- Email: [email protected]
Anodic stripping voltammetric determination of bismuth(III) using a Tosflex-coated mercury film electrode
Talanta 50 (1999) 977 – 984 www.elsevier.com/locate/talanta
Anodic stripping voltammetric determination of bismuth(III) using a Tosflex-coated mercury film electrode Hao-Yun Yang, Wen-Yin Chen, I-Wen Sun * Department of Chemistry, National Cheng-Kung Uni6ersity, Tainan, 70101, Taiwan, ROC Received 5 March 1998; received in revised form 14 June 1999; accepted 18 June 1999
Abstract A Tosflex–mercury film electrode (TMFE) was prepared by spin-coating a solution of the perfluorinated anion exchange polymer Tosflex onto a glassy carbon electrode surface followed by electrodeposition of mercury film on this electrode. This electrode was used for the determination of trace bismuth(III) which was preconcentrated onto the TMFE as anionic bismuth(III) complexes with chloride in a chloride medium. The preconcentration was carried out at a potential of−0.2 V, and the preconcentration of the bismuth(III) was enhanced significantly by the anion-exchange feature of Tosflex. The accumulated bismuth(III) was then determined by anodic square-wave stripping voltammetry (SWSV). Various parameters influencing the determination of bismuth(III) were examined in detail. With 2 min accumulation, the analytical signal versus concentration dependence was linear up to 50 ppb, and the detection limit was 0.58 ppb. This modified electrode showed good resistance to the interferences from surface-active compounds and common ions. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Bismuth; Stripping voltammetry; Electroanalysis; Tosflex
1. Introduction Anodic stripping voltammetry is a sensitive technique for the determination of traces of bismuth(III) in aqueous media [1 – 4]. For example, ng of Bi(III) could be detected using a glassy carbon rotating disk electrode with square-wave anodic stripping voltammetry [4]. This method, however, is time-consuming: a 20 min precondition procedure was required before a 5 min (or 20 min) accumulation and stripping of Bi(III) could * Corresponding author. Fax: +886-06-2740552. E-mail address: [email protected] (I.-W. Sun)
be carried out. Extensive efforts have been devoted to improve the anodic stripping voltammetric determination of Bi(III). Several carbon paste chemically modified electrodes containing a complexing reagent with chemical affinity for Bi(III) have been developed for the determination of Bi(III) [5–9]. These procedures, however, do not achieve the detection limit obtained with mercury electrode. The applicability of stripping analysis can be improved by the use of ion-exchange polymer modified electrodes. The ion-exchange polymer modified electrode exhibit enhanced preconcentration efficiency for analytes, and is less subject to
0039-9140/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 1 8 5 - X
978
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
interferences from surface-active compounds. Numerous examples on the use of cation-exchanger, Nafion, modified electrodes for the determination of cations have been published [10,21,22]. While most species of analytical interest exist in the form of cations in sample solutions, Bi(III) is mostly present in anionic forms. Apparently, an anionexchange polymer modified electrode would be particularly appealing for the determination of anionic Bi(III). Zen et al. [11] have developed a poly(4-vinylpyridine) – mercury film modified glassy carbon electrode (PVP – MFE) to determine bismuth in the form of [BiCl4]−. Although the PVP–MFE exhibits good resistance to interferences from surface-active compounds and a good detection limit, the use of PVP – MFE presents some problems. Because PVP dissolves easily in acidic solution, a cross-link agent was required to stabilize the PVP film and the PVP coated electrode needed to be heated at 90°C for about 2 h for the cross-link process to complete [12]. Moreover, the PVP – MFE is effective only in fairly acidic solution because the PVP polymer is protonated more completely in more acidic solutions [13]. Relatively recently, a new class of perfluorinated anion-exchange polymer named Tosflex (Tosoh Soda) has become commercially available [14]. Similar to Nafion, Tosflex exhibits very good stability and is easy to use. Thus, Tosflex can be an ideal alternative to PVP for the determination of anions. However, reports on the employment of Tosflex film modified electrodes for the determinations of trace metal ions are relatively limited. Only several examples including copper(I), mercury(II), thallium(III) and tellurium(IV) have been demonstrated recently [15 – 18]. This paper describes a square-wave stripping voltammetric (SWSV) procedure for the determination of bismuth(III) by using a Tosflex–mercury film modified electrode (TMFE). Anionic bismuth(III) chloride complex is first accumulated on the electrode surface by the ion exchange feature of the TMFE and followed by anodic SWSV measurement. Various factors influencing the determination of bismuth(III) were investigated.
2. Experimental
2.1. Apparatus All electrochemical experiments were performed with a Bioanalytical Systems BAS CV-50W electrochemical analyzer in conjunction with a BAS model C-2 electrochemical cell. The three-electrode system consisted of a glassy carbon disk working electrode (BAS, 3 mm diameter) coated with Tosflex and mercury, a saturated Ag–AgCl reference electrode (BAS), and a platinum spiral auxiliary electrode. All glassware was cleaned with 1:1 nitric acid and rinsed with deionized water.
2.2. Chemicals and reagents The Tosflex membrane, denoted IE-SA 48, was obtained from Tosoh Soda, Japan. Sodium perchlorate, sodium nitrate and nitric acid were of analytical grade from Riedel-de Haen (RDH). Standard metal solutions (1000 ppm) of Zn(II), Ge(IV), Cr(VI), Cu(II), and Cd(II) were from Fisher. Standard solutions (1000 ppm) of Se(IV) and bismuth(III) were from Mallinckrodt. Hg(II) standard solution (1000 ppm) was from Merck. The nonionic surfactant Triton X-100 was received from Lancaster. All the preparation and dilution of solutions were made with deionized water.
2.3. Preparation of Tosflex–mercury film electrode and sample solution The dissolution of Tosflex membrane was carried out according to the procedure described in the literature [14]. About 2.5 g of finely cut dry membrane and 10 ml of water–methanol–2propanol aqueous–alcoholic solution was heated to the boiling point at atmospheric pressure under reflux and stirring for 20–50 h. After cooling, the undissolved membrane was separated by centrifugation and a clear, yellowish solution was collected. The concentration of the dissolved polymer was determined gravimetrically from an evaporated portion of the solution. The coating solution was brought to a final concentration of 1.2 wt% by dilution with methanol.
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
After being polished with a polishing cloth to a shiny surface, the glassy carbon electrode (GCE) was rinsed with deionized water and then cleaned ultrasonically in 1:1 nitric acid and deionized water. Then, 4 ml of Tosflex coating solution was spincoated onto the GCE at a spin rate of 3000 rpm. A uniform thin film was formed by evaporating the solvent after about 3 min of spinning. Mercury was electrodeposited on the Tosflex coated GCE from 5 ml of 10 ppm mercury(II) solution containing 0.1 M sodium chloride at an applied potential of − 0.8 V versus Ag – AgCl for 4 min with stirring. The sample solution medium contained bismuth(III) and proper amounts of KCl as the supporting electrolyte. The solution pH was adjusted with 0.05 M HNO3 and 0.05 M NaOH. Groundwater and tap water were collected from the campus of National Cheng-Kung University. Seawater was collected from the beach of Tainan, Taiwan. All water samples, except seawater, were added with 0.5 M KCl as the supporting electrolyte. All sample were passed through the c 1 filter paper and then adjusted to the pH 1.4 by nitric acid. Urine samples were obtained from laboratory personnel. After being passed through a filter with 0.45 mm pore size, the urine solutions were further treated by centrifugation at 12 000 rpm for 20 min.
979
(III) solution containing 0.5 M KCl and nitric acid as the supporting electrolyte, the Osteryoung square wave stripping voltammogram (OSWSV) shown in Fig. 1(a) is obtained. In this figure, a stripping peak was observed at − 80 mV. Fig. 1(b), on the other hand, shows the OSWSV for the same solution recorded on the TMFE. In this figure the stripping peak shifts to−110 mV. As can be clearly seen in Fig. 1, the response of Bi(III) on the TMFE is significantly higher than that on the MFE. This fact implies that bismuth can be accumulated more efficiently on the TMFE than on the MFE, and thus, the TMFE is more suitable for the determination of bismuth. Fig. 2 shows the cyclic voltammograms of bismuth on the TMFE in different supporting electrolytes. As can be seen in this figure, the bismuth response detected by the TMFE in a medium containing nitrate ion is much smaller than that detected in a medium containing chloride ion
2.4. Procedure for 6oltammetric measurement In the voltammetric measurement, 5 ml of the sample solution was placed in the voltammetric cell. The solution was deaerated with argon for 5 min, and then a preconcentration potential was applied to the fleshly prepared TMFE while the solution was stirred. After the preconcentration period, the stirring was stopped, and after 10 s the Osteryoung square-wave stripping voltammogram was recorded by applying a positive-going scan. 3. Results and discussion
3.1. Electrochemical beha6ior of bismuth(III) on the Tosflex modified electrodes When a mercury film modified glassy carbon electrode (MFE) is dipped in a 10 ppm bismuth
Fig. 1. Osteryoung square wave voltammogram for a 10 ppm bismuth(III) solution containing 0.5 M KCl, (a) at a mercury film coated glassy carbon disk electrode; (b) at the Tosflex– mercury film electrode (TMFE). SWSV parameters, modulation amplitude 20 mV; modulation frequency 120 Hz.
980
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
muth stripping peak current was found to increase with the spin-coating rate, until 3000 rpm, which is the limit spin rate of the instrument. Electrodes prepared with the coating solution of 1.2 wt% of Tosflex at a 3000 rpm spin-coating rate for 3 min were therefore used in all subsequent experiments. The amount of mercury electrodeposited on to the Tosflex coated GCE is dependent on the deposition time. The effect of mercury deposition on the electrode performance was evaluated with 10 ppm mercury (II) standard solutions. The results showed that the bismuth stripping peak current increased upon increasing the mercury deposition time and reached a maximum after 4 min. Therefore, a mercury deposition time of 4 min was used in the subsequent works. Fig. 2. Cyclic voltammograms for 10 ppb bismuth(III) in 0.05M HNO3 (dashed line) and 0.5M KCl and 0.05M HNO3 (solid line) on TMFE. Scan rate was 100 mV s − 1.
(wave a in Fig. 2). Apparently, in the nitrate medium, the bismuth (III) ion is in the cationic form and thus, cannot be effectively accumulated by the anionic exchanger, Tosflex, of the TMFE during the preconcentration step. In contrast, anionic bismuth (III) complexes, are formed in the chloride medium and can be effectively accumulated into the TMFE during the preconcentration step. Moreover, the bismuth stripping peak obtained in the chloride media occurs at a potential more negative than that obtained in the nitrate media. As shown in Fig. 2, this potential shift was similar to what was observed for the oxidation of mercury (wave b in Fig. 2) which also had been described in the literature [19].
3.2. Factors affect the performance of the TMFE The thickness of the Tosflex film directly affects the electrode performance since it controls the diffusion process and the maximum loading of the bismuth(III) chloride anions in the TMFE. The film thickness was varied by preparing the electrodes with different spin-coating rate. The bis-
3.3. Effect of solution pH The dependence of the bismuth stripping peak current on the pH was studied by maintaining the chloride concentration at 0.1 M and the results obtained show that the bismuth stripping peak current gradually increases with decreasing the pH and starting level off when the pH is 1.4. This behavior suggests that the formation of bismuth(III) chloride is more favorable in an acidic environment. As a result, solutions with pH 1.4 were used in the subsequent experiments. It is worth to note that at this pH value Tosflex film was still very stable.
3.4. Effect of chloride ion concentration The detection of bismuth with TMFE relies on the effective formation of the bismuth(III) chloride anions. Fig. 3 presents the effect of changing the chloride ion concentration on the bismuth stripping peak current. The optimum concentration range of chloride ion is found to be around 0.5 M. Lowering the chloride ion concentration results in lower signals, indicating that a reasonable excess of chloride ions is required for converting the bismuth (III) into its chloride complexes. The decrease of the stripping peak
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
981
current at chloride concentration higher then 0.5 M might be attributed to failures in permselectivity, which were previously observed at high ionic strengths also for Nafion coated electrode [20]. A 0.5 M concentration of chloride ion was therefore employed in the subsequent experiments.
3.5. Effect of square-wa6e parameters The square-wave parameters that were investigated were the pulse height and the frequency. These parameters together affect the peak shape and peak current of the bismuth response. The peak current of bismuth increases with squarewave pulse height up to 20 mV, and further increase in the pulse height decreases and broadens the bismuth stripping peak. Meanwhile, considerable increase in the background current is also observed when the pulse height exceeds 20 mV. Increases in the square-wave frequency up to 120 Hz results in a substantial increase in the
Fig. 4. Effect of preconcentration time on the SWSV response for 10 ppb bismuth(III) obtained on TMFE. Ed = −0.2 V s; KCl = 0.5 M; pH=1.4.
response for bismuth. Overall, the optimum pulse height and frequency selected for SWSV were 20 mV and 120 Hz, respectively.
3.6. Effect of preconcentration potential and time
Fig. 3. Effect of chloride ion concentration on the SWSV response for 10 ppb bismuth(III) obtained on TMFE. Ed = − 0.2 V; td = 120 s; pH= 1.4.
The dependence of the bismuth stripping peak current on the preconcentration potential is studied and the results show that the peak current increases as the preconcentration potential becomes more negative between− 0.15 and − 0.4 V and starts to level off as the preconcentration potential becomes more negative than− 0.4V. Because a more negative preconcentration potential would increase the possibility of co-deposition of interfering species, a preconcentration potential of− 0.2 V was chosen in the subsequent work. The effect of preconcentration time on the SWSV response of bismuth was studied, and the results are displayed in Fig. 4. The peak current increases with increasing preconcentration time up to 6 min. This phenomenon indicates that the
982
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
ion-exchange process between Tosflex and bismuth(III) chloride anions is very good. Although this figure shows that higher sensitivity for the detection of lower bismuth concentration can be achieved by increasing the preconcentration time, a preconcentration time of 2 min was chosen for the construction of the calibration curve in this study.
3.7. Calibration A calibration graph was constructed from data taken for bismuth(III) solutions following 2 min preconcentration under optimum experimental conditions described above. Fresh sample solutions were used for each individual bismuth(III) concentration and at least three independent determinations were made for each data point. The calibration plot thus obtained showed a linear behavior between 0 and 50 ppb of bismuth(III) with slope (mA/ppb), and correlation coefficient of 2.99, and 0.995, respectively. The detection limit (S/N = 3) is 0.58 ppb. An even lower detection limit could be achieved for bismuth(III) provided that the preconcentration time is longer than 2 min.
Table 1 Influence of other ions on the response of Bi(III) at TMFEa Ions
Pb(II) Cd(II) Te(IV) Zn(II) Ge(IV) Cu(II) Sb(III) Cr(VI)
Concentration excess
Contribution (%)
Over Bi(III)
(iBi(III) =100%)
1000× 1000× 1000× 1000× 1000× 500× 100× 100×
−2 −3 −6 +3 +7 −17 +16 +3
a
(Bi(III))=10 ppb; Ep =−0.2 V; tp =120 s; pH, 1.4; KCl = 0.5 M.
the organic interferences from reaching the interface at which the deposition and stripping of bismuth takes place. In this study, the nonionic surfactant Triton X-100 was used to exemplify the effect of a typical surfactant. As shown in Fig. 5, for 10 ppb bismuth(III), the detection was found to tolerate the presence of Triton X-100 for at least up to 10 ppm with the TMFE. Compared to the same experiments performed with a bare MFE
3.8. Interferences The influences of various metal ions on the determination of bismuth were examined. The results obtained for 10 ppb bismuth with 2 min preconcentration time are summarized in Table 1. This table shows that over 1000-fold wt. excess concentrations of zinc(II), germanium(IV), tellurium(IV), lead(II), cadmium(II) and 500-fold excess of copper(II) and 100-fold excess of chromium(VI), antimony(III) only slightly interfere with the bismuth response. The only major interfering ion among the ions studied was selenium(IV). It is well documented that surface-active compounds can often adsorb on the electrode surface and reduce the analytical response of the analyte in a stripping analysis using a bare mercury electrode. Coating the electrode with Tosflex can circumvent such interferences. The Tosflex membrane coated on the electrode surface can prevent
Fig. 5. Effect of the surfactant Triton X-100 at different concentrations on the stripping response for 10 ppb bismuth(III) with the TMFE (solid line) and the bare MFE (dashed line). KCl =0.5 M, pH =1.4, td =120 s, Ed = − 0.2 V. Normalised peak currents are calculated as the ratio between signals recorded in the presence and in the absence of surfactant.
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
983
Table 2 Determination of Bi(III) in sea water, tap water, ground water, and urine samplesa
Detected value original Spiked (ppb) Detected value after spike (ppb) Recovery (%) a b
Sea water
Tap water
Ground water
Urineb
ND 4 3.889 0.04 97
ND 1 1.02 9 0.02 102
ND 2 1.99 9 0.03 99
ND 5 5.04 9 0.12 101
Number of samples assayed = 3. Deposition potential = −0.6 V.
(Fig. 5), the TMFE shows a much better resistance towards surfactant interference.
3.9. Determination of bismuth in real water samples The analytical utility of the TMFE for the determination of bismuth(III) was assessed by applying it to the determination of bismuth(III) in sea water, tap water, groundwater, and urine samples. No bismuth(III) was detected in all the four original water samples so they were spiked with appropriate amounts of bismuth(III). The results collected in Table 2 are those for the original and spiked water samples. As can be seen, the recovery of the spiked bismuth(III) is very good for all the four water samples, indicating that the proposed procedure is feasible for the determination of bismuth(III) in various water samples. Note that the amount of bismuth(III) in natural water is typically very low, and this is indeed the case in this study. The amount of bismuth(III) in the original urine samples assayed in this study can not be detected by the proposed procedure with 2 min preconcentration at a potential of − 0.2V. Nevertheless, it was found that the spiked 5 ppb bismuth(III) stripping peak could be actually seen for these samples with a precontration potential of−0.6 V, and the amount of bismuth(III) in these original water samples was therefore believed to be well below 5 ppb.
4. Conclusion The application of the TMFE for the determination of trace bismuth(III) is evident from the
above results. The presence of Tosflex polymeric coating improves the efficiency and selectivity of the preconcentration step, as well as the mechanical stability of the mercury film. Compared to the PVP–MFE procedure reported earlier [11], the TMFE offers some advantages in addition to better resistance to the interferences from common ions and organic surface-active compound. Because of the intrinsic stability of Tosflex, the preparation of the TMFE does not require any cross-liking agent nor the heating process, and thus, the time for preparing the working electrode is greatly reduced. Consequently, the TMFE would be an ideal substitute for the PVP–MFE.
Acknowledgements The authors gratefully acknowledge the financial support of the National Science Council of the Republic of China (Taiwan) under Grants NSC87-2815-C-006-073M.
References [1] T.M. Florence, J. Electroanal. Chem. 35 (1972) 237. [2] T.R. Gilbert, D.N. Hume, Anal. Chim. Acta 65 (1973) 451. [3] T.M. Florence, J. Electroanal. Chem. 49 (1974) 255. [4] S. Komorsky-Lovric, Anal. Chim. Acta 204 (1988) 161. [5] J. Lexa, K. Stulik, Talanta 32 (1985) 1027. [6] K. Kalcher, Fresenius Z. Anal. Chem. 325 (1986) 186. [7] K.L. Dong, L. Kryger, J.K. Christensen, K.N. Thomsen, Talanta 38 (1991) 101. [8] I. Svancara, K. Vytras, Anal. Chim. Acta 273 (1993) 195. [9] C. Wang, Q. Sun, H. Li, Electroanalysis 9 (1997) 645. [10] D.W.M. Arrigan, Analyst 119 (1994) 1953. [11] J.M. Zen, M.J. Chung, Anal. Chim. Acta 320 (1996) 43.
984
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
[12] B. Lindholm, M. Sharp, J. Electroanal. Chem. 198 (1986) 37. [13] J.M. Zen, M.J. Chung, Anal. Chem. 67 (1995) 3571. [14] L. Dunsch, L. Kavan, J. Weber, J. Electroanal. Chem. 280 (1990) 313. [15] P. Ugo, L.M. Moretto, G.A. Mazzocchin, Anal. Chim. Acta 273 (1993) 229. [16] P. Ugo, L.M. Moretto, G.A. Mazzocchin, Anal. Chim. Acta 305 (1995) 74.
.
[17] T.-H. Lu, I.-W. Sun, Electroanalysis 10 (1998) 1052. [18] H.-Y. Yang, I.-W. Sun, Electroanalysis 11 (1999) 195. [19] L.M. Moretto, G.A. Mazzocchin, P. Ugo, J. Electroanal. Chem. 427 (1997) 113. [20] R. Nagegeli, J. Redepenning, F.C. Anson, J. Phys. Chem. 90 (1986) 6227. [21] P. Ugo, L.M. Moretto, Electroanalysis 7 (1995) 1105. [22] H.-Y. Yang, I.-W. Sun, Anal. Chim. Acta 358 (1998) 285.
Anodic stripping voltammetric determination of bismuth(III) using a Tosflex-coated mercury film electrode Hao-Yun Yang, Wen-Yin Chen, I-Wen Sun * Department of Chemistry, National Cheng-Kung Uni6ersity, Tainan, 70101, Taiwan, ROC Received 5 March 1998; received in revised form 14 June 1999; accepted 18 June 1999
Abstract A Tosflex–mercury film electrode (TMFE) was prepared by spin-coating a solution of the perfluorinated anion exchange polymer Tosflex onto a glassy carbon electrode surface followed by electrodeposition of mercury film on this electrode. This electrode was used for the determination of trace bismuth(III) which was preconcentrated onto the TMFE as anionic bismuth(III) complexes with chloride in a chloride medium. The preconcentration was carried out at a potential of−0.2 V, and the preconcentration of the bismuth(III) was enhanced significantly by the anion-exchange feature of Tosflex. The accumulated bismuth(III) was then determined by anodic square-wave stripping voltammetry (SWSV). Various parameters influencing the determination of bismuth(III) were examined in detail. With 2 min accumulation, the analytical signal versus concentration dependence was linear up to 50 ppb, and the detection limit was 0.58 ppb. This modified electrode showed good resistance to the interferences from surface-active compounds and common ions. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Bismuth; Stripping voltammetry; Electroanalysis; Tosflex
1. Introduction Anodic stripping voltammetry is a sensitive technique for the determination of traces of bismuth(III) in aqueous media [1 – 4]. For example, ng of Bi(III) could be detected using a glassy carbon rotating disk electrode with square-wave anodic stripping voltammetry [4]. This method, however, is time-consuming: a 20 min precondition procedure was required before a 5 min (or 20 min) accumulation and stripping of Bi(III) could * Corresponding author. Fax: +886-06-2740552. E-mail address: [email protected] (I.-W. Sun)
be carried out. Extensive efforts have been devoted to improve the anodic stripping voltammetric determination of Bi(III). Several carbon paste chemically modified electrodes containing a complexing reagent with chemical affinity for Bi(III) have been developed for the determination of Bi(III) [5–9]. These procedures, however, do not achieve the detection limit obtained with mercury electrode. The applicability of stripping analysis can be improved by the use of ion-exchange polymer modified electrodes. The ion-exchange polymer modified electrode exhibit enhanced preconcentration efficiency for analytes, and is less subject to
0039-9140/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 1 8 5 - X
978
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
interferences from surface-active compounds. Numerous examples on the use of cation-exchanger, Nafion, modified electrodes for the determination of cations have been published [10,21,22]. While most species of analytical interest exist in the form of cations in sample solutions, Bi(III) is mostly present in anionic forms. Apparently, an anionexchange polymer modified electrode would be particularly appealing for the determination of anionic Bi(III). Zen et al. [11] have developed a poly(4-vinylpyridine) – mercury film modified glassy carbon electrode (PVP – MFE) to determine bismuth in the form of [BiCl4]−. Although the PVP–MFE exhibits good resistance to interferences from surface-active compounds and a good detection limit, the use of PVP – MFE presents some problems. Because PVP dissolves easily in acidic solution, a cross-link agent was required to stabilize the PVP film and the PVP coated electrode needed to be heated at 90°C for about 2 h for the cross-link process to complete [12]. Moreover, the PVP – MFE is effective only in fairly acidic solution because the PVP polymer is protonated more completely in more acidic solutions [13]. Relatively recently, a new class of perfluorinated anion-exchange polymer named Tosflex (Tosoh Soda) has become commercially available [14]. Similar to Nafion, Tosflex exhibits very good stability and is easy to use. Thus, Tosflex can be an ideal alternative to PVP for the determination of anions. However, reports on the employment of Tosflex film modified electrodes for the determinations of trace metal ions are relatively limited. Only several examples including copper(I), mercury(II), thallium(III) and tellurium(IV) have been demonstrated recently [15 – 18]. This paper describes a square-wave stripping voltammetric (SWSV) procedure for the determination of bismuth(III) by using a Tosflex–mercury film modified electrode (TMFE). Anionic bismuth(III) chloride complex is first accumulated on the electrode surface by the ion exchange feature of the TMFE and followed by anodic SWSV measurement. Various factors influencing the determination of bismuth(III) were investigated.
2. Experimental
2.1. Apparatus All electrochemical experiments were performed with a Bioanalytical Systems BAS CV-50W electrochemical analyzer in conjunction with a BAS model C-2 electrochemical cell. The three-electrode system consisted of a glassy carbon disk working electrode (BAS, 3 mm diameter) coated with Tosflex and mercury, a saturated Ag–AgCl reference electrode (BAS), and a platinum spiral auxiliary electrode. All glassware was cleaned with 1:1 nitric acid and rinsed with deionized water.
2.2. Chemicals and reagents The Tosflex membrane, denoted IE-SA 48, was obtained from Tosoh Soda, Japan. Sodium perchlorate, sodium nitrate and nitric acid were of analytical grade from Riedel-de Haen (RDH). Standard metal solutions (1000 ppm) of Zn(II), Ge(IV), Cr(VI), Cu(II), and Cd(II) were from Fisher. Standard solutions (1000 ppm) of Se(IV) and bismuth(III) were from Mallinckrodt. Hg(II) standard solution (1000 ppm) was from Merck. The nonionic surfactant Triton X-100 was received from Lancaster. All the preparation and dilution of solutions were made with deionized water.
2.3. Preparation of Tosflex–mercury film electrode and sample solution The dissolution of Tosflex membrane was carried out according to the procedure described in the literature [14]. About 2.5 g of finely cut dry membrane and 10 ml of water–methanol–2propanol aqueous–alcoholic solution was heated to the boiling point at atmospheric pressure under reflux and stirring for 20–50 h. After cooling, the undissolved membrane was separated by centrifugation and a clear, yellowish solution was collected. The concentration of the dissolved polymer was determined gravimetrically from an evaporated portion of the solution. The coating solution was brought to a final concentration of 1.2 wt% by dilution with methanol.
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
After being polished with a polishing cloth to a shiny surface, the glassy carbon electrode (GCE) was rinsed with deionized water and then cleaned ultrasonically in 1:1 nitric acid and deionized water. Then, 4 ml of Tosflex coating solution was spincoated onto the GCE at a spin rate of 3000 rpm. A uniform thin film was formed by evaporating the solvent after about 3 min of spinning. Mercury was electrodeposited on the Tosflex coated GCE from 5 ml of 10 ppm mercury(II) solution containing 0.1 M sodium chloride at an applied potential of − 0.8 V versus Ag – AgCl for 4 min with stirring. The sample solution medium contained bismuth(III) and proper amounts of KCl as the supporting electrolyte. The solution pH was adjusted with 0.05 M HNO3 and 0.05 M NaOH. Groundwater and tap water were collected from the campus of National Cheng-Kung University. Seawater was collected from the beach of Tainan, Taiwan. All water samples, except seawater, were added with 0.5 M KCl as the supporting electrolyte. All sample were passed through the c 1 filter paper and then adjusted to the pH 1.4 by nitric acid. Urine samples were obtained from laboratory personnel. After being passed through a filter with 0.45 mm pore size, the urine solutions were further treated by centrifugation at 12 000 rpm for 20 min.
979
(III) solution containing 0.5 M KCl and nitric acid as the supporting electrolyte, the Osteryoung square wave stripping voltammogram (OSWSV) shown in Fig. 1(a) is obtained. In this figure, a stripping peak was observed at − 80 mV. Fig. 1(b), on the other hand, shows the OSWSV for the same solution recorded on the TMFE. In this figure the stripping peak shifts to−110 mV. As can be clearly seen in Fig. 1, the response of Bi(III) on the TMFE is significantly higher than that on the MFE. This fact implies that bismuth can be accumulated more efficiently on the TMFE than on the MFE, and thus, the TMFE is more suitable for the determination of bismuth. Fig. 2 shows the cyclic voltammograms of bismuth on the TMFE in different supporting electrolytes. As can be seen in this figure, the bismuth response detected by the TMFE in a medium containing nitrate ion is much smaller than that detected in a medium containing chloride ion
2.4. Procedure for 6oltammetric measurement In the voltammetric measurement, 5 ml of the sample solution was placed in the voltammetric cell. The solution was deaerated with argon for 5 min, and then a preconcentration potential was applied to the fleshly prepared TMFE while the solution was stirred. After the preconcentration period, the stirring was stopped, and after 10 s the Osteryoung square-wave stripping voltammogram was recorded by applying a positive-going scan. 3. Results and discussion
3.1. Electrochemical beha6ior of bismuth(III) on the Tosflex modified electrodes When a mercury film modified glassy carbon electrode (MFE) is dipped in a 10 ppm bismuth
Fig. 1. Osteryoung square wave voltammogram for a 10 ppm bismuth(III) solution containing 0.5 M KCl, (a) at a mercury film coated glassy carbon disk electrode; (b) at the Tosflex– mercury film electrode (TMFE). SWSV parameters, modulation amplitude 20 mV; modulation frequency 120 Hz.
980
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
muth stripping peak current was found to increase with the spin-coating rate, until 3000 rpm, which is the limit spin rate of the instrument. Electrodes prepared with the coating solution of 1.2 wt% of Tosflex at a 3000 rpm spin-coating rate for 3 min were therefore used in all subsequent experiments. The amount of mercury electrodeposited on to the Tosflex coated GCE is dependent on the deposition time. The effect of mercury deposition on the electrode performance was evaluated with 10 ppm mercury (II) standard solutions. The results showed that the bismuth stripping peak current increased upon increasing the mercury deposition time and reached a maximum after 4 min. Therefore, a mercury deposition time of 4 min was used in the subsequent works. Fig. 2. Cyclic voltammograms for 10 ppb bismuth(III) in 0.05M HNO3 (dashed line) and 0.5M KCl and 0.05M HNO3 (solid line) on TMFE. Scan rate was 100 mV s − 1.
(wave a in Fig. 2). Apparently, in the nitrate medium, the bismuth (III) ion is in the cationic form and thus, cannot be effectively accumulated by the anionic exchanger, Tosflex, of the TMFE during the preconcentration step. In contrast, anionic bismuth (III) complexes, are formed in the chloride medium and can be effectively accumulated into the TMFE during the preconcentration step. Moreover, the bismuth stripping peak obtained in the chloride media occurs at a potential more negative than that obtained in the nitrate media. As shown in Fig. 2, this potential shift was similar to what was observed for the oxidation of mercury (wave b in Fig. 2) which also had been described in the literature [19].
3.2. Factors affect the performance of the TMFE The thickness of the Tosflex film directly affects the electrode performance since it controls the diffusion process and the maximum loading of the bismuth(III) chloride anions in the TMFE. The film thickness was varied by preparing the electrodes with different spin-coating rate. The bis-
3.3. Effect of solution pH The dependence of the bismuth stripping peak current on the pH was studied by maintaining the chloride concentration at 0.1 M and the results obtained show that the bismuth stripping peak current gradually increases with decreasing the pH and starting level off when the pH is 1.4. This behavior suggests that the formation of bismuth(III) chloride is more favorable in an acidic environment. As a result, solutions with pH 1.4 were used in the subsequent experiments. It is worth to note that at this pH value Tosflex film was still very stable.
3.4. Effect of chloride ion concentration The detection of bismuth with TMFE relies on the effective formation of the bismuth(III) chloride anions. Fig. 3 presents the effect of changing the chloride ion concentration on the bismuth stripping peak current. The optimum concentration range of chloride ion is found to be around 0.5 M. Lowering the chloride ion concentration results in lower signals, indicating that a reasonable excess of chloride ions is required for converting the bismuth (III) into its chloride complexes. The decrease of the stripping peak
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
981
current at chloride concentration higher then 0.5 M might be attributed to failures in permselectivity, which were previously observed at high ionic strengths also for Nafion coated electrode [20]. A 0.5 M concentration of chloride ion was therefore employed in the subsequent experiments.
3.5. Effect of square-wa6e parameters The square-wave parameters that were investigated were the pulse height and the frequency. These parameters together affect the peak shape and peak current of the bismuth response. The peak current of bismuth increases with squarewave pulse height up to 20 mV, and further increase in the pulse height decreases and broadens the bismuth stripping peak. Meanwhile, considerable increase in the background current is also observed when the pulse height exceeds 20 mV. Increases in the square-wave frequency up to 120 Hz results in a substantial increase in the
Fig. 4. Effect of preconcentration time on the SWSV response for 10 ppb bismuth(III) obtained on TMFE. Ed = −0.2 V s; KCl = 0.5 M; pH=1.4.
response for bismuth. Overall, the optimum pulse height and frequency selected for SWSV were 20 mV and 120 Hz, respectively.
3.6. Effect of preconcentration potential and time
Fig. 3. Effect of chloride ion concentration on the SWSV response for 10 ppb bismuth(III) obtained on TMFE. Ed = − 0.2 V; td = 120 s; pH= 1.4.
The dependence of the bismuth stripping peak current on the preconcentration potential is studied and the results show that the peak current increases as the preconcentration potential becomes more negative between− 0.15 and − 0.4 V and starts to level off as the preconcentration potential becomes more negative than− 0.4V. Because a more negative preconcentration potential would increase the possibility of co-deposition of interfering species, a preconcentration potential of− 0.2 V was chosen in the subsequent work. The effect of preconcentration time on the SWSV response of bismuth was studied, and the results are displayed in Fig. 4. The peak current increases with increasing preconcentration time up to 6 min. This phenomenon indicates that the
982
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
ion-exchange process between Tosflex and bismuth(III) chloride anions is very good. Although this figure shows that higher sensitivity for the detection of lower bismuth concentration can be achieved by increasing the preconcentration time, a preconcentration time of 2 min was chosen for the construction of the calibration curve in this study.
3.7. Calibration A calibration graph was constructed from data taken for bismuth(III) solutions following 2 min preconcentration under optimum experimental conditions described above. Fresh sample solutions were used for each individual bismuth(III) concentration and at least three independent determinations were made for each data point. The calibration plot thus obtained showed a linear behavior between 0 and 50 ppb of bismuth(III) with slope (mA/ppb), and correlation coefficient of 2.99, and 0.995, respectively. The detection limit (S/N = 3) is 0.58 ppb. An even lower detection limit could be achieved for bismuth(III) provided that the preconcentration time is longer than 2 min.
Table 1 Influence of other ions on the response of Bi(III) at TMFEa Ions
Pb(II) Cd(II) Te(IV) Zn(II) Ge(IV) Cu(II) Sb(III) Cr(VI)
Concentration excess
Contribution (%)
Over Bi(III)
(iBi(III) =100%)
1000× 1000× 1000× 1000× 1000× 500× 100× 100×
−2 −3 −6 +3 +7 −17 +16 +3
a
(Bi(III))=10 ppb; Ep =−0.2 V; tp =120 s; pH, 1.4; KCl = 0.5 M.
the organic interferences from reaching the interface at which the deposition and stripping of bismuth takes place. In this study, the nonionic surfactant Triton X-100 was used to exemplify the effect of a typical surfactant. As shown in Fig. 5, for 10 ppb bismuth(III), the detection was found to tolerate the presence of Triton X-100 for at least up to 10 ppm with the TMFE. Compared to the same experiments performed with a bare MFE
3.8. Interferences The influences of various metal ions on the determination of bismuth were examined. The results obtained for 10 ppb bismuth with 2 min preconcentration time are summarized in Table 1. This table shows that over 1000-fold wt. excess concentrations of zinc(II), germanium(IV), tellurium(IV), lead(II), cadmium(II) and 500-fold excess of copper(II) and 100-fold excess of chromium(VI), antimony(III) only slightly interfere with the bismuth response. The only major interfering ion among the ions studied was selenium(IV). It is well documented that surface-active compounds can often adsorb on the electrode surface and reduce the analytical response of the analyte in a stripping analysis using a bare mercury electrode. Coating the electrode with Tosflex can circumvent such interferences. The Tosflex membrane coated on the electrode surface can prevent
Fig. 5. Effect of the surfactant Triton X-100 at different concentrations on the stripping response for 10 ppb bismuth(III) with the TMFE (solid line) and the bare MFE (dashed line). KCl =0.5 M, pH =1.4, td =120 s, Ed = − 0.2 V. Normalised peak currents are calculated as the ratio between signals recorded in the presence and in the absence of surfactant.
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
983
Table 2 Determination of Bi(III) in sea water, tap water, ground water, and urine samplesa
Detected value original Spiked (ppb) Detected value after spike (ppb) Recovery (%) a b
Sea water
Tap water
Ground water
Urineb
ND 4 3.889 0.04 97
ND 1 1.02 9 0.02 102
ND 2 1.99 9 0.03 99
ND 5 5.04 9 0.12 101
Number of samples assayed = 3. Deposition potential = −0.6 V.
(Fig. 5), the TMFE shows a much better resistance towards surfactant interference.
3.9. Determination of bismuth in real water samples The analytical utility of the TMFE for the determination of bismuth(III) was assessed by applying it to the determination of bismuth(III) in sea water, tap water, groundwater, and urine samples. No bismuth(III) was detected in all the four original water samples so they were spiked with appropriate amounts of bismuth(III). The results collected in Table 2 are those for the original and spiked water samples. As can be seen, the recovery of the spiked bismuth(III) is very good for all the four water samples, indicating that the proposed procedure is feasible for the determination of bismuth(III) in various water samples. Note that the amount of bismuth(III) in natural water is typically very low, and this is indeed the case in this study. The amount of bismuth(III) in the original urine samples assayed in this study can not be detected by the proposed procedure with 2 min preconcentration at a potential of − 0.2V. Nevertheless, it was found that the spiked 5 ppb bismuth(III) stripping peak could be actually seen for these samples with a precontration potential of−0.6 V, and the amount of bismuth(III) in these original water samples was therefore believed to be well below 5 ppb.
4. Conclusion The application of the TMFE for the determination of trace bismuth(III) is evident from the
above results. The presence of Tosflex polymeric coating improves the efficiency and selectivity of the preconcentration step, as well as the mechanical stability of the mercury film. Compared to the PVP–MFE procedure reported earlier [11], the TMFE offers some advantages in addition to better resistance to the interferences from common ions and organic surface-active compound. Because of the intrinsic stability of Tosflex, the preparation of the TMFE does not require any cross-liking agent nor the heating process, and thus, the time for preparing the working electrode is greatly reduced. Consequently, the TMFE would be an ideal substitute for the PVP–MFE.
Acknowledgements The authors gratefully acknowledge the financial support of the National Science Council of the Republic of China (Taiwan) under Grants NSC87-2815-C-006-073M.
References [1] T.M. Florence, J. Electroanal. Chem. 35 (1972) 237. [2] T.R. Gilbert, D.N. Hume, Anal. Chim. Acta 65 (1973) 451. [3] T.M. Florence, J. Electroanal. Chem. 49 (1974) 255. [4] S. Komorsky-Lovric, Anal. Chim. Acta 204 (1988) 161. [5] J. Lexa, K. Stulik, Talanta 32 (1985) 1027. [6] K. Kalcher, Fresenius Z. Anal. Chem. 325 (1986) 186. [7] K.L. Dong, L. Kryger, J.K. Christensen, K.N. Thomsen, Talanta 38 (1991) 101. [8] I. Svancara, K. Vytras, Anal. Chim. Acta 273 (1993) 195. [9] C. Wang, Q. Sun, H. Li, Electroanalysis 9 (1997) 645. [10] D.W.M. Arrigan, Analyst 119 (1994) 1953. [11] J.M. Zen, M.J. Chung, Anal. Chim. Acta 320 (1996) 43.
984
H.-Y. Yang et al. / Talanta 50 (1999) 977–984
[12] B. Lindholm, M. Sharp, J. Electroanal. Chem. 198 (1986) 37. [13] J.M. Zen, M.J. Chung, Anal. Chem. 67 (1995) 3571. [14] L. Dunsch, L. Kavan, J. Weber, J. Electroanal. Chem. 280 (1990) 313. [15] P. Ugo, L.M. Moretto, G.A. Mazzocchin, Anal. Chim. Acta 273 (1993) 229. [16] P. Ugo, L.M. Moretto, G.A. Mazzocchin, Anal. Chim. Acta 305 (1995) 74.
.
[17] T.-H. Lu, I.-W. Sun, Electroanalysis 10 (1998) 1052. [18] H.-Y. Yang, I.-W. Sun, Electroanalysis 11 (1999) 195. [19] L.M. Moretto, G.A. Mazzocchin, P. Ugo, J. Electroanal. Chem. 427 (1997) 113. [20] R. Nagegeli, J. Redepenning, F.C. Anson, J. Phys. Chem. 90 (1986) 6227. [21] P. Ugo, L.M. Moretto, Electroanalysis 7 (1995) 1105. [22] H.-Y. Yang, I.-W. Sun, Anal. Chim. Acta 358 (1998) 285.