- Email: [email protected]
Adsorptive cathodic stripping voltammetric determination of gold(III) in the presence of yeast mannan
Analytica Chimica Acta 385 (1999) 393±399
Adsorptive cathodic stripping voltammetric determination of gold(III) in the presence of yeast mannan Barbara Lacka, John Duncanb, Tebello Nyokonga,* a
Department of Chemistry, Rhodes University, Grahamstown 6140, South Africa Department of Biochemistry and Microbiology, Rhodes University, Grahamstown 6140, South Africa
b
Received 27 May 1998; received in revised form 14 October 1998; accepted 18 October 1998
Abstract Adsorptive cathodic stripping voltammetric (AdCSV) studies of gold(III) on a glassy carbon electrode and in the presence of yeast mannan are reported. These studies give evidence of the formation of a complex between gold(III) and mannan in acid media as judged by the enhancement in the AdCSV currents and shift in the reduction peak of gold(III) in the presence of mannan. The AdCSV currents were linearly dependent on gold(III) concentrations ranging from 7.0 10ÿ7 to 3.0 10ÿ4 mol dmÿ3. A detection limit of 6.0 10ÿ8 mol dmÿ3 was obtained. Interferences of copper(II) were observed in the presence of mannan, but there was no signi®cant interference of silver(I). # 1999 Elsevier Science B.V. All rights reserved. Keywords: Gold; Stripping voltammetry; Mannan; Copper; Yeast
1. Introduction The accurate, sensitive and selective determination of gold continues to be of interest in analytical chemistry research. Many metals can be conveniently determined by anodic stripping voltammetry (ASV). The main problem with electrochemical stripping analysis of gold(III) is to ®nd a working electrode onto which gold can be deposited for subsequent stripping. Even though gold(III) is easily reduced, deposition of elemental gold onto carbon or platinum electrodes is hampered by the slow nucleation process [1,2]. However, at relatively high concentrations *Corresponding author. Tel.: +27-46-6038260; fax: +27-46-6225109; e-mail: [email protected]
(>200 mg lÿ1), gold(III) may be determined with some reproducibility on these electrodes [1]. The electrodeposition of gold onto carbon or platinum electrodes is affected by several factors such as pH, complexing agents, electrolyte and other metal ions present in solution [3±7]. Even though both platinum and carbon electrodes have been shown to be the preferred electrode material for the determination of gold(III) ions in solution, the later suffers from high background current when low gold(III) concentrations are analyzed [8]. In order to improve on the ef®ciency of gold deposition on carbon electrodes, co-deposition with other metals and activation of the electrode with small amounts of gold have been reported [1,7]. Gold(III) is readily reduced to gold(I) at pH's greater than 2 [9]. Thus, the stripping voltammetry
0003-2670/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved. PII: S0003-2670(98)00736-3
394
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
of gold(III) is generally performed in highly acid media. The stripping solution has to provide a medium in which gold can be reversibly oxidized and the stripping peak for gold is well resolved far from probable interferents. Hydrochloric acid or a combination of hydrochloric and nitric acids are preferred for the stripping of gold [3,7]. The width and area of the stripping peak have been found to be independent of HCl concentration [1], but high concentrations of HCl are preferred when there is a possibility of the presence of silver ions in solution. Any silver chloride that may be deposited on to the working electrode is rapidly dissolved by the high concentrations of HCl. The use of biomass in metal ion recovery from industrial waste waters continues to attract considerable attention [8±13]. In particular, yeast and algae biomass can accumulate signi®cant quantities of metal ions from aqueous solutions. In general, the binding of metal ions to biomass is believed to involve complexation with ligands in or on the cell wall surface. Thus, the determination of the actual binding sites of the metal ions is of importance. One of the components of the yeast cell wall is mannan, a polysaccharide which consists of chains of mannose units (I),
linked together at a1,6 positions to form the main chain [14]. We report in this paper the adsorptive cathodic stripping voltammetry (AdCSV) of gold in the presence of mannan extract from the yeast, Saccharomyces cerevisiae. The ®rst step in AdCSV is the formation of a metal±chelate complex, followed by its controlled accumulation (deposition) onto a working electrode and its reduction during the stripping step. Cathodic stripping voltammetry frequently offers
lower sensitivity than anodic stripping voltammetry, but has better selectivity. Since the two most serious interferences likely to be present in a gold sample are copper(II) and silver(I), we studied the effects of these ions on the adsorptive cathodic stripping voltammograms of gold(III). The peak for the reduction of Au3 is about 200 mV more negative on glassy carbon electrode (GCE) than on the Pt electrode [8], resulting in interferences from copper to be more important on the GCE than on the Pt electrodes. A glassy carbon working electrode and concentrated nitric acid electrolyte are used in this work for the AdCSV determination of gold(III) in the presence of mannan. 2. Experimental 2.1. Methods Adsorptive cathodic stripping voltammograms were obtained with the BioAnalytical Systems (BAS) 100B/W Electrochemical Workstation. A 3 mm diameter glassy carbon electrode (GCE) was employed as a working electrode. A silver/silver chloride (KCl 3 mol dmÿ3) and a platinum wire served as reference and auxiliary electrodes, respectively. Prior to use, the glassy carbon electrode was washed with dilute nitric acid and then cleaned by polishing with alumina (0.05 mm) on a Buehler feltpad, followed by rinsing in water and ®nally in the electrolyte used for the electrochemical experiments. For adsorptive cathodic stripping voltammetry, a known volume of Au3 (as HAuCl4), dissolved in a 1 : 1 mixture of HCl (0.1 mol dmÿ3) and HNO3 (0.1 mol dmÿ3) was introduced into the electrochemical cell, then a known amount of the yeast mannan (in water) was added and the solution in the cell was made up to the required volume with 0.1 mol dmÿ3 HNO3. The solution was deaerated for 5 min. The Au(III)± mannan chelate is expected to have formed at this stage. A deposition potential of 1.05 V was then applied to the GCE for a known period of time to effect the adsorption of the Au(III)±mannan chelate onto the GCE. The voltammograms for the AdCSV of Au3 were recorded from the deposition potential of 1.05 V to ÿ0.3 V versus Ag/AgCl at a scan rate of 100 mV sÿ1 during the stripping step. For the experi-
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
395
ments in which Cu2 or Ag were added to Au3 solutions, voltammograms were recorded from the deposition potential to ÿ0.6 V versus Ag/AgCl. In order to ensure that all of the Au(III)±mannan chelate had been removed from the glassy carbon electrode before the next AdCSV scan, it was necessary to rinse the electrode in dilute nitric acid followed by scrubbing on a Buechler felt-pad containing alumina, and washing with distilled deionized water, in between scans. Good reproducibility (to within 5%) of the voltammetric responses was obtained for each solution. For experiments involving interferences with silver and copper ions, known amounts of Ag and Cu2 were added to the Au3 solutions and adsorptive cathodic stripping voltammograms were then recorded. 2.2. Reagents The gold(III) stock solution was prepared from tetrachloroauric(III) acid (HAuCl43H2O, Aldrich), in an acid mixture (1 : 1) of HCl (0.1 mol dmÿ3) and HNO3 (0.1 mol dmÿ3). Silver(I) and copper(II) solutions were prepared from AgNO3 (Holpro, South Africa) and CuNO3 (UNILAB, South Africa), respectively. Yeast mannan from Saccharomyces cerevisiae was obtained from Sigma. Stock solutions of mannan were prepared in water. Distilled deionized water was used for all electrochemical experiments. All glassware was soaked for at least 24 h in nitric acid, followed by washing in water before use. 3. Results and discussion 3.1. Adsorptive cathodic stripping voltammetry of gold(III) Even though differential pulse stripping voltammetry (DPSV) is widely used for the determination of metals, we have chosen linear sweep stripping voltammetry for our studies of the interaction of Au(III) with mannan. Our attempts to perform stripping studies using DPSV resulted in signals that were more complex than observed for linear sweep, as has been observed before [8]. Fig. 1 shows the AdCSV of 1.0 10ÿ4 mol dmÿ3 of Au3 in the absence (Fig. 1(a)) and presence
Fig. 1. The adsorptive cathodic stripping voltammograms obtained for Au3 (1.0 10ÿ4 mol dmÿ3) on GCE in the absence (a) and presence (b) of 0.30 mg lÿ1 of yeast mannan. The AdCSV of 0.11 mg lÿ1 yeast mannan in the absence of metal is shown in (c). Scan rate 100 mV sÿ1.
(Fig. 1(b)) of yeast mannan (0.3 mg lÿ1). Addition of mannan resulted in the enhancement of the stripping peak of Au3, indicative of the in situ formation of the Au±mannan chelate. The voltammetric response of surface con®ned species is directly related to their surface concentration and at low adsorbate concentrations, the surface concentration is directly proportional to the concentration in the bulk solution [15]. The adsorptive cathodic stripping voltammogram of yeast mannan alone, in the absence of Au3, was observed at ÿ0.55 V versus Ag/AgCl (Fig. 1(c)). The adsorptive cathodic stripping peak
396
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
of Au3 in the absence of mannan was observed at 0.29 V versus Ag/AgCl (Fig. 1(a)). In dilute aqua regia electrolyte, a single peak assigned to the reduction of Au3 to Au was observed before at 0.400 V versus Ag/AgCl, on GCE [8]. The reduction of Au3 has been observed at 0.45 V versus Ag/AgCl on GCE in the presence of acetate buffer and NaCl [16]. On modi®ed GCE, the reduction peak for Au3 was observed at potentials as low as 0.15 V [16] and on unmodi®ed carbon paste electrodes, the reduction of Au3 to Au occurred at 0.32 V versus saturated calomel electrode [3]. These observations show that the reduction of Au3 is highly dependent on the nature of the electrode, electrolyte and complexing ions. On an unmodi®ed GCE, only a single reduction peak associated with the three-electron reduction of Au3 to Au, has been observed [16]. Fig. 1(a) shows that two reduction peaks were observed for Au3 in the absence of mannan under our experimental conditions, suggesting a step-wise reduction. The second reduction peak was observed near ÿ0.09 V and was much weaker than the peak observed at 0.29 V. There have been reports of the observation of two reduction peaks for the reduction of Au3 on carbon electrodes [17]. We suggest that a two step reduction of Au3 via Au occurs under our conditions, leading to two reduction peaks. As discussed in Section 1, gold(III) is readily reduced to gold(I) at pH greater than 2. Thus, the adsorptive cathodic stripping voltammetry of gold(I), presumed to be formed when the pH of gold(III) solutions was adjusted to values greater than 2, gave a reduction peak due to gold(I) at ÿ0.02 V, a potential close to that associated with the reduction of gold(I) in Fig. 1(a). The peak height for the second reduction at ÿ0.09 V (Fig. 1(a)) is smaller than that for the ®rst reduction implying that a fewer number of electrons are involved in the former. We suggest that the more positive peak at 0.29 V is due to the reduction of the Au3 and the peak at ÿ0.09 V is due to the reduction of the Au. In the presence of mannan the voltammetric peak observed at 0.29 V shifted to 0.33 V. The magnitude of the shift in the metal reduction peak on addition of a ligand is indicative of the stability of the metal±ligand complex. A positive shift of 0.04 V for the Au3 reduction peak on addition of mannan, shows that the resulting Au±mannan chelate is more easily reduced than the Au3 ions. The shift of 0.04 V is
relatively small showing that the complex formed between Au3 and mannan is not very stable. The peak observed near ÿ0.09 V, which we associate with the Au reduction, was also enhanced on addition of mannan to Au3 solutions. In the presence of mannan, the Au reduction peak shifted to a more negative potential (ÿ0.15 V versus Ag/AgCl), showing that the resulting complex is less readily reduced than the Au species. This potential represents a shift of ÿ0.06 V from the reduction potential of the Au species, which is slightly larger than the shift observed for the Au(III)±mannan complex, suggesting that Au forms slightly more stable complexes with mannan than Au3. To establish the optimum conditions for the detection of gold, we studied the different parameters that affect the formation of the metal±mannan chelate, its adsorption onto the electrode and its stripping from the electrode. Fig. 2 shows the variation of the AdCSV currents for the peak assigned to the reduction of Au3 in the presence of mannan, as a function of the time used for the deposition (accumulation) of the Au(III)± mannan chelate onto the electrode. A rapid increase in the cathodic stripping current with deposition time was observed. As is expected for adsorption processes, the dependence of the peak current on the deposition time is limited by the saturation of the electrode, resulting in a peak of the currents for the plot of the deposition time versus current (Fig. 2). The maxi-
Fig. 2. The dependence of AdCSV peak currents on the deposition time for Au3 (1 10ÿ4 mol dmÿ3) in the presence of 0.30 mg lÿ1 of mannan. Scan rate 100 mV sÿ1. Deposition potential 1.05 V vs. Ag/AgCl.
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
Fig. 3. The dependence of AdCSV peak currents on the deposition potential for Au3 (1 10ÿ4 mol dmÿ3) in the presence of 0.30 mg lÿ1 of mannan. Scan rate 100 mV sÿ1. Deposition time 250 s.
mum currents occurred at an deposition time of about 40 s, hence this time was used for all AdCSV studies. The potential required for the deposition of the Au(III)±mannan chelate onto the electrode, the deposition potential, varied with the adsorptive cathodic stripping voltammetry currents as shown in Fig. 3. An optimum deposition potential of 1.05 V was obtained and used for further AdCSV studies. The AdCSV currents of the peak assigned to Au3 varied linearly with the concentration of Au3, for concentrations ranging from 7.0 10ÿ7 to 3.0 10ÿ4 mol dmÿ3, in the presence of a constant concentration of mannan (0.04 mg lÿ1). A plot of the variation of AdCSV currents in the Au3 concentration ranging from 5.0 10ÿ5 to 3.0 10ÿ4 mol dmÿ3 is linear with a regression equation of y 1.7x 5.6 (y peak height and x concentration) and a correlation coef®cient of 0.995 was obtained for the linear plot. The relative standard deviation for 1.0 10ÿ4 mol dmÿ3 and a deposition time of 40 s was 3%. The detection limit is 6.0 10ÿ8 mol dmÿ3 for the determination of Au(III) under our experimental conditions. The AdCSV currents increased with increase in the concentration of the mannan at low concentrations, as shown in Fig. 4. At high mannan concentrations, the current response decreased as the concentration of the mannan increased, implying that the activity of the electrode decreases at high concentrations of the ligand due to the full coverage of the electrode, since the ligand, mannan, may occupy free sites on the
397
Fig. 4. The influence of yeast mannan concentration on the adsorptive cathodic stripping currents of 1 10ÿ4 mol dmÿ3 Au3. Scan rate 100 mV sÿ1.
electrode, competing with the metal±mannan complex. Mannan consists of units of mannose, and the most likely site for coordination of the metal is at the OH groups. In a manner similar to the interactions of metals to catechols [18], we propose that the following scheme applies for the interaction of mannan with, for example, Au3. Au3 Manaq ! Au Manaq 3H
(1)
Au Manaq ! Au Manads
(2)
Au Manads 3H 3eÿ ! Au Man
(3)
Where Man represents mannan. 3.2. Interferences of copper and silver ions Interferences for the adsorptive cathodic stripping voltammetric determination of Au3 could either be due to peak overlap or the formation of intermetallic compounds. In the presence of complexing ligands such as mannan, interferences could also be due to the competition of the metal ions for the ligands. Many metal ions have only a slight effect on the anodic stripping voltammetry determination of gold(III) even when they are present in a large excess (120-fold) with respect to gold [3]. However, silver is known to interfere signi®cantly with the gold determination, a 10-fold excess of silver(I) can seriously obscure the stripping voltammogram of gold [1]. Even though the stripping peak for copper is
398
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
Fig. 5. The adsorptive cathodic stripping voltammograms of Au3 (6.0 10ÿ5 mol dmÿ3) and Cu2 (1.2 10ÿ4 mol dmÿ3) in the presence (a) and absence (b) of (0.24 mg lÿ) of mannan. Scan rate 100 mV sÿ1.
several hundred millivolts more cathodic to the gold peak, gold and copper are known to form intermetallic compounds when the copper(II) concentration is much higher (>100-fold) than that of gold(III) [1]. Fig. 5(a) shows the AdCSV of Au3 (6.0 10ÿ5 mol dmÿ3) in the presence of Cu2 (1.2 10ÿ4 mol dmÿ3) and mannan (0.24 mg ÿ1). The Au(III)±mannan peak was observed at 0.33 V when only Au3 and mannan were present in solution (Fig. 1(b)). On addition of Cu2 the potential of the peak did not change signi®cantly (Fig. 5(a)). There was however, a large increase in background current of the adsorptive cathodic stripping voltammograms in the presence of Cu2, hence the peak associated with Au was not observed in Fig. 5. The adsorptive cathodic stripping voltammograms of a solution containing Au3 in the presence of Cu2 and mannan, showed a new reduction peak at ÿ0.36 V (Fig. 5(a)). A weak reduction peak was observed for Cu2 alone in the presence of mannan near ÿ0.27 V. The peak for the reduction of Cu2 in the absence of mannan but in the presence of Au3 was observed at ÿ0.30 V. These observation suggest that the peak at ÿ0.36 V in Fig. 5(a) in the adsorptive cathodic stripping voltammetry of a solution containing Au3, Cu2 and mannan is due to the complex between copper(II) and
mannan. This peak becomes sharper and more enhanced in the presence of Au3. The currents for this peak and its sharpness increased with increase in Au3 concentration. Only a small increase in peak current was observed for the peak at ÿ0.36 V on increasing the concentration of Cu2. In fact, addition of large concentrations of Cu2 (>10ÿ3 mol dmÿ3) resulted only in the broadening of this peak and not in the increase in current. The observation that the Cu(II)±mannan complex, with a reduction potential at ÿ0.36 V, was sensitive to the addition of Au3 ions, suggests the formation of intermetallic compounds involving copper and gold. However, the currents for the reduction of the Cu(II) complex were much smaller than for the Au(III) complex considering similar concentrations of the metal ions and mannan ligand. The adsorptive cathodic stripping voltammogram of a solution containing Au3 and Cu2 in the absence of mannan is shown in Fig. 5(b). On addition of mannan (Fig. 5(a)) a much larger enhancement is observed for the peak associated with the reduction of the gold(III) complex than for the copper(II) complex, showing that Au3 is more sensitive to the presence of mannan than copper. In the presence of both Au3 and Cu2, the shift of 0.04 V in the potential of the peak due to the reduction of Au3 was smaller than the shift of ÿ0.06 V observed for the reduction of Cu2 on addition of mannan. These observations suggest that Cu2 may form a slightly more stable complex with mannan than Au3. Ag is known to cause serious interferences for the determination of Au3 using stripping voltammetry [8]. Au3 also causes interferences for the determination of Ag [19]. It has been shown before that cathodic stripping voltammogram peaks of Au3 become sharper and the peak potential shifts anodically in the presence of Ag [16]. Ag increases the sensitivity toward gold determination on modi®ed glassy carbon electrodes [16]. However, we observed no signi®cant changes in the adsorptive cathodic stripping voltammetry peak of Au3 ÿ4 ÿ3 (1.0 10 mol dm ) in the presence of mannan on addition of concentrations of Ag ranging from 1.0 10ÿ4 to 1.0 10ÿ3 mol dmÿ3. Also we did not observe any new peaks which could be associated with the Ag(I)±mannan complex, suggesting that Ag does not compete with Au3 for the coordination of mannan under the present experimental conditions.
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
399
4. Conclusions
References
We have shown in this work that yeast mannan has an af®nity for Au3 and Cu2, thus Cu2 may compete with Au3 for the coordination to mannan resulting in signi®cant interferences for the accumulation of gold by yeast mannan. However, yeast mannan was found to be more sensitive to Au3 than to Cu2. The detection limit observed in this work compares with the limits reported in the literature for stripping voltammetry of Au3 using linear sweep voltammetry. Even though high sensitivity is an important requirement from the analytical point of view, it must be accompanied by high selectivity. The observation that Ag does not compete with Au3 for mannan is signi®cant since Ag is known to cause serious interferences for the determination of Au3. The development of ligands that are speci®c for Au(III) is important in the analysis and recovery of gold. The study of the interaction of Au(III) with mannan is an important step towards understanding the accumulation of gold by biomass.
[1] H. Huilang, D. Jagner, L. Renman, Anal. Chim. Acta 208 (1988) 301. [2] E.S. Jacobs, Anal. Chem. 35 (1963) 2112. [3] X. Cai, K. Kalcher, C. Neuhold, W. Diewald, R.J. Magee, Analyst 118 (1993) 53. [4] G.H. Schmid, G.W. Bolger, Clin. Chem. 19 (1973) 1002. [5] P. Tuzhi, L. Huiping, W. Shuwen, Analyst 118 (1993) 1321. [6] S. Dong, X. Yang, Talanta 43 (1996) 1109. [7] M. Korolczuk, Fresenius J. Anal. Chem. 356 (1996) 480. [8] A.M. Bond, S. Kratsis, S. Mitchell, J. Macak, Analyst 122 (1997) 1147. [9] G. Trejo, A.F. Gil, I. GonzaÂlez, J. Appl. Electrochem. 26 (1996) 1287. [10] D.W. Darnall, B. Greene, M.T. Henzl, J.M. Hosea, R.A. McPherson, J. Sneddon, M.D. Alexander, Environ. Sci. Technol. 20 (1986) 206. [11] B. Greene, M.T. Henzl, J.M. Hosea, D.W. Darnall, Biotech. Bioeng. 28 (1986) 764. [12] D. Brady, A.D. Stoll, J.R. Duncan, Biotech. Bioeng. 44 (1994) 297. [13] B.S. Wilhelmi, Ph.D. Thesis, Rhodes University, 1997. [14] W.N. Haworth, E.L. Hirst, F.A. Isherwood, J. Chem. Soc. (1937) 784. [15] J. Wang, in: A.J. Bard (Ed.), Electroanalytical Chemistry, vol. 16, Wiley, New York, 1989, p. 1. [16] I. Turgan, D. Mandler, Fresenius J. Anal. Chem. 349 (1994) 491. [17] L.N. Vasileva, T.A. Koroleva, Anal. Chem. (USSR) 28 (1973) 1875. [18] L. Limson, T. Nyokong, Anal. Chim. Acta 344 (1997) 87. [19] I. SÏvancara, K. Kalcher, W. Diewald, K. VytrÏas, Electroanalysis 8 (1996) 336.
Acknowledgements This work was supported by Rhodes University and the Foundation for Research Development (South Africa).
Adsorptive cathodic stripping voltammetric determination of gold(III) in the presence of yeast mannan Barbara Lacka, John Duncanb, Tebello Nyokonga,* a
Department of Chemistry, Rhodes University, Grahamstown 6140, South Africa Department of Biochemistry and Microbiology, Rhodes University, Grahamstown 6140, South Africa
b
Received 27 May 1998; received in revised form 14 October 1998; accepted 18 October 1998
Abstract Adsorptive cathodic stripping voltammetric (AdCSV) studies of gold(III) on a glassy carbon electrode and in the presence of yeast mannan are reported. These studies give evidence of the formation of a complex between gold(III) and mannan in acid media as judged by the enhancement in the AdCSV currents and shift in the reduction peak of gold(III) in the presence of mannan. The AdCSV currents were linearly dependent on gold(III) concentrations ranging from 7.0 10ÿ7 to 3.0 10ÿ4 mol dmÿ3. A detection limit of 6.0 10ÿ8 mol dmÿ3 was obtained. Interferences of copper(II) were observed in the presence of mannan, but there was no signi®cant interference of silver(I). # 1999 Elsevier Science B.V. All rights reserved. Keywords: Gold; Stripping voltammetry; Mannan; Copper; Yeast
1. Introduction The accurate, sensitive and selective determination of gold continues to be of interest in analytical chemistry research. Many metals can be conveniently determined by anodic stripping voltammetry (ASV). The main problem with electrochemical stripping analysis of gold(III) is to ®nd a working electrode onto which gold can be deposited for subsequent stripping. Even though gold(III) is easily reduced, deposition of elemental gold onto carbon or platinum electrodes is hampered by the slow nucleation process [1,2]. However, at relatively high concentrations *Corresponding author. Tel.: +27-46-6038260; fax: +27-46-6225109; e-mail: [email protected]
(>200 mg lÿ1), gold(III) may be determined with some reproducibility on these electrodes [1]. The electrodeposition of gold onto carbon or platinum electrodes is affected by several factors such as pH, complexing agents, electrolyte and other metal ions present in solution [3±7]. Even though both platinum and carbon electrodes have been shown to be the preferred electrode material for the determination of gold(III) ions in solution, the later suffers from high background current when low gold(III) concentrations are analyzed [8]. In order to improve on the ef®ciency of gold deposition on carbon electrodes, co-deposition with other metals and activation of the electrode with small amounts of gold have been reported [1,7]. Gold(III) is readily reduced to gold(I) at pH's greater than 2 [9]. Thus, the stripping voltammetry
0003-2670/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved. PII: S0003-2670(98)00736-3
394
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
of gold(III) is generally performed in highly acid media. The stripping solution has to provide a medium in which gold can be reversibly oxidized and the stripping peak for gold is well resolved far from probable interferents. Hydrochloric acid or a combination of hydrochloric and nitric acids are preferred for the stripping of gold [3,7]. The width and area of the stripping peak have been found to be independent of HCl concentration [1], but high concentrations of HCl are preferred when there is a possibility of the presence of silver ions in solution. Any silver chloride that may be deposited on to the working electrode is rapidly dissolved by the high concentrations of HCl. The use of biomass in metal ion recovery from industrial waste waters continues to attract considerable attention [8±13]. In particular, yeast and algae biomass can accumulate signi®cant quantities of metal ions from aqueous solutions. In general, the binding of metal ions to biomass is believed to involve complexation with ligands in or on the cell wall surface. Thus, the determination of the actual binding sites of the metal ions is of importance. One of the components of the yeast cell wall is mannan, a polysaccharide which consists of chains of mannose units (I),
linked together at a1,6 positions to form the main chain [14]. We report in this paper the adsorptive cathodic stripping voltammetry (AdCSV) of gold in the presence of mannan extract from the yeast, Saccharomyces cerevisiae. The ®rst step in AdCSV is the formation of a metal±chelate complex, followed by its controlled accumulation (deposition) onto a working electrode and its reduction during the stripping step. Cathodic stripping voltammetry frequently offers
lower sensitivity than anodic stripping voltammetry, but has better selectivity. Since the two most serious interferences likely to be present in a gold sample are copper(II) and silver(I), we studied the effects of these ions on the adsorptive cathodic stripping voltammograms of gold(III). The peak for the reduction of Au3 is about 200 mV more negative on glassy carbon electrode (GCE) than on the Pt electrode [8], resulting in interferences from copper to be more important on the GCE than on the Pt electrodes. A glassy carbon working electrode and concentrated nitric acid electrolyte are used in this work for the AdCSV determination of gold(III) in the presence of mannan. 2. Experimental 2.1. Methods Adsorptive cathodic stripping voltammograms were obtained with the BioAnalytical Systems (BAS) 100B/W Electrochemical Workstation. A 3 mm diameter glassy carbon electrode (GCE) was employed as a working electrode. A silver/silver chloride (KCl 3 mol dmÿ3) and a platinum wire served as reference and auxiliary electrodes, respectively. Prior to use, the glassy carbon electrode was washed with dilute nitric acid and then cleaned by polishing with alumina (0.05 mm) on a Buehler feltpad, followed by rinsing in water and ®nally in the electrolyte used for the electrochemical experiments. For adsorptive cathodic stripping voltammetry, a known volume of Au3 (as HAuCl4), dissolved in a 1 : 1 mixture of HCl (0.1 mol dmÿ3) and HNO3 (0.1 mol dmÿ3) was introduced into the electrochemical cell, then a known amount of the yeast mannan (in water) was added and the solution in the cell was made up to the required volume with 0.1 mol dmÿ3 HNO3. The solution was deaerated for 5 min. The Au(III)± mannan chelate is expected to have formed at this stage. A deposition potential of 1.05 V was then applied to the GCE for a known period of time to effect the adsorption of the Au(III)±mannan chelate onto the GCE. The voltammograms for the AdCSV of Au3 were recorded from the deposition potential of 1.05 V to ÿ0.3 V versus Ag/AgCl at a scan rate of 100 mV sÿ1 during the stripping step. For the experi-
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
395
ments in which Cu2 or Ag were added to Au3 solutions, voltammograms were recorded from the deposition potential to ÿ0.6 V versus Ag/AgCl. In order to ensure that all of the Au(III)±mannan chelate had been removed from the glassy carbon electrode before the next AdCSV scan, it was necessary to rinse the electrode in dilute nitric acid followed by scrubbing on a Buechler felt-pad containing alumina, and washing with distilled deionized water, in between scans. Good reproducibility (to within 5%) of the voltammetric responses was obtained for each solution. For experiments involving interferences with silver and copper ions, known amounts of Ag and Cu2 were added to the Au3 solutions and adsorptive cathodic stripping voltammograms were then recorded. 2.2. Reagents The gold(III) stock solution was prepared from tetrachloroauric(III) acid (HAuCl43H2O, Aldrich), in an acid mixture (1 : 1) of HCl (0.1 mol dmÿ3) and HNO3 (0.1 mol dmÿ3). Silver(I) and copper(II) solutions were prepared from AgNO3 (Holpro, South Africa) and CuNO3 (UNILAB, South Africa), respectively. Yeast mannan from Saccharomyces cerevisiae was obtained from Sigma. Stock solutions of mannan were prepared in water. Distilled deionized water was used for all electrochemical experiments. All glassware was soaked for at least 24 h in nitric acid, followed by washing in water before use. 3. Results and discussion 3.1. Adsorptive cathodic stripping voltammetry of gold(III) Even though differential pulse stripping voltammetry (DPSV) is widely used for the determination of metals, we have chosen linear sweep stripping voltammetry for our studies of the interaction of Au(III) with mannan. Our attempts to perform stripping studies using DPSV resulted in signals that were more complex than observed for linear sweep, as has been observed before [8]. Fig. 1 shows the AdCSV of 1.0 10ÿ4 mol dmÿ3 of Au3 in the absence (Fig. 1(a)) and presence
Fig. 1. The adsorptive cathodic stripping voltammograms obtained for Au3 (1.0 10ÿ4 mol dmÿ3) on GCE in the absence (a) and presence (b) of 0.30 mg lÿ1 of yeast mannan. The AdCSV of 0.11 mg lÿ1 yeast mannan in the absence of metal is shown in (c). Scan rate 100 mV sÿ1.
(Fig. 1(b)) of yeast mannan (0.3 mg lÿ1). Addition of mannan resulted in the enhancement of the stripping peak of Au3, indicative of the in situ formation of the Au±mannan chelate. The voltammetric response of surface con®ned species is directly related to their surface concentration and at low adsorbate concentrations, the surface concentration is directly proportional to the concentration in the bulk solution [15]. The adsorptive cathodic stripping voltammogram of yeast mannan alone, in the absence of Au3, was observed at ÿ0.55 V versus Ag/AgCl (Fig. 1(c)). The adsorptive cathodic stripping peak
396
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
of Au3 in the absence of mannan was observed at 0.29 V versus Ag/AgCl (Fig. 1(a)). In dilute aqua regia electrolyte, a single peak assigned to the reduction of Au3 to Au was observed before at 0.400 V versus Ag/AgCl, on GCE [8]. The reduction of Au3 has been observed at 0.45 V versus Ag/AgCl on GCE in the presence of acetate buffer and NaCl [16]. On modi®ed GCE, the reduction peak for Au3 was observed at potentials as low as 0.15 V [16] and on unmodi®ed carbon paste electrodes, the reduction of Au3 to Au occurred at 0.32 V versus saturated calomel electrode [3]. These observations show that the reduction of Au3 is highly dependent on the nature of the electrode, electrolyte and complexing ions. On an unmodi®ed GCE, only a single reduction peak associated with the three-electron reduction of Au3 to Au, has been observed [16]. Fig. 1(a) shows that two reduction peaks were observed for Au3 in the absence of mannan under our experimental conditions, suggesting a step-wise reduction. The second reduction peak was observed near ÿ0.09 V and was much weaker than the peak observed at 0.29 V. There have been reports of the observation of two reduction peaks for the reduction of Au3 on carbon electrodes [17]. We suggest that a two step reduction of Au3 via Au occurs under our conditions, leading to two reduction peaks. As discussed in Section 1, gold(III) is readily reduced to gold(I) at pH greater than 2. Thus, the adsorptive cathodic stripping voltammetry of gold(I), presumed to be formed when the pH of gold(III) solutions was adjusted to values greater than 2, gave a reduction peak due to gold(I) at ÿ0.02 V, a potential close to that associated with the reduction of gold(I) in Fig. 1(a). The peak height for the second reduction at ÿ0.09 V (Fig. 1(a)) is smaller than that for the ®rst reduction implying that a fewer number of electrons are involved in the former. We suggest that the more positive peak at 0.29 V is due to the reduction of the Au3 and the peak at ÿ0.09 V is due to the reduction of the Au. In the presence of mannan the voltammetric peak observed at 0.29 V shifted to 0.33 V. The magnitude of the shift in the metal reduction peak on addition of a ligand is indicative of the stability of the metal±ligand complex. A positive shift of 0.04 V for the Au3 reduction peak on addition of mannan, shows that the resulting Au±mannan chelate is more easily reduced than the Au3 ions. The shift of 0.04 V is
relatively small showing that the complex formed between Au3 and mannan is not very stable. The peak observed near ÿ0.09 V, which we associate with the Au reduction, was also enhanced on addition of mannan to Au3 solutions. In the presence of mannan, the Au reduction peak shifted to a more negative potential (ÿ0.15 V versus Ag/AgCl), showing that the resulting complex is less readily reduced than the Au species. This potential represents a shift of ÿ0.06 V from the reduction potential of the Au species, which is slightly larger than the shift observed for the Au(III)±mannan complex, suggesting that Au forms slightly more stable complexes with mannan than Au3. To establish the optimum conditions for the detection of gold, we studied the different parameters that affect the formation of the metal±mannan chelate, its adsorption onto the electrode and its stripping from the electrode. Fig. 2 shows the variation of the AdCSV currents for the peak assigned to the reduction of Au3 in the presence of mannan, as a function of the time used for the deposition (accumulation) of the Au(III)± mannan chelate onto the electrode. A rapid increase in the cathodic stripping current with deposition time was observed. As is expected for adsorption processes, the dependence of the peak current on the deposition time is limited by the saturation of the electrode, resulting in a peak of the currents for the plot of the deposition time versus current (Fig. 2). The maxi-
Fig. 2. The dependence of AdCSV peak currents on the deposition time for Au3 (1 10ÿ4 mol dmÿ3) in the presence of 0.30 mg lÿ1 of mannan. Scan rate 100 mV sÿ1. Deposition potential 1.05 V vs. Ag/AgCl.
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
Fig. 3. The dependence of AdCSV peak currents on the deposition potential for Au3 (1 10ÿ4 mol dmÿ3) in the presence of 0.30 mg lÿ1 of mannan. Scan rate 100 mV sÿ1. Deposition time 250 s.
mum currents occurred at an deposition time of about 40 s, hence this time was used for all AdCSV studies. The potential required for the deposition of the Au(III)±mannan chelate onto the electrode, the deposition potential, varied with the adsorptive cathodic stripping voltammetry currents as shown in Fig. 3. An optimum deposition potential of 1.05 V was obtained and used for further AdCSV studies. The AdCSV currents of the peak assigned to Au3 varied linearly with the concentration of Au3, for concentrations ranging from 7.0 10ÿ7 to 3.0 10ÿ4 mol dmÿ3, in the presence of a constant concentration of mannan (0.04 mg lÿ1). A plot of the variation of AdCSV currents in the Au3 concentration ranging from 5.0 10ÿ5 to 3.0 10ÿ4 mol dmÿ3 is linear with a regression equation of y 1.7x 5.6 (y peak height and x concentration) and a correlation coef®cient of 0.995 was obtained for the linear plot. The relative standard deviation for 1.0 10ÿ4 mol dmÿ3 and a deposition time of 40 s was 3%. The detection limit is 6.0 10ÿ8 mol dmÿ3 for the determination of Au(III) under our experimental conditions. The AdCSV currents increased with increase in the concentration of the mannan at low concentrations, as shown in Fig. 4. At high mannan concentrations, the current response decreased as the concentration of the mannan increased, implying that the activity of the electrode decreases at high concentrations of the ligand due to the full coverage of the electrode, since the ligand, mannan, may occupy free sites on the
397
Fig. 4. The influence of yeast mannan concentration on the adsorptive cathodic stripping currents of 1 10ÿ4 mol dmÿ3 Au3. Scan rate 100 mV sÿ1.
electrode, competing with the metal±mannan complex. Mannan consists of units of mannose, and the most likely site for coordination of the metal is at the OH groups. In a manner similar to the interactions of metals to catechols [18], we propose that the following scheme applies for the interaction of mannan with, for example, Au3. Au3 Manaq ! Au Manaq 3H
(1)
Au Manaq ! Au Manads
(2)
Au Manads 3H 3eÿ ! Au Man
(3)
Where Man represents mannan. 3.2. Interferences of copper and silver ions Interferences for the adsorptive cathodic stripping voltammetric determination of Au3 could either be due to peak overlap or the formation of intermetallic compounds. In the presence of complexing ligands such as mannan, interferences could also be due to the competition of the metal ions for the ligands. Many metal ions have only a slight effect on the anodic stripping voltammetry determination of gold(III) even when they are present in a large excess (120-fold) with respect to gold [3]. However, silver is known to interfere signi®cantly with the gold determination, a 10-fold excess of silver(I) can seriously obscure the stripping voltammogram of gold [1]. Even though the stripping peak for copper is
398
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
Fig. 5. The adsorptive cathodic stripping voltammograms of Au3 (6.0 10ÿ5 mol dmÿ3) and Cu2 (1.2 10ÿ4 mol dmÿ3) in the presence (a) and absence (b) of (0.24 mg lÿ) of mannan. Scan rate 100 mV sÿ1.
several hundred millivolts more cathodic to the gold peak, gold and copper are known to form intermetallic compounds when the copper(II) concentration is much higher (>100-fold) than that of gold(III) [1]. Fig. 5(a) shows the AdCSV of Au3 (6.0 10ÿ5 mol dmÿ3) in the presence of Cu2 (1.2 10ÿ4 mol dmÿ3) and mannan (0.24 mg ÿ1). The Au(III)±mannan peak was observed at 0.33 V when only Au3 and mannan were present in solution (Fig. 1(b)). On addition of Cu2 the potential of the peak did not change signi®cantly (Fig. 5(a)). There was however, a large increase in background current of the adsorptive cathodic stripping voltammograms in the presence of Cu2, hence the peak associated with Au was not observed in Fig. 5. The adsorptive cathodic stripping voltammograms of a solution containing Au3 in the presence of Cu2 and mannan, showed a new reduction peak at ÿ0.36 V (Fig. 5(a)). A weak reduction peak was observed for Cu2 alone in the presence of mannan near ÿ0.27 V. The peak for the reduction of Cu2 in the absence of mannan but in the presence of Au3 was observed at ÿ0.30 V. These observation suggest that the peak at ÿ0.36 V in Fig. 5(a) in the adsorptive cathodic stripping voltammetry of a solution containing Au3, Cu2 and mannan is due to the complex between copper(II) and
mannan. This peak becomes sharper and more enhanced in the presence of Au3. The currents for this peak and its sharpness increased with increase in Au3 concentration. Only a small increase in peak current was observed for the peak at ÿ0.36 V on increasing the concentration of Cu2. In fact, addition of large concentrations of Cu2 (>10ÿ3 mol dmÿ3) resulted only in the broadening of this peak and not in the increase in current. The observation that the Cu(II)±mannan complex, with a reduction potential at ÿ0.36 V, was sensitive to the addition of Au3 ions, suggests the formation of intermetallic compounds involving copper and gold. However, the currents for the reduction of the Cu(II) complex were much smaller than for the Au(III) complex considering similar concentrations of the metal ions and mannan ligand. The adsorptive cathodic stripping voltammogram of a solution containing Au3 and Cu2 in the absence of mannan is shown in Fig. 5(b). On addition of mannan (Fig. 5(a)) a much larger enhancement is observed for the peak associated with the reduction of the gold(III) complex than for the copper(II) complex, showing that Au3 is more sensitive to the presence of mannan than copper. In the presence of both Au3 and Cu2, the shift of 0.04 V in the potential of the peak due to the reduction of Au3 was smaller than the shift of ÿ0.06 V observed for the reduction of Cu2 on addition of mannan. These observations suggest that Cu2 may form a slightly more stable complex with mannan than Au3. Ag is known to cause serious interferences for the determination of Au3 using stripping voltammetry [8]. Au3 also causes interferences for the determination of Ag [19]. It has been shown before that cathodic stripping voltammogram peaks of Au3 become sharper and the peak potential shifts anodically in the presence of Ag [16]. Ag increases the sensitivity toward gold determination on modi®ed glassy carbon electrodes [16]. However, we observed no signi®cant changes in the adsorptive cathodic stripping voltammetry peak of Au3 ÿ4 ÿ3 (1.0 10 mol dm ) in the presence of mannan on addition of concentrations of Ag ranging from 1.0 10ÿ4 to 1.0 10ÿ3 mol dmÿ3. Also we did not observe any new peaks which could be associated with the Ag(I)±mannan complex, suggesting that Ag does not compete with Au3 for the coordination of mannan under the present experimental conditions.
B. Lack et al. / Analytica Chimica Acta 385 (1999) 393±399
399
4. Conclusions
References
We have shown in this work that yeast mannan has an af®nity for Au3 and Cu2, thus Cu2 may compete with Au3 for the coordination to mannan resulting in signi®cant interferences for the accumulation of gold by yeast mannan. However, yeast mannan was found to be more sensitive to Au3 than to Cu2. The detection limit observed in this work compares with the limits reported in the literature for stripping voltammetry of Au3 using linear sweep voltammetry. Even though high sensitivity is an important requirement from the analytical point of view, it must be accompanied by high selectivity. The observation that Ag does not compete with Au3 for mannan is signi®cant since Ag is known to cause serious interferences for the determination of Au3. The development of ligands that are speci®c for Au(III) is important in the analysis and recovery of gold. The study of the interaction of Au(III) with mannan is an important step towards understanding the accumulation of gold by biomass.
[1] H. Huilang, D. Jagner, L. Renman, Anal. Chim. Acta 208 (1988) 301. [2] E.S. Jacobs, Anal. Chem. 35 (1963) 2112. [3] X. Cai, K. Kalcher, C. Neuhold, W. Diewald, R.J. Magee, Analyst 118 (1993) 53. [4] G.H. Schmid, G.W. Bolger, Clin. Chem. 19 (1973) 1002. [5] P. Tuzhi, L. Huiping, W. Shuwen, Analyst 118 (1993) 1321. [6] S. Dong, X. Yang, Talanta 43 (1996) 1109. [7] M. Korolczuk, Fresenius J. Anal. Chem. 356 (1996) 480. [8] A.M. Bond, S. Kratsis, S. Mitchell, J. Macak, Analyst 122 (1997) 1147. [9] G. Trejo, A.F. Gil, I. GonzaÂlez, J. Appl. Electrochem. 26 (1996) 1287. [10] D.W. Darnall, B. Greene, M.T. Henzl, J.M. Hosea, R.A. McPherson, J. Sneddon, M.D. Alexander, Environ. Sci. Technol. 20 (1986) 206. [11] B. Greene, M.T. Henzl, J.M. Hosea, D.W. Darnall, Biotech. Bioeng. 28 (1986) 764. [12] D. Brady, A.D. Stoll, J.R. Duncan, Biotech. Bioeng. 44 (1994) 297. [13] B.S. Wilhelmi, Ph.D. Thesis, Rhodes University, 1997. [14] W.N. Haworth, E.L. Hirst, F.A. Isherwood, J. Chem. Soc. (1937) 784. [15] J. Wang, in: A.J. Bard (Ed.), Electroanalytical Chemistry, vol. 16, Wiley, New York, 1989, p. 1. [16] I. Turgan, D. Mandler, Fresenius J. Anal. Chem. 349 (1994) 491. [17] L.N. Vasileva, T.A. Koroleva, Anal. Chem. (USSR) 28 (1973) 1875. [18] L. Limson, T. Nyokong, Anal. Chim. Acta 344 (1997) 87. [19] I. SÏvancara, K. Kalcher, W. Diewald, K. VytrÏas, Electroanalysis 8 (1996) 336.
Acknowledgements This work was supported by Rhodes University and the Foundation for Research Development (South Africa).