Talanta 72 (2007) 1392–1399
A mercury-free electrochemical sensor for the determination of thallium(I) based on the rotating-disc bismuth film electrode E.O. Jorge a , M.M.M. Neto a,b,∗ , M.M. Rocha a a
Departamento de Qu´ımica e Bioqu´ımica, Centro de Ciˆencias Moleculares e Materiais, Faculdade de Ciˆencias, Universidade de Lisboa, Campo Grande, Ed. C8, 1749-016 Lisboa, Portugal b Departamento de Qu´ımica Agr´ıcola e Ambiental, Instituto Superior de Agronomia, TULisbon, Tapada da Ajuda, 1349-017 Lisboa, Portugal Received 25 October 2006; received in revised form 5 January 2007; accepted 19 January 2007 Available online 30 January 2007
Abstract A bismuth film electrode was tested and proposed as an environmentally friendly sensor for the determination of trace levels of Tl(I) in nondeoxygenated solutions. Determination of thallium was made by anodic stripping voltammetry at a rotating-disc bismuth film electrode plated in situ, using acetate buffer as the supporting electrolyte. The stripping step was carried out by a square wave potential-time excitation signal. A univariate optimisation study was performed with several experimental parameters as variables. Under the selected optimised conditions, a linear calibration plot was obtained in the submicromolar concentration range, allowing the electrochemical determination of thallium in trace amounts; the calculated detection limit was 10.8 nM and the relative standard deviation for 15 measurements of 0.1 M Tl(I) was ±0.2%, for a 120 s accumulation time. Interference of other metals on the response of Tl(I) was investigated. Application to real environmental samples was tested. The bismuth film electrode appears to be a promising tool for electroanalytical purposes, ensuring the use of clean methodology. © 2007 Elsevier B.V. All rights reserved. Keywords: Thallium; Bismuth film electrode; Square wave voltammetry; Anodic stripping voltammetry; Mercury-free electrodes
1. Introduction Thallium is present in the environment in trace amounts. Emissions from cement plants and combustion of fossil-fuel are the main causes of thallium pollution [1]. Thallium is a non-biological element, but it enters cells via K+ transport systems, since the ionic form of Tl+ is similar to K+ due to their resemblance in ionic radii (Tl+ : 164 pm, K+ : 152 pm). Thallium binds more tightly than potassium to N and S ligands [2] and is very harmful for living organisms even at very low concentrations [3]. It is therefore important to develop sensitive and accurate analytical methods to determine trace levels of thallium in environmental and food samples. Although spectrophotometric techniques have been widely used to determine traces of thallium in different matrices [4–7],
∗ Corresponding author at: Departamento de Qu´ımica e Bioqu´ımica, Centro de Ciˆencias Moleculares e Materiais, Faculdade de Ciˆencias, Universidade de Lisboa, Campo Grande, Ed. C8, 1749-016 Lisboa, Portugal. Tel.: +351 962763055. E-mail address:
[email protected] (M.M.M. Neto).
0039-9140/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2007.01.047
electroanalytical methods also appear very suitable [8–13], because they are reliable and sensitive and require less expensive equipment. Anodic stripping voltammetry (ASV) is recognized as a powerful technique for measurement of heavy metals in trace amounts [14]. For this purpose, mercury electrodes, namely the hanging mercury drop and the mercury film electrode (MFE), have been commonly used, due to the high hydrogen overvoltage and the ability to form amalgams. Thallium is very soluble in mercury (43.0 at%) [15] and a good sensitivity is achieved with mercury electrodes because thallium ions are readily reduced and amalgamated during the deposition step of the ASV analytical protocol [16]. However, because of its high toxicity, mercury should be replaced by less toxic or, preferably, non-toxic electrode materials. In several countries, Hg-containing instruments and electrical components have been phased out and the use of mercury compounds has even been banned [17]. A number of alternative and new electrode materials has been investigated, including various forms of carbon, gold, iridium or boron-doped diamond, but none has approached the favourable electrochemical behaviour of mercury [18–22].
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
Recently, the bismuth film electrode (BFE) has been developed as a very attractive sensor for ASV of heavy metals to replace mercury electrodes, with an equivalent performance [23,24]. Bismuth can be electroplated onto an inert substrate (e.g. glassy carbon) using the in situ plating procedure introduced by Florence [25] to prepare the MFE. In addition to its low toxicity, the BFE offers other advantages over the MFE. It is less susceptible to oxygen interference, exhibiting lower background currents, not only in square wave stripping voltammetry but also in linear scan voltammetric experiments in non-deaerated media [23]. After measurements, the renewal of the BFE is electrochemically achieved. It must be stressed that the removal of mercury films from glassy carbon is not an easy task [26–28]. Whereas mercury films on glassy carbon consist of finely divided mercury droplets, the bismuth deposits have a completely different morphology [23,29,30]. Using scanning electron microscopy (SEM), Wang et al. observed a highly porous, three-dimensional fibril-like network for a bismuth deposit on glassy carbon, from acetate buffer solutions [23]. Thus, BFE exhibits good mechanical stability, which enhances its potential usefulness to operate under hydrodynamic conditions, e.g. in flow systems. All these attractive features make this sensor a promising replacement for the mercury electrodes in electroanalysis, with additional advantages. A significant number of reports on the analytical utility of the BFE has already been published [31–36]. Glassy carbon [24] and carbon paste [36] have been used as substrates to the development of bismuth-coated electrodes for thallium analysis by ASV. Other mercury-free sensors for the detection of thallium by ASV have also been proposed. Using a chemically modified glassy carbon electrode with Langmuir–Blodget film of p-allylcalix [4] arene, Dong et al. determined thallium in environmental water samples [37]. Graphite microelectrodes were applied for the same purpose with low detection and quantification limits, 0.01 and 0.03 g L−1 , respectively [10]. The quantification of thallium in the presence of lead and cadmium was carried out at a silver–gold alloy electrode by subtractive ASV [38]. It is well-known that ASV with hydrodynamic sensors is very advantageous [14,39,40]. The mass transport is controlled by well-defined hydrodynamics, leading to high sensitivity. The rotating-disc electrode is one of the most widely used electrodes of this type. Some electroanalytical studies involving bismuth film electrodes in hydrodynamic configurations have been reported in the literature for analytical purposes [33,34,41]. However, no application of the rotating-disc bismuth film electrode for the quantification of thallium in real samples seems to have been previously published. In this work, the analytical potentialities of a rotating-disc bismuth film electrode are explored for the determination of Tl(I) in environmental matrices by ASV, using a square wave voltammetric scan in the determination step. Square wave voltammetry enables a fast analysis with high scan rates, and minimises the problems arising from the presence of dissolved oxygen. After the optimisation of the experimental parameters, the sensor was successfully applied to the quantification of thallium in river water and soil samples.
1393
2. Experimental 2.1. Apparatus Electrochemical experiments were performed using a computer-controlled Autolab Basic PGSTAT 20 potentiostat (Eco Chemie, Utrecht, The Netherlands) driven by GPES software (Version 4.7). A M Series AA atomic absorption spectrophotometer and GF95Z graphite furnace system with sample dispenser (Thermo Electron Corporation, Cambridge; UK) was employed for the photometric determination of thallium. Electrochemical measurements were made in a glass cell (100 mL) equipped with a platinum wire auxiliary electrode, an Ag|AgCl|3 M KCl reference electrode (Metrohm), and a rotating-disc glassy carbon working electrode (area: 0.385 cm2 ) from Oxford Electrodes. Bismuth films were prepared in situ by electrochemical deposition of bismuth onto the surface of the glassy carbon disc. A combined pH electrode pHC 30058 with a Radiometer MeterLab PHM201 was used to make pH measurements. All the glassware was carefully cleaned by soaking the pieces, overnight, in a 2% diluted RBS 25 detergent, washed several times with distilled water and finally with Millipore water. 2.2. Reagents and solutions All solutions were made from analytical grade reagents and Millipore Milli-Q ultrapure water (conductivity <0.1 S cm−1 ) and stored in Pyrex glass flasks. The supporting electrolyte was a 0.1 M acetate buffer (pH 4.6) prepared from acetic acid and sodium acetate (Aldrich). The bismuth(III), thallium(I), lead(II) and cadmium(II) stock solutions were prepared from Bi(NO3 )3 ·5H2 O, TlNO3 , Pb(NO3 )2 and Cd(NO3 )2 ·4H2 O, respectively, all supplied by Merck. The working solutions were obtained by adding the appropriate amount of stock solution to the supporting electrolyte. 2.3. Procedure Prior to each experiment, the glassy carbon electrode was polished on a polishing pad with Alpha Micropolish alumina 0.3 m particle size (B¨uehler), rinsed with Milli-Q water and dried. The electrode surface was then activated by continuous potentiodynamic cycling (n = 10), at 50 mV s−1 , between −1.0 V and +0.8 V, in supporting electrolyte without the removal of oxygen. The bismuth film was prepared in situ, by co-deposition of bismuth and the target species at −1.4 V for 120 s, from 1.9 M bismuth(III) nitrate in acetate buffer (pH 4.6). After an equilibration period of 5 s, a positive going square wave scan to a final potential of −0.5 V was applied on the BFE while the anodic stripping voltammogram was being recorded. The electrode was continuously rotating at a rotation speed of 240 rpm. The selected square wave parameters were: frequency 40 Hz, amplitude 80 mV and scan increment 5 mV. After measurements, the electrode potential was held at +0.3 V for 30 s to clean its surface electrochemically. All potentials are referred to Ag|AgCl reference electrode.
1394
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
In cyclic voltammetry tests, solutions were previously deaerated with oxygen-free nitrogen for 15 min. The standard addition method was used to evaluate the content of Tl(I) in real samples. All experiments were carried out at temperatures of 25 ± 1 ◦ C. 2.4. Real samples pre-treatment Water samples were analysed after a minor pre-treatment, consisting of filtration through Whatman filter no. 1, followed by pH adjustment to 4.6 with acetate buffer. The soil samples were simply washed, since thallium compounds are very soluble in water. Owing to their high volatility [42], it was decided to avoid digestion procedures. Air-dried soil (50 g) was thus shaken for 1 h at room temperature with 150 mL of Milli-Q ultrapure water, using magnetic stirring. The liquid phase was then decanted and successive filtrations under vacuum with numbers 1, 41 and 42 Whatman filters were carried out and followed by a final filtration with a Minisart 0.20 m filter. The pH of the collected colourless filtrate was then adjusted to 4.6 with acetate buffer solution. Prior to voltammetric determinations, test solutions were transferred to the cell and an aliquot of bismuth(III) salt solution was added to obtain a Bi(III) concentration of 1.9 M. 3. Results and discussion 3.1. Electrochemical behaviour of Tl(0)/Tl(I) at the BFE The accessible potential window of the bismuth-coated electrode in acetate buffer was firstly investigated. The current response of a rotating-disc glassy carbon electrode, immersed in a solution containing 1.9 M bismuth(III) nitrate, was analysed after holding the potential at four different values (−0.8, −1.0, −1.2 and −1.4 V) for 120 s deposition time (tdep ). The recorded anodic stripping voltammograms indicate that the stripping potential of bismuth (Ep = −0.23 V) was not affected, but the peak current increased as the deposition potential became more negative, denoting that a better coverage of the electrode surface was attained at −1.4 V. The potential window of the bismuth-coated glassy carbon electrode is limited in the anodic region by the oxidation of bismuth, starting around −0.4 V; the hydrogen overvoltage, which limits the cathodic range, is quite high (starting around −1.2 V). Cyclic voltammetry of thallium in the presence of 1.9 M Bi(III) was performed in 0.1 M acetate buffer solutions (pH 4.6), to characterize the voltammetric signal of thallium at an in situ plated bismuth film electrode. A well-defined stripping peak (Ep = −0.78 V) due to the oxidation of thallium at the BFE was obtained. The shape of a typical voltammogram, which is illustrated in Fig. 1, reflects the irreversibility of the electrode process. The effect of potential scan rate (ranging between 10 and 150 mV s−1 ) on the peak potential and the peak current of thallium was studied. A slightly positive shift in the peak potential was observed, which confirms the irreversibility of the redox process. The anodic peak heights markedly increased with increasing scan rate, yielding a linear relationship between peak
Fig. 1. Typical cyclic voltammogram of 100 M Tl(I) + 1.9 M Bi(III) in 0.1 M acetate buffer (pH 4.6) at a glassy carbon rotating-disc electrode. Ei = −0.5 V (vs. Ag|AgCl), sweep rate: 100 mV s−1 , rotation speed: 240 rpm.
current and the scan rate up to 100 mV s−1 , with a correlation coefficient of 0.998. As expected for film electrodes, the signal of the target species is supposed to be affected by the film thickness. This can be controlled by varying the Bi(III) concentration in the test solution, at a fixed deposition time. Bismuth was co-deposited with thallium on the rotating-disc glassy carbon electrode at −1.4 V, for 120 s, after adding different bismuth(III) concentrations at the micromolar level to an acetate solution containing 0.1 M Tl(I). As shown by the obtained square wave stripping voltammograms exhibited in Fig. 2a, the thallium peak potential remained constant (at −0.76 V) up to 1.9 M Bi(III), and deviates to more negative values for higher Bi(III) concentrations. In which concerns peak current, Ip , it increased with increasing bismuth concentration between 0.6 and 1.9 M, and then decreased at higher concentrations (Fig. 2b). This behaviour, and also the peak shape observed for the higher bismuth concentrations, strongly suggest the occurrence of multilayer deposition, that may cause a weak adhesion of bismuth to the glassy carbon surface. A fixed Bi(III) concentration of 1.9 M, capable of generating reproducible bismuth film electrodes, which enable good sensitivity, was used in the analytical experiments. 3.2. Optimization of experimental conditions for anodic stripping voltammetric determination of thallium(I) at a BFE Aiming to establish the most suitable experimental conditions for the square wave anodic stripping voltammetric measurements of thallium(I) at the rotating-disc bismuth film electrode in acetate buffer, a univariate optimization study was performed with the deposition time, the deposition potential, the electrode rotation speed and the square wave parameters as variables. Square wave anodic stripping voltammograms of 0.1 M Tl(I) at the BFE were recorded, after holding the electrode at 6 different deposition times, starting from 60 s. Peak current increased rapidly with deposition time up to 120 s, and increased
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
1395
Fig. 2. (a) Square wave stripping voltammograms of 0.1 M Tl(I) at a glassy carbon rotating-disc electrode, in the presence of increasing levels of Bi(III). Supporting electrolyte: 0.1 M acetate buffer (pH 4.6). Edep : −1.4 V; tdep : 120 s; teq : 5 s; rotation speed: 240 rpm, square wave amplitude: 25 mV, Es : 5 mV, square wave frequency: 20 Hz. (b) The influence of Bi(III) concentration on thallium peak current.
more slightly for longer accumulation periods, reflecting the gradual saturation of the electrode surface. Hence, a deposition time of 120 s was adopted. The influence of the deposition potential was also analysed. A remarkable enhancement in sensitivity was observed upon changing the deposition potential from −0.9 to −1.5 V. At more negative potentials, the reduction of the positively charged thallium(I) ions is more effectively accomplished. A deposition potential of −1.4 V was chosen in all subsequent work. This value is in agreement with the suitable potential for co-deposition of bismuth films. The effect of increasing mass transport to the in situ plated BFE was examined. The flux of solution to the electrode surface, generated by forced convection, causes a significant improvement in sensitivity. For example, an 8-fold enhancement of the thallium peak is observed at an electrode rotation speed of 480 rpm, when compared with the one obtained at static conditions. According to Levich equation [43], which predicts that the thickness of the diffusion layer decreases linearly with the increment of the square root of the electrode rotation speed, ω, a linear plot of thallium stripping peak current versus ω1/2 was obtained from 60 to 1600 rpm, with a correlation coefficient of 0.998, confirming that the overall diffusion rate of the target species to the electrode surface is controlled by convective mass transfer. However, at high rates of convective mass transport (ω > 1800 rpm), the electrode response became poorly reproducible, exhibiting oscillations in current. Despite the enhanced sensitivity achieved using a hydrodynamic electrode, these results indicate that, at high rotation speeds, the BFE is more susceptible to mechanical disruption during convective accumulation with obvious implications on its analytical performance. The speed of 240 rpm was selected as giving the best compromise between sensitivity, mechanical stability and reproducibility of the bismuth film.
The normal square wave parameters frequency, scan increment and amplitude were also optimised in order to perform fast analysis with high sensitivity. The variation of peak current with square wave frequency is illustrated in Fig. 3. There is a considerable increase up to 40 Hz, but at higher frequencies the current starts to increase slowly, tending to level off, owing to the influence of the capacitive background current on the total measured current. As a result of this hindering capacitive contribution, sensitivity decreases. A square wave frequency of 40 Hz was chosen so as to provide large peaks with a good definition and fast scan rates. Scan increment, Es , was varied from 1 to 10 mV. No significant improvement was achieved above 5 mV; although peak currents were slightly increased, peak broadening occurred.
Fig. 3. The dependence of the stripping response of Tl(I) at a BFE on the square wave frequency. [Bi(III)] = 1.9 M. Other conditions as in Fig. 2.
1396
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
The best value of scan increment consistent with both adequate peak definition and evaluation was therefore 5 mV. In combination with the selected frequency, a scan rate of 200 mV s−1 was produced. The effect of square wave amplitude was studied in the range 10–140 mV. Greater peak currents were recorded for higher square wave amplitudes. Although a linear plot was obtained, deviation from linearity was observed at amplitudes above 90 mV. The selected square wave parameters for quantitative measurements of traces of thallium(I), in 0.1 M acetate buffer, by square wave ASV at an in situ plated rotating-disc bismuth film electrode were: square wave frequency, 40 Hz; scan increment, 5 mV; square wave amplitude, 80 mV. Equilibration time (teq ) values up to 5 s proved to have a very slightly positive effect on the electrode signal (about 4%); higher teq gave rise to lower peak heights. An equilibration time of 5 s was chosen.
important for the reliability of data in future analytical applications. For this purpose, a test solution containing 0.1 M Tl(I) in 1.9 M Bi(III)/0.1 M acetate buffer was employed. The electrode was continuously rotating at 240 rpm. A series of 15 consecutive measurements performed every 30 min, after 120 s co-deposition of thallium and bismuth at −1.4 V, yielded a very stable response for thallium with a relative standard deviation of 0.2%, with a mean peak current of 29.5 A. Between runs, the electrode was potentiostatically conditioned at +0.3 V for 30 s, to remove the metallic film, before renewing it. Well-shaped and reproducible peaks have been obtained with no apparent deterioration of the signal quality for at least 7 h. We can conclude that the rotating-disc bismuth film electrode plated in situ has very good reproducibility and allows a satisfactory repeatability of the results.
3.3. Calibration curve and precision
Possible interferents include electroactive trace metals. Lead(II) and cadmium(II) are generally considered as the major interferences in the determination of thallium on mercury [9,45] and carbon electrodes [10] by ASV. The use of the BFE significantly contributes to overcome this problem. Successive additions of lead acetate were performed in a background electrolyte solution containing 0.1 M Tl(I) and 1.9 M Bi(III). The stripping voltammograms have exhibited excellent peak resolution, with stripping potentials occurring at −0.79 and −0.62 V, for Tl(I) and Pb(II), respectively. No interference has been found in ASV of thallium up to 200 M Pb(II). The interference of cadmium was found to be more problematic. Cadmium and thallium peaks partially overlap, since the corresponding oxidation potentials are in close proximity to one another. However, the oxidation processes of Cd and
Under the experimental conditions selected on the optimisation studies, a calibration plot of thallium stripping peak current, measured at −0.8 V, against concentration was constructed using standard additions (Fig. 4). There is a close fit to linearity from 12 to 150 nM. Least-squares treatment yielded a slope of 188 ± 4.01 A/M, an intercept of 0.855 ± 0.337 A and a correlation coefficient of 0.998 (N = 12). For more diluted solutions, the plotted values do not fit the straight line. For concentrations under 1.2 nM the results are poorly reproducible. The calculated limits of detection (LOD) and quantification (LOQ) [44] were 10.8 and 36.1 nM, respectively. The assessments of the reproducibility of the generated bismuth film electrode itself as well as its response are very
3.4. Interferences of Pb(II) and Cd(II)
Fig. 4. Square wave voltammetry at a BFE for increasing concentrations of Tl(I). Ei = −1.0 V to Ef = −0.5 V (vs. Ag|AgCl), square wave frequency = 40 Hz, square wave amplitude = 80 mV. Other conditions as in Fig. 2. Also shown on the right is the resulting calibration plot.
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
1397
Fig. 5. Square wave ASV of 0.1 M Tl(I) + 0.1 M Pb(II) at a BFE (a) in the absence (solid line) and presence (broken lines) of Cd(II) (b) in the presence of 0.19 M Cd(II) + 1.0 mM EDTA. Other conditions as in Fig. 4.
Tl at the BFE, in acetate buffer, lead to a slight improvement in peak resolution when compared with that obtained at bare carbon electrodes [10]. As seen in Fig. 5a, stripping voltammograms registered after making successive additions of Cd(II) in a solution containing fixed concentrations of Tl(I) and Pb(II) show that Pb(II) peak remains practically inalterable, while Tl(I) peak raises with increasing cadmium concentration. Although exhibiting poor resolution, it should be pointed out that no shift of peak potentials occurred for thallium and cadmium. The interference of Pb(II) in the ASV of Tl(I) is actually largely overcome with the use of the proposed BFE. However, when cadmium is present in the solutions under analysis, the
addition of complexing agents is recommended [10,14,45] to minimize the interference of cadmium in ASV measurements of thallium. As shown in Fig. 5b, the addition of 1.0 mM EDTA in acetate buffer enabled the detection of 0.1 M Tl(I), with no cadmium and lead interferences. The presence of excess of Cd(II) and Pb(II) in the quantification of 98 nM Tl(I) at the BFE was allowed for ratios of Cd:Tl and Pb:Tl of 2500:1 and 10,000:1, respectively, which mean an important improvement in comparison with the data obtained using different electrode materials [9,10,45]. The interfering effects of Cu(II) and Zn(II) were also examined. In the presence of 1.0 mM EDTA, for 98 nM Tl(I), zinc is tolerated when present at 1000-fold
Fig. 6. Standard additions for (a) river water and (b) soil extract samples in acetate buffer. Curves: A, sample; B, C, D and E, sample +98, 120, 180 and 240 nM, respectively, of Tl(I). Other conditions as in Fig. 4.
1398
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399
more than the Tl(I) concentration, and copper over 100-fold more.
toxic chemical species in environmental matrices using a clean methodology.
3.5. Application to real samples
Acknowledgements
The analytical utility of the rotating-disc BFE for the determination of Tl(I) in environmental samples was tested with river Tagus water and soil samples collected from polluted sites of the river bank on the outskirts of Lisbon, where a cement plant is in operation nearby. It is interesting to check the thallium levels at this place, since there is some population density together with some agricultural and fishing activities. The anthropogenic activities may cause enrichment of Tl in the environment, leading to abnormally high levels in natural matrices, as reported elsewhere [46]. The suitability of the BFE for the determination of thallium in the samples is illustrated in Fig. 6. The thallium peak in river water and soil extract samples can thus be easily quantified following four standard additions of thallium nitrate solution. The considerably high thallium concentration found in the river water and soil samples, 2.2 ± 0.1 g L−1 and 15.0 ± 0.6 ng g−1 , respectively, has confirmed the suspicion of heavy pollution in this case study. These results demonstrated good agreement compared with independent analysis by graphite furnace atomic absorption spectroscopy (AAS): 2.5 ± 0.64 g L−1 and less than 30 ng g−1 , for the water and soil samples, respectively. No appreciable cadmium content was found using the reference method (LOQ for the AAS method: 0.5 g L−1 Cd). Each analytical measurement was replicated five times using both the ASV and AAS methods. The high lead content found in the river water (not calculated) is clearly shown by the presence of a sharp stripping peak at ∼−0.6 V in the recorded voltammograms (see Fig. 6a). It is well-known that lead is a contaminant usually present in environmental matrices. The significantly smaller amount of lead shown by the voltammogram obtained for the soil extract was attributed to the very low solubility of lead compounds in water. Thus, lead was not easily transferred to the aqueous phase during the soil treatment procedure employed in the present work.
The authors acknowledge FCT (Fundac¸a˜ o para a Ciˆencia e a Tecnologia) for support. We thank Dr. M.A. Trancoso from INETI for the AAS measurements.
4. Conclusions We have demonstrated that BFE’s are very suitable to determine trace amounts of thallium in environmental samples by square wave ASV, avoiding the use of mercury. The presented data strongly suggest that BFE will play an important role in electroanalysis in the near future, allowing direct measurements in non-deaerated solutions. The potentialities to operate in flow systems are reinforced by the success of its performance under hydrodynamic conditions, based on the use of the rotating-disc electrode. Ion-exchanger polymeric coatings on the bismuth films [9] should be considered in order to improve the voltammetric signal, lowering detection limits. The combination of low toxicity of bismuth with good performance of the rotating-disc BFE in voltammetric measurements makes it an attractive and promising sensor to monitor
References [1] I. Liem, G. Kaiser, M. Sager, G. T¨olg, Anal. Chim. Acta 158 (1984) 179. [2] R.J.P. Williams, J.J.R. Fra´usto da Silva, The Natural Selection of the Chemical Elements, Clarendon Press, Oxford, 1996, p. 626. [3] A. Leonard, G.B. Gerner, Mutat. Res. 387 (1997) 47. [4] R.C. Carpenter, Anal. Chim. Acta 125 (1981) 209. [5] M.A. Floyd, V.A. Fassel, R.K. Winge, J.M. Katzenberger, A.P.D. Silva, Anal. Chem. 52 (1980) 431. [6] W. Schmidt, V. Dietl, Fresenius Z. Anal. Chem. 315 (1983) 690. [7] P. Schramel, I. Wendler, J. Angerer, Int. Arch. Occup. Environ. Health 69 (1997) 219. [8] A. Ciszewski, Z. Lukaszewski, Anal. Chim. Acta 146 (1983) 51. [9] T.-H. Lu, H.-Y. Yang, I. Wen Sun, Talanta 49 (1999) 59. [10] N. Spano, A. Panzanelli, P.C. Piu, M.I. Pilo, G. Sanna, R. Seeber, A. Tapparo, Anal. Chim. Acta 553 (2005) 201. [11] A.K. Das, M. Dutta, M.L. Cervera, M. de la Guardia, Microchem. J. 86 (2007) 2. [12] R. Cleven, L. Fokkert, Anal. Chim. Acta 289 (1994) 215. [13] I. Svancara, P. Ostapczuk, J. Arunachalam, H. Emons, K. Vytras, Electroanalysis 9 (1997) 26. [14] J. Wang, Stripping Analysis, VCH, New York, 1995. [15] T.R. Copeland, R.K. Skogerboe, Anal. Chem. 46 (1974) 1257A. [16] B. Krasnodebska-Ostrega, E. Stryjewska, Chem. Anal. (Warsaw) 49 (2004) 519. [17] E. Gustafsson, Water Air Soil Pollut. 80 (1995) 99. [18] E.P. Achterberg, C. Braungardt, Anal. Chim. Acta 400 (1999) 381. [19] O.M.S. Filipe, C.M.A. Brett, Talanta 61 (2003) 643. [20] J. Wang, B. Tian, Anal. Chem. 65 (1993) 1529. [21] M.A. Nolan, S.P. Kounaves, Anal. Chem. 71 (1999) 3567. [22] Y.-C. Tsai, B.A. Coles, K. Holt, J.S. Foord, F. Marken, R.G. Compton, Electroanalysis 13 (2001) 831. [23] J. Wang, J. Lu, S.B. Hocevar, P.A.M. Farias, Anal. Chem. 72 (2000) 3218. [24] J. Wang, J. Lu, U.A. Kirgoz, S.B. Hocevar, B. Ogorevc, Anal. Chim. Acta 434 (2001) 29. [25] T.M. Florence, J. Electroanal. Chem. 27 (1970) 273. [26] W. Frenzel, Anal. Chim. Acta 273 (1993) 123. [27] N.F. Zakharchuk, Kh.Z. Brainina, Electroanalysis 10 (1998) 379. [28] F. Ribeiro, M.M.M. Neto, M.M. Rocha, I.T.E. Fonseca, Anal. Chim. Acta 579 (2006) 227. [29] A. Kr´olicka, A. Bobrowski, Electrochem. Commun. 6 (2004) 99. [30] M. Yang, Z. Hu, J. Electroanal. Chem. 583 (2005) 46. [31] J. Wang, J. Lu, S.B. Hocevar, B. Ogorevc, Electroanalysis 13 (2001) 13. [32] Z. Guo, F. Feng, Y. Hou, N. Jaffrezic-Renault, Talanta 4 (2005) 1052. [33] G. Kefala, A. Economou, A. Voulgaropoulos, M. Sofoniou, Talanta 61 (2003) 603. [34] M. Morfobos, A. Economou, A. Voulgaropoulos, Anal. Chim. Acta 519 (2004) 57. [35] G. Kefala, A. Economou, A. Voulgaropoulos, Analyst 129 (2004) 1082. [36] I. Svancara, L. Baldrianova, E. Tesarova, S.B. Hocevar, S.A.A. Elsuccary, A. Economou, S. Sotiropoulos, B. Ogorevc, K. Vytras, Electroanalysis 18 (2005) 177. [37] H. Dong, H. Zheng, L. Lin, B. Ye, Sens. Actuators B 115 (2006) 303. [38] Y. Bonfil, M. Brand, E. Kirowa-Eisner, Electroanalysis 15 (2003) 1369. [39] M.M.G.S. Rocha, M.M.P.M. Neto, M.O. Torres, A. de Varennes, Electroanalysis 9 (1997) 145. [40] M.M.P.M. Neto, M.M.G.S. Rocha, C.M.A. Brett, Talanta 41 (1994) 1597.
E.O. Jorge et al. / Talanta 72 (2007) 1392–1399 [41] E. Chatzitheodorou, A. Economou, A. Voulgaropoulos, Electroanalysis 16 (2004) 1745. [42] B. Griepink, M. Sager, G. T¨olg, Pure Appl. Chem. 60 (1988) 9. [43] V.G. Levich, Physicochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs, NJ, 1962.
1399
[44] J.N. Miller, J.C. Miller (Eds.), Statistics and Chemometrics for Analytical Chemistry, fourth ed., Prentice-Hall, Upper Saddle River, NJ, 2000 (Chapter 5). [45] J.-M. Zen, J.-W. Wu, Electroanalysis 9 (1997) 302. [46] A.L. John Peter, T. Viraraghavan, Environ. Int. 31 (2005) 493.