Electrochemical impedance study and sensitive voltammetric determination of Pb(II) at electrochemically activated glassy carbon electrodes Jyh-Myng Zen,* Hsieh-Hsun Chung, Govindasamy Ilangovan and Annamalai Senthil Kumar Department of Chemistry, National Chung-Hsing University, Taichung 402, Taiwan. E-mail: [email protected] Received 7th March 2000, Accepted 10th April 2000 Published on the Web 22nd May 2000

An electrochemically oxidized glassy carbon electrode was found to work excellently in the trace determination of Pb(II) using square-wave anodic stripping voltammetry. The associated phenomena were unravelled using chronocoulometry, linear scan voltammetry and electrochemical impedance spectroscopy. The chronocoulometry results showed that the double layer charge enormously increased on preanodization due to the creation of phenolic and quinolic oxide functional groups on the surface. The ac impedance analysis indicated that the duplex-layer model could illustrate the oxidized glassy carbon electrode behaviour quantitatively. The experimental factors, including preanodization condition, square-wave parameter, supporting electrolyte and pH, were optimized. Using the optimum parameters, the calibration curve was constructed and the detection limit was found to be 0.7 ppb (S/N = 3). The electrode is quite stable for repetitive measurements. The practical application was demonstrated to estimate trace Pb(II) in groundwater with very good precision.

Introduction There is a constant demand for improved analytical methods for the sensitive and selective determination of Pb(II) in environmental, biological and medical fields, as high dosages of this toxic heavy metal ion lead to various physiological complications, including carcinogenesis, allergic reaction, hypertension, etc.1–4 Stripping analysis using mercury film electrodes is the most effective approach for Pb(II) determination of conventional methods.5–11 The development of polymer/mercury composite film electrodes is reported to lead to an improvement of selectivity, and these electrodes have proved highly sensitive.12–14 However, the difficulty in depositing Hg and its physical hazards mean that it is not ideal for further practical applications in biological samples. We report here a relatively simple method for the sensitive detection of Pb(II) using squarewave voltammetry (SWV) on a preanodized glassy carbon electrode (GCE). The sensitivity achieved is on a par with the polymer/mercury composite film electrodes. GCE is one of the most widely used electrodes in electroanalytical applications, with the advantage of a wide working window in both the anodic and cathodic directions. However, the electron transfer event on this dynamic electrode heavily depends on the prior history of pretreatment before it is actually used in analysis.15 Various pretreatment (or activation) procedures, depending on the kind of analysis and nature of the redox system, have been adopted to achieve faster electron transfer rates and more reproducible results.16–33 In the case of electrochemical activation, preanodization at very positive potentials has been accepted as the prime activating procedure, although other pretreatments, such as preanodization followed by short-time cathodization16,18–21 and exclusive precathodization,27,28 have also been reported. Such pretreatments are believed to result in an increase of surface quinone and other functional groups, which can then catalyse the oxidation/ reduction of the analyte. Consequently, another advantage of using GCE is that one can tune its surface to favour the detection of the species of interest. Nevertheless, a particular kind of activation pretreatDOI: 10.1039/b001848l

ment with respect to one redox reaction may indeed lead to deactivation towards another redox reaction. For example, preanodization of GCE at a potential more positive than 1.5 V (vs. SCE) was required in order to activate a freshly polished electrode for hydroquinone redox reaction; however, the same pretreatment deactivated the hexacyanoferrate(II) reaction.16 Hence, it becomes inevitable that, before studying any new redox system on GCE, a suitable activation method for that particular system needs to be formulated at first hand. Although extensive studies of the electron transfer dynamics of GCE after various pretreatments have been documented, less attention has been paid to the effective utilization of the activated electrodes in analytical applications. A few examples were demonstrated for oxidative amperometric flow injection determination of cyanide, oxalate, sulfite and EDTA.34–37 The electrochemically activated GCE with the surface functional groups generated during preanodization can also be used to preconcentrate metal ion species, as demonstrated for Cu(II) determination.38 Recently, we have investigated the effect of preanodization on the electron transfer dynamics of various carbon electrodes and clays for the sensitive determination of various biologically important compounds.39–41 In this paper, we further describe the determination of Pb(II) on a preanodized GCE. The mechanism of such anodic activation can also be understood by chronocoulometry (CC) and ac impedance measurements. Instead of obtaining a convoluted response, different charging components of the modified surface were obtained. An extensive impedance study on the preanodized GCE with reference to its performance in Pb(II) detection is discussed.

Experimental Chemicals and apparatus The standard solution of Pb(II) and all supporting electrolytes and standard solutions used in this study were obtained from Merck (Darmstadt, Germany). Aqueous solutions were preAnalyst, 2000, 125, 1139–1146

This journal is © The Royal Society of Chemistry 2000

1139

pared with doubly distilled and deionized water. Electrochemical measurements were carried out with a CHI 660 electrochemical workstation (Austin, TX, USA). A conventional three electrode set-up was used both for square-wave anodic stripping voltammetry (SWASV) and ac impedance experiments. A BAS model (West Lafayette, IN, USA) VC-2 electrochemical cell was employed in these experiments. The three-electrode set-up consisted of either a bare or preanodized GCE as working electrode, an Ag/AgCl reference electrode and a platinum wire auxiliary electrode. The electrochemical impedance of the preanodized GCE was measured at 10 discrete frequencies per decade from 10 mHz to 100 kHz at an amplitude of 10 mV (rms) using an AUTOLAB FRA frequency response analyser (3508 AD Utrecht, The Netherlands) with a FRA2 module controlled by an IBM compatible PC. Impedance measurements were performed at a constant applied potential (Edc) with respect to the equilibrium potential. The acquired data were analysed based on equivalent electrical circuits by a weighted, non-linear, least-squares method using the fitting software, EQCRT, elaborated by Boucamp.42 Fitting constraints were imposed such that further iterations were stopped when the chi-square (c2) change was less than 0.001% compared to the previous iteration. The goodness of fit was assessed from minimum c2, correlation matrix and relative error distribution plots. Less than 5% fluctuations between the experimental and fitted data were assumed to be satisfactory in confirming the validity of the selected fitting circuit. Procedure A GCE of 3 mm in diameter was polished to a mirror finish with a polishing kit (BAS) and subsequently subjected to ultrasonic cleaning to remove the polishing debris. As it is known that preanodization forms a very thick oxide film on the surface, the cleaning procedure was repeated before every experiment to ensure that a fresh surface was exposed. The polished and cleaned electrode was subjected to preanodization at a static potential for a desired time in the respective solutions. During preconcentration, the test solution was stirred for effective accumulation in a required preconcentration time. Quantitative determinations were then performed in the SWASV mode. After stripping, the electrode was regenerated immediately and the renewed electrode was then checked in the respective base electrolyte to ascertain no peak response before the next measurement. A groundwater sample was collected and prepared as reported previously and the content of Pb(II) was evaluated by the standard addition method.13,14

carbonyl-modified GCE revealed that the electron transfer reaction with cations, such as Fe3+/2+, Eu3+/2+ and V3+/2+, undergoes an inner-sphere charge transfer mechanism through the activation of surface functional groups. A similar explanation can also be applied in the present case for the increased current in (d). It is therefore expected that the obvious increase in the detecting signal of Pb(II) upon preanodization of GCE is due to the participation of certain surface functional groups. Further studies on linear sweep voltammetry at various sweep rates show a surface-confined behaviour up to 50 mV s21 for Pb(II) on the preanodized GCE; beyond this scan rate, the behaviour becomes diffusion-controlled. Since the metal can deposit on the preanodized GCE at different access sites, i.e. either in surface or interfacial porous sites,45–47 the fact that the corresponding stripping signals change with the scan rate is thus acceptable. Nevertheless, this observation may also indicate that the Pb deposits form in a three dimensional matrix as will be discussed later.41 Previous studies from ellipsometry and SIMS suggest that electrochemical oxidation treatment of GCE causes the surface formation of an oxide layer containing > CNO and –C–OH functional groups to a depth of around 200 Å.23,42 The signal observed here may result from the Pb(II) species incorporated into the porous oxide layer since the phenolic and quinolic surface functionalities were reported to trigger complexation for high-valent metal cation.38 The fact that there is a detectable signal observed even at open circuit for the preanodized GCE [(c) in Fig. 1] suggests that Pb(II) ions have a definite interaction with the functional groups in the electrode surface. A similar observation was noticed earlier for the palladium–carbon fibre electrocatalytic system in acid solutions, where palladium deposition was carried out through electro-oxidation of carbon fibre followed by palladium plating.45 An important message from this system, however, is that the double layer capacitance was found to proportionally increase with electro-oxidation.45 The double layer charge at various electrode conditions in the present system was thus evaluated by CC and the results obtained are summarized in Table 1. An attractive feature of CC is that the double layer charge and the charge due to the faradic

Results and discussion Voltammetric characteristics of Pb(II) on the preanodized GCE Fig. 1 shows the linear scan voltammograms of 1 ppm Pb(II) in pH 5.5 phosphate buffer solution (PBS) at various electrode conditions. As can be seen, no signal is observed on a polished GCE without preconcentration (a), whereas a very small anodic stripping signal is observed after preconcentration (b). On the polished GCE after preanodization (c), a signal at Epa = 20.57 V with a current function (if) of 13.08 A V21 s mol21 cm3 was detected. A very pronounced signal with Epa and if corresponding to 20.54 V and 37.4 A V21 s mol21 cm3, respectively, was obtained for the preanodized GCE with preconcentration (d). The if obtained value in (d) is around 15 and 3 times higher than the if value of (b) and (c), respectively. Meanwhile, the shift of the redox potential to a more positive value in (d) is a good indication of the feasible deposition of Pb on the GCE. Recent studies by McCreery and coworkers43,44 on 1140

Analyst, 2000, 125, 1139–1146

Fig. 1 Linear scan voltammograms for 1 ppm Pb(II) in pH 5.5 PBS at a bare GCE (a, b) and a preanodized GCE (c, d). In (b, d) the GCE was preconcentrated at –1.2 V for 30 s. Scan rate, 100 mV s21; Pa = 2.3 V; ta = 3 s in 0.01 M H2SO4.

process of the surface-adsorbed species can be differentiated in a single experiment.48 Typically, the CC intercept in the reverse step of a q vs. t1/2 plot (Anson plot) corresponds to the double layer charge, while the intercept in the forward step yields the sum of the charge associated with reaction of the surface confined species and the double layer charge. From the difference between these two intercepts, one can measure the amount of redox species on the surface. As illustrated in Table 1, the double layer charge considerably increased upon preanodization. The intercept of the reverse sweep, and hence the double layer charge, increases by six times after preanodization; on the other hand, the slope changes little. The excess charge in the forward step for the preanodized GCE can be attributed to the faradic reactions of surface functional groups.20 Interestingly, whether the GCE is preanodized or not, both the forward and reverse intercepts in the presence of Pb(II) are similar to the results in pure buffer solution. This dictates the following two important conclusions. First, the increased intercept in the forward step in Pb(II) solution is due to the redox reactions of the surface functional groups on the preanodized GCE. Second, the Pb stripping process is strictly under a diffusion-controlled mechanism. The observation of a diffusion-controlled regime indicates that the Pb deposit may form in a three-dimensional matrix. The most appropriate candidate for the surface functional groups on the preanodized GCE, obviously, is the formation of a multilayer oxide matrix during preanodization.41 Since the difference in the forward and reverse intercepts, 70.1 mC, is due to the faradic process associated with the surface functional groups, the total surface coverage concentration can thus be calculated according to: > CNO + H+ + e2 ? > C–OH.20,49 Based on Q = nFAGt (Gt, surface coverage; A, electrode area; n, number of electrons transferred), the corresponding Gt is found to be 1.05 3 1028 mol cm22. Note that this Gt value is comparable with those reported for polymer-modified electrodes50 and further confirms the formation of the multilayer oxide film on preanodization. Table 1 also contains CC results for the preanodized GCE after polarization at 21.2 V for 60 s. As can be seen, the intercept in the forward step is largely reduced from 95.4 to 11.4. Obviously, the surface functional groups produced during preanodization are reducible and the potential of 21.2 V is enough to reduce the > CNO groups. To understand the influence of preanodization on the GCE surface, an ac impedance study was executed, and a quantitative correlation between these surface functionalities and the performance of the electrode is established in the following section. Preanodized GCE surface features by impedance analysis Electrochemical impedance measurements for the preanodized GCE were made over a wide frequency range from 100 kHz to 10 mHz, so that related parameters associated with the electrode/electrolyte interface could be accurately evaluated.51 Since the optimum response was observed at Pa = 2.3 V, the

impedance data were acquired for electrodes under this preanodized condition in different blank solutions and at various applied potentials (Edc) in the window of 0–0.4 V. The three-dimensional Edc–ZA–log(f) plots obtained are shown in Fig. 2. Similar plots were also advocated previously to explain the overall changes in impedance data as a function of the experimental conditions.51–54 Significant changes are observed in the lower frequency region, indicating that the processes occurring in this frequency region differ to a considerable extent in varied preanodization media and with varying Edc. The fact that the ZA values obtained were found to be higher at the bare than at the preanodized GCE indicates the formation of electrically active surface sites (i.e. surface oxides). The formation of surface oxide depends on the preanodization medium in the following order: KCl > Na2SO4 > HCl > H2SO4. Furthermore, the ZA values obtained were also found to decrease when Edc was more positive in all cases. Compared to the analytical results, the order is exactly opposite to the peak currents observed, i.e. H2SO4 > HCl > Na2SO4 > KCl. Obviously, the ac impedance results can provide an explanation to the analytical responses. As reported earlier, the effect of surface treatment on electron transfer events cannot be simply interpreted by considering the double layer alone, and the role of surface functional groups should also be taken into consideration.15 The complex plane impedance spectra were thus analysed using the equivalent circuits as depicted in Fig. 3. Fig. 3A was used to fit the impedance data for GCE preanodized in H2SO4, HCl and Na2SO4 and Fig. 3B for that in KCl. Note that other suggested circuits do not fit the data obtained very well.55 In the proposed circuit, Rs represents the resistance of the electrolyte, and the various capacitances proposed in the circuit are assumed as a constant phase element (CPE) instead of a perfect capacitor. Since the Nyquist plot shows various transformations in the frequency window studied, the double layer is represented by the combination of two CPEs in the proposed circuit. A duplexlayer model was used to explain the preanodized GCE interface by assuming the presence of a compact inner layer (Qdl) characterized by high field conduction and an outer porous oxide layer (Qd) where ionic transport takes place. Of course, the model is an idealized description of the thick porous oxide film on the preanodized GCE surface. Rct and Rf represent the charge transfer resistance and the reaction impedance due to the surface functional groups, respectively. The capacitance of the surface functional groups is accounted for by the parallel set-up of Rf and Qf similar to the impedance generated by the adsorbed species on the surface.52 In KCl medium, a drastic improvement in fitting was noticed after connecting an additional resistor, Rd, with Qd in parallel connection. The strange behaviour in KCl medium is due to the poor availability of [H+] and the high specific adsorption of Cl2 ions. The adsorption of Cl2 ions into the microporous oxide layer was reported earlier for a Pt–O layer in neutral medium.56,57 This behaviour also has phenomenological similarity to the accumulation charges in the microchannels of Nafion film and to the impedance results obtained.58 However,

Table 1 Chronocoulometrya study of various electrodes in different solutions Slope (31026)

Intercept/mC Electrode system

Forward

Reverse

Excess

Forward

Reverse

GCE (in buffer solution) 6.325 24.706 1.619 1.786 21.626 GCE (in Pb2+ solution) 5.828 24.166 1.662 1.588 21.542 4.786 23.166 1.620 1.814 21.855 GCE (in Pb2+ solution)b 98.920 227.970 70.950 3.696 21.393 GCE (in buffer solution)c GCE (in Pb2+ solution)c 95.380 224.090 71.290 3.330 21.299 11.380 22.997 8.383 1.984 22.404 GCE (in Pb2+ solution)bc a Experimental conditions: E = 0.1 V; E = 21.2 V; pulse, 1500 ms; [Pb2+] = 2 ppm in pH 1 PBS. b Preconcentration conditions: P = 21.2 V; t = 60 s. i f p p c Preanodization conditions: P = 2.3 V; t = 3 s in H SO . a a 2 4

Analyst, 2000, 125, 1139–1146

1141

in the case of HCl, the same anion of Cl2 did not alter the duplex layer. Good fits of the experimental data in both the complex plane spectrum and Bode’s plots were observed as shown in Fig. 4. The phase angle variation with frequency actually shows the overlap of two peaks and was evident especially for Edc = 0.1 V. The peak observed at higher frequency represents the

duplex layer and that observed at lower frequency corresponds to the surface reaction impedance. Usually, the capacitance line in the Nyquist plot in the lower frequency region inclines constantly by an angle a between 0° and 45°. The deviation from vertical for an ideal capacitor is attributed to the dispersion of frequency due to the rough

Fig. 2 Three-dimensional E–log(f)–ZA plots for (A) bare GCE and (B) GCE preanodized at 2.3 V in 0.01 M H2SO4. (C–E) are the same as (B) except in Na2SO4 (C), HCl (D) and KCl (E). Impedance data were acquired in pH 5.5 PBS.

1142

Analyst, 2000, 125, 1139–1146

surface that is often associated with a solid electrode.59 The inhomogeneities of the electrode surface manifest themselves in the equivalent circuits by a CPE instead of a pure capacitor. ZCPE = 1/T(jw)r, where T represents a pure capacitor only when r = 1, and r is related to a by a = (1 2 r)90°. So, a = 0 and r = 1 represents a perfect capacitor, and lower r values directly reflect the roughness of the electrode. A careful examination of the estimated parameters in Table 2 reveals that various trends can be observed and correlated with the performance of the electrode. First, the higher Qf values obtained in acids than in neutral media demonstrate the generation of more surface functional groups during preanodization. The highest Qf value with low Rct values for the preanodized GCE in HCl medium shows some specific faradic interaction of Cl2 with surface functional groups. The exact features of this interaction are still unknown. Second, the relatively high Qd observed in KCl solution implies the accumulation of Cl2 in the interfacial sites of the preanodized GCE. However, no faradic interaction is expected as in HCl since the Rct value in KCl is considerably higher. Third, this explanation is further confirmed by the respective r values obtained. The fact that the observed r values typically lower than 0.5 indicate a distributed capacitance (note that r < 0.5 for an extremely porous surface and 1 for an ideal

Fig. 3 (A) Equivalent circuits used to fit the impedance data for GCE preanodized in H2SO4, HCl and Na2SO4. (B) The same as (A) except in KCl.

smooth surface). The rd values for electrodes preanodized in H2SO4, HCl and Na2SO4 are lower than that in KCl. Moreover, in the case of KCl, the rd value is close to 1, indicating a perfect capacitance, and it is expected that Cl2 ions occupy the active surface sites with the formation of a parallel plate condenser near the electrode surface. The rd value around 0.5 in the other media suggests free transport inside the micropores. Fourth, the parameter Qdl also shows that the electrode preanodized in KCl

Fig. 4 (A) Nyquist plots of impedance acquired in pH 5.5 PBS on GCE preanodized in 0.01 M H2SO4 (a), Na2SO4 (b) and HCl (c). (B) Bode’s plots of the data presented in (A). Points show the experimental data and the full line is calculated from the optimized parameters.

Table 2 Fitting parameters obtained from the proposed equivalent circuit in various media used for preanodization Medium (0.01 M)

Edc/V

Rs/W

H2SO4

0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4 0.0 0.1 0.2 0.3 0.4

114.0 124.0 106.0 97.6 101.6 118.4 128.4 121.4 119.4 120.4 118.6 121.6 118.0 118.0 114.0 121.7 145.3 121.0 123.5 121.0

HCl

Na2SO4

KCl

Rd/W

Qd/mF cm22 rd

Qdl/mF cm22 rdl

Rct/kW

Rf/MW

Qf/mF cm22

rf

270.0 732.0 598.0 689.0 690.0

13.841 25.029 12.090 1.273 7.361 6.389 2.737 4.423 4.354 5.966 14.264 4.837 0.538 31.757 10.263 140.929 34.943 18.914 30.186 27.685

27.129 25.943 22.157 31.543 29.271 40.529 28.571 24.314 23.314 29.314 16.086 13.771 15.557 15.614 10.996 3.170 1.607 1.466 2.399 1.947

54.60 51.50 90.40 133.80 121.40 3.10 2.85 2.38 2.98 3.91 12.01 10.10 52.10 67.10 83.10 86.70 91.70 46.80 68.40 69.40

1.622 1.615 1.470 0.949 0.463 0.715 0.856 0.681 0.678 0.303 2.638 1.743 1.312 0.747 0.344 2.614 2.235 1.812 1.074 1.039

17.829 16.186 28.514 16.257 2.191 169.429 158.429 179.143 57.143 94.257 0.046 0.007 0.008 0.119 0.377 4.264 2.203 0.498 1.338 0.652

0.625 0.622 0.622 0.623 0.550 0.665 0.661 0.674 0.644 0.647 0.487 0.507 0.492 0.508 0.527 0.592 0.575 0.539 0.554 0.572

0.434 0.521 0.392 0.340 0.426 0.653 0.589 0.667 0.584 0.554 0.431 0.379 0.680 0.680 0.700 0.999 0.985 0.975 0.930 0.940

0.933 0.949 0.938 0.932 0.928 1.000 1.000 1.000 0.987 0.982 0.917 0.907 0.898 0.898 0.876 0.873 0.861 0.871 0.874 0.862

Analyst, 2000, 125, 1139–1146

1143

has the smallest value. Thus, the total capacitance of the oxide layer in KCl is mainly contributed by the oxide/anion interface for the preanodized GCE. On the other hand, in other cases, the oxide/solution interface contribution is more important than the oxide/anion interface. The rdl values are close to 1 in all cases, testifying that the frequency distribution occurs to a lesser extent at the oxide/solution interface. Analytical optimization and characterization For the sensitive determination of Pb(II), both the electrode conditions and SW parameters were carefully optimized. The preanodization potential (Pa) and preanodization time (ta) are considered as electrode parameters, while the modulation frequency, modulation amplitude and step width are examined as SW parameters. The effects of Pa and ta were examined first and the results are shown in Figs. 5A and 5B, respectively. The peak current increased steeply when Pa was more positive than 2.0 V and then remained steady up to 2.5 V. The observed result

Fig. 5 Effect of Pa (A) and ta (B) on the peak current and background charge for 2 ppm Pb(II) in pH 5.5 PBS. Preconcentration conditions as in Fig. 1.

1144

Analyst, 2000, 125, 1139–1146

is consistent with the literature evidence that the surface oxidation for GCE occurs beyond 1.8 V.16,23,49 Moreover, such a saturation of peak current beyond 2.5 V is also in agreement with our previous results for uric acid determination.39 Preanodization at potentials beyond 2.5 V is detrimental due to the formation of an insulating oxide layer on the surface which is too thick. The observed peak current for Pa > 2.5 V was lower than that of a bare GCE testifying to the formation of the insulating layer. Similarly, ta also has an enormous influence on the observed peak current (Fig. 5B), which increased initially and then decreased slightly at longer ta. These results imply that both Pa and ta are very critical in the sensitive determination of Pb(II). Typical Pa and ta values of 2.3 V and 3 s were selected as the best parameters in the present study. The background charge (Qb) obtained from the integration of the background current during the stripping process is also plotted in Fig. 5 for comparison. Interestingly, Qb was also found to increase steeply when Pa > 2.0 V. The major difference, however, is that Qb increases even beyond 2.5 V, while the Pb(II) stripping peak current decreases. The increase in Qb even beyond 2.5 V obviously indicates a continuous creation of polar functional groups. To address this problem, CC was carried out for GCE preanodized over a wide range of Pa (1.6–3.0 V) in base electrolyte with 2 ppm Pb(II) under basic conditions as shown in Table 1. The intercept in the reverse step corresponding to the double layer charge was found to increase with Pa. Meanwhile, the intercept of the forward step increases rapidly, resulting in a steady increase in the difference between these two intercepts. The surface concentrations of redox groups, obtained from these differences, are plotted as a function of Pa as shown in Fig. 6. Unlike the trend in the stripping current, the surface coverage increases steadily even when Pa > 2.5 V. It is therefore confirmed that the linear increase in Qb is due to the continuous formation of polar functional groups on the surface. Nevertheless, the constructive interaction of Pb(II) with > CNO is limited by diffusion complications. As mentioned earlier, it is detrimental when Pa > 2.5 V due to the formation of an insulating oxide layer on the surface which is too thick. The fact that the observed difference between the forward and reverse intercepts is abruptly increased from 132.9 to 300.4 at Pa = 2.5 V and Pa = 2.6 V, respectively, provides strong support for this explanation. The influence of the medium used for preanodization on the performance of the electrode was also studied in detail. The

Fig. 6 Effect of Pa on the calculated Gt with ta = 3 s in 0.01 M H2SO4. Other conditions as in Fig. 5.

medium solutions of H2SO4, Na2SO4, HCl, KCl and PBS with various pH and anion were selected for study. Fig. 7 shows the effect of the medium used for preanodization on the stripping current for 0.5 ppm Pb(II). As can be seen, there is an inverse relation between Qb and peak current for both the H2SO4/ Na2SO4 and HCl/KCl pairs. It seems as if the anion of the medium has a greater influence on the stripping current than does the pH value. The much smaller peak current on the electrode preanodized in KCl medium is likely to be due to the specific adsorption of Cl2 into the pores of the oxidized layer, which is supported by the observation of a large Qb value in this medium. However, compared to the response in HCl medium, Cl2 adsorption is more severe in neutral medium. Note that the ac impedance result discussed earlier directly corresponds to this phenomenon. Since SWV was found to be much more sensitive than linear scan voltammetry in the detection of Pb(II) using the preanodized GCE, the influence of various SW parameters was studied next. These SW parameters are interlinked and have a combined influence on the peak current. The SWASV peak current was observed to increase with increasing in SW frequency. After reaching a maximum around 400 Hz, the peak current decreased slightly. Similarly, the SWASV peak current increased with amplitude and saturated around 40 mV. A further increase in amplitude yielded a sloping background rendering measurement difficult. However, very sharp changes were noticed in the peak current with an increase in potential sweep. The scan increment together with frequency defines the effective scan rate. A very sharp increase in the peak current was noticed with an increase in step increment. After reaching a maximum around 5 mV, a very steep decrease in the peak current was observed. Overall, the best signal to background characteristics can be obtained with the following optimum SW parameters: modulation frequency, 400 Hz; modulation amplitude, 40 mV; modulation increment, 5 mV. Finally, using the above SW parameters, the effects of the preconcentration potential (Pp) and preconcentration time (tp) in SWASV for the determination of 1 ppm Pb(II) were studied. The peak currents were found to increase very quickly up to 21.0 V and 100 s for Pp and tp, respectively. The preconcentration parameters can thus be selected based on the experimental requirement and convenience. Repetitive preconcentration– measurement–regeneration cycles were then performed to characterize the reproducibility of the electrode performance.

Fig. 7 Effect of the preanodization medium on the peak current and background charge. Other conditions as in Fig. 1.

Transferring the electrode after Pb(II) detection to PBS and then scanning in the same potential window achieved regeneration. Initially, a small trace of a voltammetric peak corresponding to Pb(II) could be detected. However, repeated stripping in base electrolyte for, normally, five cycles regenerated the original surface and it could subsequently be used for analysis. The electrode regeneration depends on the pH of the stripping solution. The electrode could be perfectly renewed in more acidic solutions. The results of 20 successive cycles for 1 ppm Pb(II) showed a relative standard deviation of 1.71% and 7.83% in pH 1 and pH 2 PBS, respectively. Since pH 1 PBS was proved to work effectively for electrode renewal, for convenience, the same solution was also chosen as the supporting electrolyte in detection. Fig. 8A shows the stripping voltammograms in pH 1 PBS on GCE preanodized in 0.01 M H2SO4 solution. At all concentrations, the stripping peak was observed at the same peak potential. It was interesting to observe a small stripping peak at a higher potential around 20.35 V. This is similar to the observation for catechol oxidation on preanodized GCE, where the catechol absorbed into the micropores is energetically

Fig. 8 (A) SW voltammograms and (B) calibration plot for Pb(II) determination in pH 1 PBS on the preanodized GCE. SW parameters: frequency, 400 Hz; amplitude, 40 mV; step height, 5 mV. Other conditions as in Fig. 1.

Analyst, 2000, 125, 1139–1146

1145

different and requires different potentials to be oxidized.20 Thus, the small peak observed at higher potential can be attributed to Pb stripping from the deep pores with energetically different sites. Nevertheless, the larger peak at 20.48 V was used for calibration purposes. A linear calibration curve was obtained in the concentration range 0–2 ppm (Fig. 8B) with a slope (mA ppm21) and correlation coefficient of 0.205 and 0.999, respectively. The detection limit was calculated to be 0.7 ppb (S/N = 3). This value was comparable with the earlier result on a Nafion/copper–mercury film electrode (0.08 ppb with tp = 150 s) and a poly(4-vinylpyridine)/mercury film electrode (0.3 ppb with tp = 30 s).13,14 The analytical utility of the proposed method was demonstrated by applying it to the determination of Pb(II) in groundwater. Using the standard addition method with 10 ppb per spike, a linear equation (y = 0.940 + 0.544x, r = 0.9988) was obtained. Based on this equation, the concentration of Pb(II) was found to be 2.46 ppb in groundwater and the recovery in each spike was very good ranging from 96.0% to 102.7%. The detected values for the same samples from the Nafion/copper–mercury film electrode and poly(4-vinylpyridine)/mercury film electrode were 2.36 and 2.26 ppb, respectively. These results provide sufficient evidence for the feasibility of using the proposed method in practical applications.

Conclusions The modification of GCE by simple electrochemical preanodization works excellently to increase the sensitivity of Pb(II) estimation. The preanodized surface yields specific interaction of Pb(II) with the functional groups generated during preanodization. The ac impedance analysis of the preanodized GCE verifies the formation of surface functional groups in different preanodization media. The counter-anion in the preanodization medium plays a relatively more important role than the pH of the medium. The electrodes preanodized in SO42 media produce more surface functional groups; thus, the charging component of the surface species is higher. On the other hand, for the GCE preanodized in Cl2 media, relatively smaller Qf values were observed, indicating that the surface functionalities are generated to a lesser extent. It is proposed that, due to specific adsorption of Cl2 in the pores of the oxidized layer, further oxidation of the surface is prohibited. A duplex-layer model interprets the double layer of the preanodized GCE very well. More importantly, the nature of the oxide layer depends on the anion of the pretreating medium. Extensive studies using ac impedance should give detailed insights into these aspects and are currently under investigation in our laboratory.

Acknowledgements

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

The authors gratefully acknowledge financial support from the National Science Council of the Republic of China.

References

49 50 51

1 2 3 4 5 6 7

1146

G. Erla, S. Papa, M. Aicta, A. Tartarone, R. Murgo, D. Seripa, G. Di Giorgio, G. Lelli and V. M. Fazio, Eur. J. Can. A, 1998, 34, S73. M. Pokrajac, B. Brzakovic, Z. Levic and V. M. Varagic, Electroenceph. Clin. Neurophys., 1996, 99, 326. E. Reichrtova, J. Gajdosova, M. Ursinyova, V. Hladikova, V. Prachar, M. Weningerova and J. Holikova, Toxicol. Lett., 1995, 78, 69. V. A. Granadillo, J. E. Tahan and R. A. Romero, Clin. Chim. Acta, 1995, 233, 47. D. Sancho, M. Vega, L. Deban, R. Pardo and G. Gonzalez, Analyst, 1997, 122, 727. D. Jagner, D.-G. Lars and K. Aren, Anal. Chim. Acta, 1979, 106, 15. D. Jagner and S. Westerlund, Anal. Chim. Acta, 1980, 117, 159.

Analyst, 2000, 125, 1139–1146

52 53 54 55 56 57 58 59

D. Jagner, M. Josefson and S. Westerlund, Anal. Chim. Acta, 1981, 129, 153. D. Jagner, E. Sahlin and L. Rehman, Talanta, 1994, 41, 515. M. A. Baldo, C. Bragato and S. Daniele, Analyst, 1997, 122, 1. Q. Wu and G. E. Batley, Anal. Chim. Acta, 1995, 309, 95. K. Yokoi, A. Yamaguchi, M. Mizumachi and T. Koide, Anal. Chim. Acta, 1995, 316, 363. J.-M. Zen and J.-W. Wu, Anal. Chem., 1996, 68, 3966. J.-M. Zen and Y.-S. Ting, Anal. Chim. Acta, 1996, 332, 59. R. L. McCreery, in Electroanalytical Chemistry ed. A. J. Bard, Marcel Dekker, New York, 1991, vol. 17, p. 221. R. C. Engstrom, Anal. Chem., 1982, 54, 2310. N. Cenas, J. Rozgaite, A. Pocius and J. Kulys, J. Electroanal. Chem., 1983, 154, 121. R. C. Engstrom and V. A. Strasser, Anal. Chem., 1984, 56, 136. G. E. Cabaniss, A. A. Diamantis, W. R. Murphy, Jr., R. W. Linton and T. J. Meyer, J. Am. Chem. Soc., 1985, 107, 1845. T. Nagoaka and T. Yoshino, Anal. Chem., 1986, 58, 1037. T. Nagoaka, T. Fukunaga, T. Yoshino, I. Watanabe, T. Nakayama and S. Okazaki, Anal. Chem., 1988, 60, 2766. P. M. Kovach, M. R. Deakin and R. M. Wightman, J. Phys. Chem., 1986, 90, 4612. L. J. Keply and A. J. Bard, Anal. Chem., 1988, 60, 1459. M. L. Bowers and B. A. Yenser, Anal. Chim. Acta, 1991, 243, 43. C. Barbero, J. J. Silber and L. Sereno, J. Electroanal. Chem., 1988, 248, 321. L. Bodalbhai and A. Brajter-Toth, Anal. Chem., 1988, 60, 2557. G. Ilangovan and K. Chandrasekara Pillai, Langmuir, 1997, 13, 566. G. Ilangovan and K. Chandrasekara Pillai, J. Electroanal. Chem., 1997, 431, 11. I.-F. Hu, D. H. Karweik and T. Kuwana, J. Electroanal. Chem., 1985, 188, 59. N. M. Pontikos and R. L. McCreery, J. Electroanal. Chem., 1992, 324, 229. R. J. Rice, N. M. Pontikos and R. L. McCreery, J. Am. Chem. Soc., 1990, 112, 4617. G. W. Hance and T. Kuwana, Anal. Chem., 1987, 59, 131. M. Poon, R. L. McCreery and R. C. Engstrom, Anal. Chem., 1988, 60, 1725. A. G. Fogg and R. M. Alonsa, Analyst, 1987, 112, 1071. A. G. Fogg, R. M. Alonsa and M. A. Fernandez-Arciniega, Analyst, 1986, 111, 249. A. G. Fogg, M. A. Fernandez-Arciniega and R. M. Alonsa, Analyst, 1985, 110, 851. A. G. Fogg, M. A. Fernandez-Arciniega and R. M. Alonsa, Analyst, 1985, 110, 1201. K.-K. Shiu and K. Shi, Electroanalysis, 1998, 10, 959. J.-M. Zen, J.-J. Jou and G. Ilangovan, Analyst, 1998, 123, 1345. Y. Shih and J.-M. Zen, Electroanalysis, 1999, 11, 229. J.-M. Zen and P.-J. Chen, Anal. Chem., 1997, 69, 5087. B. Boucamp, Solid State Ionics, 1986, 20, 21. P. Chen, M. A. Frying and R. L. McCreery, Anal. Chem., 1995, 67, 3115. P. Chen and R. L. McCreery, Anal. Chem., 1996, 68, 3958. A. D. Jannakoudakis, D. Jannakoudakis, E. Theodoridou and J. O. Besenhard, Synth. Metals, 1992, 53, 47. J. Wang, Stripping Analysis, VCH, Deerfield Beach, FL, 1985. M. L. Bowers, J. Heffer, D. L. Dugger and R. Wilson, Anal. Chim. Acta, 1991, 248, 127. A. J. Bard and L. R. Faulkner, Electrochemical Methods, John Wiley, New York, 1980. A. L. Beilby and A. Carlsson, J. Electroanal. Chem., 1988, 248, 283. R. W. Murray, in Electroanalytical Chemistry, ed. A. J. Bard, Marcel Dekker, New York, 1984, vol. 13, p. 191. J. R. MacDonald, Impedance Spectroscopy, Wiley, New York, 1987. S. Ye, A. K. Vijh and L. H. Dao, J. Electroanal. Chem., 1996, 415, 115. J. R. MacDonald, J. Electroanal. Chem., 1987, 223, 25. L. Bai and B. E. Conway, Electrochim. Acta, 1993, 38, 1803. P. Heiduschka, A. W. Munz and W. Gopel, Electrochim. Acta, 1994, 39, 2207. P. Kulesza, T. Jedral and Z. Galus, J. Electroanal. Chem., 1980, 109, 141. J. Kawaik, T. Jedral and Z. Galus, J. Electroanal. Chem., 1983, 145, 163. G. Pourcelly, A. Oikonomou and C. Gavach, J. Electroanal. Chem., 1990, 287, 43. B. Piela and P. K. Wrona, J. Electroanal. Chem., 1995, 388, 69.