Thermodynamics, a suitable reporter in the design of mercury (II) ion selective electrodes

2.5.1 Hg (II)-ISEs cell assembly

The potential measurements (EMF, mV) of the fabricated Hg (II)-ISEs were recorded in a Fluke 73 Series III Digital Multimeter. The apparatus reads over a wide measurement range (1000 V) with a good precision. The multimeter was conjugated to an external reference electrode (silver/silver chloride that is equipped with a Flexible Connector (MF-2052, RE-5B) filled with sodium chloride solution (3 mol dm⁻³) and the designed internal reference electrode was a silver wire (6 cm × 0.15 mm) coated with a thin layer of silver chloride by an electroplating. The electrochemical cell assembly was as follows, Ag/AgCl/internal solution (1.0 × 10⁻² mol dm⁻³ Hg(II)(nitrate as counter-ion) in 0.1 M KCl)/PVC membrane/test solution/KCl (std.)/Ag/AgCl.

The procedure used in the fabrication of mercury (II) selective membrane electrodes was previously described in the literature (Meier, 1982; Gale, 2000). Composition of PVC based membranes used in this work are summarised in Table 1.

2.5.2 Potential measurements of the Hg (II)-ISEs based on 1 and 2

In order to obtain the optimal potential response, the performance of the developed Hg (II)-ISEs was determined by the potential recorded across the prepared membrane during the addition of mercury nitrate solutions to a container with de-ionised water (75 cm³) at room temperature under a constant stirring rate (180 rpm). The analyte solutions were prepared in de-ionised water in the concentration range of 6.67 × 10⁻⁸ to 4.32 × 10⁻² mol dm⁻³. Addition of the prepared standard solutions was made every three minutes from low to high concentrations, where emf values (mV) were recorded following each addition. Then the calibration curve was constructed by plotting the variation of the recorded potential vs the logarithm of the Hg(II) activity (log a_Hg^II).

2.5.3 Effects of soaking, pH, dynamic response time and life-span on the Hg (II)-ISEs potential

The developed electrodes were soaked in aqueous solution of Hg(NO₃)₂ (1 × 10⁻² mol dm⁻³) at room temperature for different periods of time (6, 12, 24, 30, 36 and 48 h). Calibration curves were established following each period and slope values were determined from each plot. These values (slope/mV pHg⁻¹) were plotted as a function of time (hour).

The response of the Hg (II)-ISEs based on 1 and 2 over a wide range of pH (2–12) was examined at two different concentrations of Hg(NO₃)₂ (1.0 × 10⁻² and 1.0 × 10⁻³ mol dm⁻³). The pH was adjusted by the addition of small drops of nitric acid (0.1 mol dm⁻³) and sodium hydroxide (0.1 mol dm⁻³). EMF values (mV) were recorded and plotted against the pH of the solution.

This experiment was achieved in the same electrochemical cell but at different concentrations of the Hg (II) cation salt (1 × 10⁻⁷, 1 × 10⁻⁶, 1 × 10⁻⁵, 1 × 10⁻⁴, 1 × 10⁻³ and 1 × 10⁻² mol dm⁻³). EMF values (mV) were recorded after the addition of each concentration of the Hg (II) cation salt and the membrane potential was plotted against the dynamic response time to Hg (II). On the other hand, the electrode was immersed at two different concentrations of aqueous solution of Hg(NO₃)₂ (1 × 10⁻⁴ and 1 × 10⁻³ mol dm⁻³) respectively, and potential values were plotted as a function of response time to demonstrate the reversibility of the Hg (II) selective electrodes.

The life time of the Hg (II)-ISEs was examined over a period of four months. This was achieved my running a calibration experiment that involved the addition of the Hg(II) salt (6.67 × 10⁻⁸ to 4.32 × 10⁻² mol dm⁻³ concentration range) over a period of time (1, 7, 20, 40, 60, 90 & 120 days) where the potential values were recorded. Slope values were calculated from the calibration curves after which they were plotted as a function of time (days).

2.5.4 Selectivity measurements: effect of interfering ions

The matched potential method (MPM) recommended by IUPAC (International union of pure and applied chemistry, 2000), was used to assess the potentiometric selectivity coefficients of the membrane electrode. Thus, the concentration of the reference solution (Hg(NO₃)₂) was 1 × 10⁻⁴ mol dm⁻³, while that of the primary ion was 1 × 10⁻² mol dm⁻³ whereas the foreign ions concentration was 1 × 10⁻¹ mol dm⁻³. The potential change (ΔE) was recorded for the primary ion ( $a_{\mathit{Hg}(II)}^{\text{'}}$ ) as the addition of the interfering ion ( $a_{M^{n+}}$ ) was made until the same potential difference (ΔE) was obtained.

2.5.5 Analytical application of 1 and 2 based Hg(II) ISEs

The use of the fabricated Hg(II)-ISEs based on the ionophores 1 and 2 as indicator electrodes for the potentiometric determination of Hg(II) ion in the dental amalgam filling samples were carried out as previously reported (Danil de Namor et al., 2015).

3.3.1 Potentiometric response of carrier mediated ISEs towards Hg(II) as function of the membrane composition

The potentiometric characteristics (slope, linear range, detection limit and response time) of the electrodes towards Hg(II) were investigated for five membranes with different compositions and mixing ratios of 1 and 2, plasticizer and additive while a fixed amount of PVC was used.

Data on the response characteristics for the investigated Hg(II)-ISEs are listed in Table 4 and calibration plots for 1 (E1- E5) and 2 (E6-E10) are presented in Fig. 3a and b respectively.

Close

Fig. 3

Effect of different membrane compositions on the potential responses of ISEs based on t ionophores 1 and 2. Number of trial (N) = 4.

Export to PPT

In the absence of the ionophore (E1 and E6, blank experiments) the response characteristics displayed no selectivity towards the Hg(II) cation. In the absence of additive (E 2 and E 7) the response characteristics (slope, linear range, detection limit and response time) of the Hg(II)-ISE along with the selectivity towards Hg(II) were negatively affected. As previously shown (Tavakkoli and Shamsipur, 1996; Fakhari et al., 1997) the presence of the lipophilic additive can reduce the resistance of the membrane, reduce ionic interference and thereby improve the sensitivity towards Hg(II) cation. However, this additive (oleic acid) was selected for this study since unlike sodium tetraphenylborate (Danil de Namor et al., 2015). It showed no interaction with Hg(II) cation.

The nature of the plasticiser was examined using membrane electrodes either with dioctyl sebacate, DOS (E-3 and E8) or o-nitrophenyl octyl ether, o-NPOE (E-5 and E10). These plasticisers showed no interactions with Hg(II) (Danil de Namor et al., 2015). This is a very important point to emphasise because if there is an interaction of the plasticizer with Hg(II) this is most likely to have an impact on the selectivity coefficient and would reflect not only the selectivity of the receptor for Hg(II) but also that of the plasticiser for this cation. Results showed the best response behaviour of the electrode was that involving the o-NPOE plasticiser (E-5 and E10) showing for 1 a slope value of 29.4 ± 0.3 mV decade⁻¹ (near Nernstian behaviour of a slope value of 29.5 mV decade⁻¹), a linear range of 6.78 × 10⁻⁷- 1.34 × 10⁻³ mol dm⁻³, a detection limit of 3.16 × 10⁻⁷ mol dm⁻³ and response time of 11 ± 1 sec. For 2 the best performance in the ISE was seen in E-10 with membrane composition of 2 (1%), o-NPOE (68%), PVC(30%) and OA (1%). The slope value for this membrane was 29.7 ± 0.3 mV decade⁻¹, close to the value expected from Nernst equation, detection limit of 6.78 × 10⁻⁷ mol dm⁻³, linear range of 1.13 × 10⁻⁶-1.34 × 10⁻³ mol dm⁻³ Hg(II) and a response time of 10 ± 1 sec. So far advantages of the E-5 membrane containing 1 relative to E-10 containing 2 are the wider linear range and the lower activities that can be detected for the former relative to the latter. However the electrode response time is approximately the same for both membranes. E5 and E 10 were selected for further investigations. It is generally found that the use of o-NPOE as plasticizer leads to a longer linear section in the calibration curve and a lower detection limit than the use of DOS as plasticizer (Bedlechowicz and Hulanicki, 2002).

3.3.2 Effect of solution pH at different internal solution concentration, soaking, response time and reversibility of the Hg(II)ISEs

See Supplementary Material for interpretation.

3.3.3 Selectivity of 1 and 2 based Hg(II)ISEs

Selectivity coefficients ( $k_{{\mathit{Hg}}^{(II)},M^{n+}}^{\mathit{pot}}$ ) of the Hg(II)-ISEs based on 1 and 2 determined by the Matched Potential Method recommended by IUPAC are shown in Table 5.

It can be observed from these values that the electrode showed a high selectivity to Hg(II) with no significant interference from other cations particularly for the ISE containing receptor 2. In fact for the ISE containing 1, the selectivity coefficients for Ni(II), Pb(II) and Cd(II) follow the same pattern as that found from ¹HNMR in that the significant chemical shifts indicating ion-receptor interactions leads to a decrease (less negative) in the selectivity coefficients for these cations relative to Hg(II). This also applies to the ISE containing 2 in that among the bivalent cations investigated in solution, this receptor interacts only with Hg(II).

The following correlation (Eq. (7)) involving the difference between the stability constant of a given cation and the selectivity coefficient of the same cation, both relative to mercury (II) was previously demonstrated for two calix[4] based receptors and now this correlation is shown for 1 (Table 6). As pointed out before (Danil de Namor et al., 2018), validity of Eq. (7) requires that the conformation of the receptor is not altered from that in solution by its incorporation in the IS membrane. We have also explained previously that i) the transfer of the ion from water to the membrane phase must be related to the response time of the electrode rather than to the equilibrium potential and ii) ion pair formation in the membrane phase will lead to a non- linear response of the electrode. Therefore these two issues do not seem to present any problem regarding the selectivity of the electrode in acetonitrile and in the membrane phase (Danil de Namor et al., 2018).

A direct implication of Eq. (7) is that selectivity coefficients of carrier mediated ISEs involving calix[4] receptors can be obtained from thermodynamic investigations without the necessity of constructing the electrode which as stated previously is a time consuming process. In order to explore further the validity of this equation with other calix[4] based carrier ISEs, Eq. (7) is applied using stability constant data for 1 and Hg(II) (Table 3) and selective coefficients for the same receptor from Table 5. Stability constants for other metal (II) complexes can be calculated.

As far as the Hg(II) ISE based on receptor 2 is concerned, application of Eq. (7) using the stability constant data previously reported in the literature (log K_s = 4.5 in MeCN at 298.15 K, Danil de Namor and Pawlowski, 2011) and the selectivity coefficients for this receptor and metal cations relative to Hg(II) given in Table 5, approximate values for stability constants of other cations and 2 are calculated and these are shown in Table 7. These data show that very weak complexes are formed to an extent that no chemical shift changes were observed in ¹H NMR studies previously reported in CD₃CN. Again log Ks values lower than 2 are below the lower limit of detection of calorimetry.

3.3.4 SEM-EDX studies of ISE membranes based on 1 and 2

SEM micrographs along with EDX spectra of the PVC membranes containing ionophore 1 or 2 unloaded (4a & 4c) and loaded with Hg(II) (4b & 4d). The SEM micrograph (4a) validates a uniform microporous membrane with honeycomb structure before being superimposed in the ISE with no particles or dark patches. EDX spectrum shows the bulk elemental composition of the membrane grafted with the receptor. The relative peak heights of these elements in the membrane are correlated with their corresponding concentrations. The scanning electron micrograph (4b) shows distorted micropores indicating ageing of the membrane. The EDX spectrum shows a remarkable reduction in the heights of O Kα, N Kα, and Cl Kα with the appearance of new characteristic peaks of Hg, which provides clear evidence of the cation assembly to the hydrophilic binding sites of the ionophore. The reduction in the Cl peak is justified by the leaching of anionic impurities in the membrane support (Park et al., 2001).

As far as SEM images of PVC membranes with ionophore 2 is concerned, significant morphological changes in both micrographs (4c & 4d) are clearly seen with uniform micropores and no dark areas in the control membrane (c). On the other hand, white patches spreading over the membrane surface was observed in Fig. 4d indicating the presence of Hg(II) cation. The EDX analysis of the scanning electron micrograph (Fig. 4d) displayed mercury characteristic peaks in the membrane components confirming the presence of the analyte in the IS membrane.

Close

Fig. 4

SEM micrographs coupled with EDX spectra of control membranes (a) & (d) and membranes used for sensing Hg(II) as nitrate salts in water (b). Aluminium peak (Al) in the spectra referred to the aluminium stub used for samples’ mounting.

Export to PPT

3.3.5 Testing the electrodes

A number of analytical techniques are available for the quantitative determination of Hg (II). Among them are Cold-Vapour Atomic Absorption Spectrometry, CV-AAS, Ion Coupled Plasma Spectroscopy, ICP-OES and others (Bagheri and Gholami, 2001; Ferrúa et al., 2007; Martinis et al., 2009; Jiang et al., 2010; Zavvar Mousavi et al., 2010). These techniques are not always available in Developing Countries due to their high cost and maintenance. In addition samples need to be transported from the site to the Laboratory. This is not the case with ISEs which also have the advantage of analysing samples ‘in situ’. The Ion Selective electrodes described in this paper were tested by measuring the amount of Hg(II) in a filling dental amalgam provided by Nordiska Dental AB Industries (Sweden) and the results compared with those obtained by Cold Vapour-Atomic Absorption Spectrometry (CV-AAS) are shown in Table 8. The findings obtained by both electrodes were found to be in good agreement with those obtained by the CV-AA. Accordingly, the electrodes are good alternative for Hg(II) quantitative determination in real dental amalgam samples.

3.3.6 Response characteristics comparison of 1 and 2 and with other reported calix[4]pyrroles, calix[4]arenes based Hg(II) ISEs and other macrocycles

As mentioned earlier, 1 and 2 have the same pendant arms (thioacetamide) but the former is fully functionalised while the latter is partially functionalised and their fundamental studies revealed high selectivity for Hg(II) particularly for 1 in the formation of the 1:1 complex. In terms of response characteristics of Hg(II)-ISE based on these two ionophores, both electrodes (E-5 & E-10) displayed a Nernstian behaviour. However, E-5 based on 1 showed a slightly lower detection limit (3.16 × 10⁻⁷ mol dm⁻³) and a wider linear range (6.78 × 10⁻⁷ to 1.34 × 10⁻³ mol dm⁻³) than E-10 based on 2. The response time of both electrodes is practically the same (11 ± 1 sec for E-5 and 10 ± 1 sec for E-10). Having stated it the ISE based on 2 has the advantage of interacting weakly with other bivalent cations relative to Hg(II) and therefore interference with other cations are unlikely to occur.

Calix[4]pyrrole based Hg(II) ISEs are rarely reported in the literature. Calix[2]thieno[2]pyrrole and meso-tetramethyl-tetrakis-(4 N,N-diethylacetamide)phenoxymethylcalix[4]pyrrole were used as ionophores for Hg(II)-ISEs (Danil de Namor, 2007; Abbas, 2012; Danil de Namor et al., 2015). However, the former was excluded from the comparative study due to the use of NaBPh₄ additive in the PVC membrane, which as reported before (Danil de Namor et al., 2015), interacts with Hg(II). Therefore it is not possible to assess whether or not the electrode response to Hg(II) is due entirely to the ionophore or is the result of the contribution (positive or negative) of the additive and the cation. On the other hand, the calix[4]pyrrole amide ionophore based Hg(II)-ISE showed similar response characteristics for the slope value (30 ± 1 mV decade⁻¹) and response time (10 sec.) with 1 based Hg(II)-ISE (E-5) but a slightly lower detection limit and a slightly wider working concentration range than 2 (see Table 4).

The Hg(II)ISEs based on a thia-bridge modified calix[4]arene developed by Mahajan and co-workers (Mahajan et al., 2008) had a detection limit of 10⁻⁶ mol dm⁻³ and a response time of 20 sec. Moreover, Tyagi and co-workers (Tyagi et al., 2010) constructed a calix[4]arene thioether derivative based Hg(II) ISE using NaBPh₄ additive in the preparation of PVC membrane. Though the authors reported a wide concentration range (7.2 × 10⁻⁸-1.0 × 10⁻¹  mol dm⁻³) and a fast time response (14 sec.) for it, this cannot be included in the comparative study for the reasons stated above. Another Hg(II)-ISE based on calix[4]arene derivative with thiazole moieties was reported by Lu and co-workers (Lu et al., 2003) with a slope of 28.79 ± 0.3 mV decade⁻¹ and concentration range of 5 × 10⁻⁶ to 1x 10⁻² mol dm⁻³. As far as other macrocycles are concerned a number of Hg(II) ISEs have been reported using crown ether derivatives as carriers. Among these 1,4 -dithia-12-crown −4 and 15-crown-5 have been reported, linear range of 10⁻⁶-10⁻² mol dm⁻³ which works at a pH ≥ 2 (Lai and Shih, 1986), hexa-thia-18-crown-6 tetraone with a linear range of 4. 10⁻⁶-1.10⁻² mol dm⁻³, life-time, 3 months and a pH range 0.5–2 (Fakhari et al., 1997), dibenzo-diazathia 18-crown-6- dione, linear range 8. 10⁻⁶-1.10⁻² mol dm⁻³, response time, 10 s, life-time, 3 months, pH range, 0.5–2.5 (Javanbakht et al.,1999), pentathia15-crown-5, linear range, 2.51. 10⁻⁵-1.10⁻¹ mol dm⁻³ (Gupta et al., 1997). To a much lesser extent ISEs based on a cryptand (as defined by the authors) have also been reported, linear range 1.10⁻⁶-1.10⁻¹ mol dm⁻³, detection limit, 5.10⁻⁶ mol dm⁻³.slope, 30.2, stable in 3–6 pH range (Patel et al., 2009) as well as that based on polyaniline-cryptand 222, linear range, 10⁻¹²-1.10⁻⁸ mol dm⁻³ (Muthukumar et al.,2007). The above calix[4] based ISEs offer advantages relative to crown ethers and cryptands in the design of carrier mediated ISEs.