PVC-CPA matrix before and aer it has been in contact with Hg (NO 3 ) 2 .A detectable transformation in the membrane shows the interaction with the cation.The EDAX analysis displays Hg(II) peaks, and this depicts the selective recognition and binding of CPA to the Hg(II) cation.Effect of pH of the aqueous solution on the response of the CPA based Hg(II)-ISE Fig. 10 shows the effect of conditioning of the ISE for different periods of time on the slope values obtained by plotting E(mV) as a function of log[Hg(II)].These results revealed that the soaking time of the ISE has a considerable effect on its response characteristics.It was found that an optimum soaking time was 24 hours, the slope value (27 AE 2 mV dec À1 ) obtained is close to

Fig. 5
Fig. 5 Effect of membrane composition on the potentiometric response characteristics of the CPA based Hg(II)-ISE.Membrane compositions are in Table2.

Fig. 6
Fig. 6 (a) Scanning electron microscopy image of a CPA ionophore-PVC porous membrane; E2 (b) EDAX spectrum showing the membrane elemental composition.

Fig. 7
Fig. 7 (a) SEM of CPA ionophore-PVC membrane (E2) after it has been used in the ISE for Hg(II) selectivity; (b) EDAX analysis of the membrane showing the presence of Hg(II).

Fig. 8
Fig. 8 Effect of the pH on the response of the CPA based electrode (E2).

Fig. 9
Fig. 9 Effect of internal solution concentration on the response of the ISE (E2).
studied.Also log K plot Hg 2þ ;M nþ values for most ions examined except for the Ag(I) ion were of the order of À1.38 or smaller.

Fig. 10
Fig. 10 Effect of soaking time on the Nernstian behaviour of the Hg ISE at 298 K.

Fig. 11
Fig. 11 Dynamic response of the CPA based Hg(II)-ISEs for step changes in mercury(II) concentration at 298 K.

aFig. 14
Fig. 14 Potentiometric titration of Hg(NO 3 ) 2 with EDTA using the CPA based electrode as indicator electrode at 298 K.

Table 1 1
H NMR data for NaPh 4 B and its interaction with the Hg(II) cation in D 2 O at 298 K

Table 2
Composition of membranes and response characteristics of the CPA based ISE for Hg(II) at 298 K

Table 3
Selectivity coefficients of metal cations relative to mercury(II) at 298 K using two methods (for ISE E2)

Table 4
Application of Hg(II)-ISEs based on the ionophores CPA determination of mercury ion a filling dental amalgam alloy sample mercury using an amalgam as a real sample