Simultaneous determination of copper, lead, cadmium and zinc using

Analytica Chimica Acta 411 (2000) 201–207

Simultaneous determination of copper, lead, cadmium and zinc using differential pulse anodic stripping voltammetry in a flow system J.F. van Staden∗ , M.C. Matoetoe Department of Chemistry, University of Pretoria, Pretoria 0002, South Africa Received 26 April 1999; received in revised form 17 January 2000; accepted 18 January 2000

Abstract The use of differential pulse anodic stripping voltammetry (DPASV) in a flow system for the simultaneous determination of Cu, Cd, Fe, Pb and Zn was evaluated for applicability with a glassy carbon electrode. Simultaneous quantitative analysis for these elements is possible at low concentrations of Fe when using pyrophosphate as a supporting electrolyte (pH 4.0). An application of the proposed method is reported for the determination of these elements in water samples. The working ranges are 50–500 ␮g l−1 for Cu, Cd and Pb and 100–400 ␮g l−1 for Zn. The Fe content was negligible in these samples. Recovery studies and metallic interferences are reported. Detection limits of 39, 9.6, 16.6 and 14.7 ␮g l−1 for Cu, Cd, Pb and Zn, respectively, were obtained. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Differential pulse anodic stripping voltammetry; Glassy carbon electrode; Flow injection analysis

1. Introduction There is an increased need to measure several analytes simultaneously and rapidly in the same sample with a single injection and/or measurement in various fields like clinical chemistry and environmental and/or industrial control. Typical simultaneous analyses have usually been based on chromatography [1,2] having detection systems based on spectrometry, anodic stripping voltammetry or potentiometric stripping. Although chromatographic techniques were successfully used, they are tedious, expensive and the separation step involved is usually time consuming and not precise, with the result that these techniques ∗ Corresponding author. Tel.: +27-12-420-2515; fax: +27-12-362-5297. E-mail address: [email protected] (J.F. van Staden)

are not convenient for routine analysis of large sets of samples where rapid analysis is needed. Alternatively multielement analysis for several analytes has successfully been done with energy dispersive X-ray fluorescence spectrometry [3], atomic spectrometry [4] and electrochemical techniques. These techniques have several advantages over chromatography as they are faster, have reduced sample treatment time and the removal of interferences is not strictly required in many instances. Recently, with the development of flow systems, the suitability and efficiency of alternative techniques have increased. Flow systems offer an opportunity to avoid contamination by working with closed systems, thus solving a major problem of simultaneous analysis. In addition it enhances sample throughout frequency as well as saving on sample and reagents when miniaturised flow systems are used. In the

0003-2670/00/$ – see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 3 - 2 6 7 0 ( 0 0 ) 0 0 7 8 5 - 6

202

J.F. van Staden, M.C. Matoetoe / Analytica Chimica Acta 411 (2000) 201–207

literature, application of flow systems to simultaneous analysis have focussed on the following [5]: • Multidetection in series or parallel involving more than one detector. • Sequential injection. • Splitting of the stream with two or more cells aligned in the same optical path [6]. • Use of two or more synchronised or special injectors for samples and/or reagents when a single detector is used [7]. Owing to its capacity to perform simultaneous multielement analysis and the ability to display the information on all elements at the same time, anodic stripping voltammetry (ASV) could offer a viable alternative for simultaneous determination of inorganic elements. The potential of ASV in trace and multielement analysis for heavy metals has been demonstrated [8–11]. The application of ASV for the determination of analytes in complex sample matrices is complicated because of the adsorption of interfering components onto the electrode surface with the result that electrode activity is lost and non-reproducible results are obtained. Furthermore difficulties associated with actual electrochemical multicomponent sample analysis, as a result of the complicated polarograms, severely decreases the application of this technique, since the information for each particular sample is rarely tabulated or described. The other problem involves the handling of the data compilation in order to design a suitable analytical method for each specific problem. The choice of a suitable combination of supporting electrolytes has increased the capacity of electrochemical techniques for multielement analysis [12–14]. Unfortunately this requires the handling of a lot of data concerning the electrochemical behaviour of metals in different media and also taking into account chemical interference between an individual element and each sample constituent, mutual interferences between the ions present and overlapping peaks or peak potential displacements as well as minimal analysis time and effort. Automation of sample handling and experimental execution allows the electrochemical response to be characterised more efficiently and thus more thoroughly. This is enhanced by employing a glassy carbon electrode instead of a hanging mercury drop electrode. Flow injection analysis (FIA) has estab-

lished itself as a well-defined automatic sampling technique operating under non-equilibrium conditions [15–17], which suited the elimination of interferences with DPASV as the detection technique. Recently an urgent need was conveyed to our laboratory for the rapid simultaneous determination of low levels of copper, lead, cadmium and zinc in water used for boiler feeds and in power stations. Among the requirements set forward were a robust low cost analyser involving no complex sample preparation, the minimum number of reagents, easy to operate, reliable and adaption for on-line monitoring. DPASV as the detection system with the main advantage of simultaneous multielements detection with FIA as the sample manipulation system seems an ideal combination to fulfil this requirement. This paper reports the simultaneous determination of Cu, Pd, Cd and Zn where a flow system is coupled with DPASV as detection technique.

2. Experimental 2.1. Chemicals and reagents All reagents were prepared from analytical reagent grade chemicals unless specified otherwise. Deionised water from a Modulab system (Continental Water Systems, San Antonio, TX) was used for dilution. The solutions were prepared as follows: cadmium, copper and iron stock solutions were prepared by dissolving the corresponding nitrates and a zinc stock solution by dissolving the sulphate in deaerated 2% (v/v) HNO3 . The working standard solutions were prepared daily by suitable dilution of this stock solution in the matrix required. High grade nitrogen was used for deaeration of samples, solvents and during analysis. Fresh stock solutions were prepared monthly. A 0.100 mol l−1 lead stock solution was obtained from Merck (Darmstadt, Germany). 2.2. Procedure The continuous flow-through trace analyser system including the ␮ autolab system (Ecochemie, Utrecht, The Netherlands) connected to a model 663 VA stand (Metrohm, Herisau, Switzerland), cell design and three electrodes arrangement used were described in


detail previously [18]. The reference electrode was an Ag/AgCl (3 mol l−1 KCl) and all potentials are quoted with respect to this. The working electrode (glassy carbon disk, 3 mm) was polished daily to a mirror finish using 0.05 ␮m alumina powder. From preliminary studies the following procedure was adopted. A supporting electrolyte mixture of 0.1 mol l−1 Na4 P2 O7 and 0.25 mol l−1 acetate buffer at pH 3.5 was prepared by separately sucking the solutions into the deaeration chamber at 4.70 and 1.12 ml min−1 , respectively. The solutions were mixed and deaerated with nitrogen and passed through the flow-though cell at 4.5 ml min−1 . Samples were introduced into the supporting electrolyte mixture at 2.73 ml min−1 just after deaeration. The solution merged, reacted and flowed through the cell. Subsequently the deposition potential (Dp , −1.25 V) was applied for 3 min. After deposition the flow was stopped and a rest period of 5 s was set before a positive differential pulse scan of 30 mV s−1 was initiated. Before starting the next run, the electrode was conditioned at 1.0 V for 1 min. After about 2 h of continuous use, the electrode was cleaned by rubbing it with a filter paper soaked in ethanol then on a dry filter paper followed by electrochemically cleaning with at least three scans. Furthermore, the electrode was polished on a daily basis with the alumina powder. 2.3. Analysis of samples and recovery experiments A 15 ml water sample, 2.0 ml of concentrated HNO3 and a suitable amount of a standard solution containing Cd, Cu, Pb and Zn were mixed in a 50 ml volumetric flask. After dilution to the mark with ultra pure water, analysis was performed as described before. Three standard additions were made for each sample.

203

For recovery experiments, a standard solution containing all elements was added in a 50 ml volumetric flask containing a water sample. The mixture was treated and analysed as described above.

3. Results and discussions 3.1. Preliminary studies The main difficulty associated with the simultaneous determination of several elements in a single voltammogram is the preparation of sample solutions. For this purpose it was necessary to find suitable electrolytes. The properties of the supporting electrolyte used in the carrier solution significantly influence the stripping process through redox reactions, formation of complex compounds and of poorly soluble compounds readily adsorbing on the electrode surface and frits used in the cells. The stripping peak potentials of the analyte elements in the different supporting electrolytes tested are shown in Table 1. An appropriate electrolyte was selected by a trial that indicates distinct peak potential (Ep ) values for the different elements studied. To enable adequate separation for simultaneous measurement an Ep difference of at least 100 mV is desirable. Resolution between Cu and Pb signals in a KSCN carrier electrolyte was insufficient owing to the vicinity of their stripping peak potentials. A similar problem was observed between Fe and Pb in acetate buffer at pH 4.0. Slowing the potential scans may solve the problem, but significantly longer analysis times will be required. Table 1 shows that (NH4 )3 C6 H5 O7 , EDTA, and Na4 P2 O7 have the sufficient Ep difference needed to resolve all five elements in their decreasing sequence. Therefore, the

Table 1 Peaks potential of the elements using different ‘electrolyte’ solutionsa Electrolyte

0.1 mol l−1 EDTA (at pH 4.0) 0.25 mol l−1 KSCN 0.1 mol l−1 Na4 P2 O7 0.25 mol l−1 acetate buffer (at pH 4.0) 0.25 mol l−1 (NH4 )3 C6 H5 O7 a

Stripping peak potential (V) Cu

Pb

Fe

Cd

Zn

0.043 −0.364 0.054 0.109 0.172

−0.461 −0.389 −0.492 −0.223 −0.412

−0.306 −0.789 −0.594 −0.296 −0.065

−0.682 −0.581 −0.804 −0.546 −0.879

−1.129 −0.829 −1.050 −0.998 −1.017

Concentration of the different ions are 4×10−5 mol l−1 .

204


Fig. 1. Voltammogram of Cu, Cd, Fe, Pb and Zn in ammonium citrate solution (pH 4.0).

best results were expected with (NH4 )3 C6 H5 O7 . Its use was made more attractive by the successful utilization of a hanging mercury electrode (HME) and the greater peak height obtained compared to the other electrolytes. The voltammogram obtained for a solution of all elements in (NH4 )3 C6 H5 O7 solution is shown in Fig. 1. A good separation was obtained as expected. The voltammograms, however, were not reproducible. Trials with various pH values and buffers were not able to correct this problem. The loss of electrode activity was not due to adsorption, because the second scan produced no peaks, indicating that all elements were stripped completely in the first scan. EDTA was used as a ligand as well as an ‘electrolyte’ to differentiate electrochemically be-

tween the different metal ions. Since many metal ions form EDTA complexes whose strength depends on pH it was important to establish if the Ep values of the complexes formed with EDTA change with pH. The effect of pH on the formation of complexes between EDTA and Cu, Pb, Fe, Cd and Zn is given in Table 2. The Ep values of all elements were found to be pH dependent (Table 2). When glassy carbon is used as the indicator electrode, the best pH value is 4.0 to minimise the effect of hydrogen evolution on the Zn stripping peak. The results obtained unfortunately show that it was not possible to obtain reproducible results with this electrolyte due to the following: • At the optimum pH (>4) used to minimize the hydrogen evolution effect on Zn determination,

Table 2 Ep values of the elements in EDTA solution at different pH valuesa pH

Pb

Cu

Cd

Fe

Zn

5.0 4.0 3.0

No peak −0.474 (1.332) −0.461 (1.642)

−0.336 (1.274) −0.043 (3.543) 0.043 (0.798)

No peak −0.731 (2.952) −0.682 (1.893)

−0.356 (0.949) −0.306 (0.328) −0.247 (0.897)

−0.998b −1.063 (1.116) −1.129 (0.725)

a b

Peak currents (␮A) are shown in brackets. Concentration of the different ions are 150 ␮g l−1 . The peak was not reproducible.


205

Fig. 2. Voltammogram of Cu, Cd, Pb and Zn in solution containing pyrophosphate in acetate buffer (pH 4.0); (A) without and (B) with 0.005% gelatin in the solution.

the Pb and Cd stripping peaks were not detected. Furthermore, the Zn peak was not reproducible. • EDTA precipitated when lower pH values were used. • The negative cut-off cathodic potential of a glassy carbon electrode (−800 mV) shifted to a more negative value with increase in pH and as a result the determination of Zn at lower pH values became problematic. On the basis of the previous studies [18], the influence of pyrophosphate in acetate buffer at pH 4.0 as supporting electrolyte was studied. A voltammogram of a solution containing all the studied elements is shown in Fig. 2A. An adequate separation of peaks was obtained with almost all elements except for the peaks of Pb and Fe. The addition of 0.005% gelatin solution as supporting electrolyte was investigated (Fig. 2B). The results obtained show that the Pb and Fe peaks were separated in the presence of gelatin, but unfortunately the height of the Cd and Zn peaks decreased at the same time. Furthermore, the electrode also required more frequent cleaning with rather than without the addition of gelatin, which is not suitable for routine work. The selection of the deposition potential for the ions of all five elements (Cu, Pb, Fe, Cd and Zn) involved in the simultaneous determination depends on the metal that has the most negative peak potential, which happens to be zinc at −1.050 V. The zinc signal increased as the deposition potential became more negative, but unfortunately hydrogen evolution occurred at negative

potentials. A reduction potential of −1.25 V was chosen as the best compromise for the zinc signal and hydrogen evolution. The deposition time was controlled by the zinc and copper signals because they had low sensitivity. A deposition time of 3 min and a pulse amplitude of 50 mV were preferred due to the good sensitivity and separation of peaks that were achieved for all the metals. 3.2. Intermetallic effects To determine if there are any intermetallic effect between the different elements, simultaneous deposition varying between 3 to 13 min of four elements (Pb, Zn, Cd and Cu) was studied. Slope values for the peak heights with the same deposition time for the simultaneous determination of Pb, Zn (both 100 ␮g l−1 ), Cd and Cu (both 50 ␮g l−1 ) in the same solution were 0.285, 0.108, 0.268 and 0.121 ␮A min−1 , respectively. When solutions containing single cations Zn (100 ␮g l−1 ) Cu (50 ␮g l−1 ) were studied, the slopes for these solutions were 0.134 ␮A min−1 for Zn and 0.104 ␮A min−1 for Cu. The slope values obtained are representative of the sensitivity of this technique for each metal. It could be concluded from these values that the sensitivity for the cadmium assay is better than that for copper (mCd/mCu=2.214) for the same concentration level. It also followed from the results obtained that the sensitivity for the determination of lead is better than that of zinc (mPb/mZn=2.639). The results also revealed that the dissolution slope

206


Table 3 Linear range, slope and % relative standard deviation (RSD) obtained for Cu, Cd, Pb and Zn with a deposition time of 3 min, pulse amplitude of 50 mV, flow rate of 4.75 ml min−1 and a scan rate of 30 mV s−1 Metal

Linear range (␮g l−1 )

Slope (␮A ␮g l−1 )

Detection limits (␮g l−1 )

RSD (%)

Cu Cd Pb Zn

50–500 50–500 50–500 100–400

0.49 0.74 0.030 0.009

39 9.6 16.6 14.7

14 5.6 13.2 13.6

for the copper assay using a solution containing only the copper cation is smaller than that for copper in the presence of other elements, while the opposite was observed for zinc. These differences are more pronounced in solutions containing higher concentrations of the elements; this is due to the formation of an alloy between zinc and copper, which has a peak potential closer to that of copper. The final effect is an increase in the copper signal and a decrease in the zinc signal. This intermetallic effect has been reported previously in the determination of zinc and copper using mercury as indicator electrodes. The presence of iron affects the Cu, Cd and Pb peaks. The iron effect on copper is a result of the formation of copper iron(III) pyrophosphate [Cu3 Fe3 (P2 O7 )3 ·12H2 O] [19]. Although iron has a peak separate from both Cd and Pb, the simultaneous determination of these three elements is not easy, due to the vicinity of their stripping peaks. These effect may also be a result of changes that occur on the electrode surface which then affects its properties. Despite these problems the presence of low concentrations of iron cannot interfere due to its low sensitivity. It was observed that the simultaneous determination of four metal ions requires careful attention to the concentration ratio at which the metal ions are present, especially cadmium and lead. A large excess of one metal ion may affect the response of the others due to

a change in the coverage on the electrode surface, and the increase in peak areas due to the closeness of their stripping peaks. In extreme cases it was observed that the potentials of the peaks not only shifted, but that the shapes of peaks also changed. 3.3. Applications To evaluate the proposed system, some analytical parameters such as linear range and detection limits were determined. The detection limits, taken as the concentration that gave a signal three times the standard deviation of the blank signal were calculated from the calibration slopes, for the simultaneous determination. The results are shown in Table 3. The detection limits and linearity can be improved by increasing the deposition time and pulse amplitude. The proposed system was tested for the simultaneous determination of Cu, Cd, Pb and Zn in water samples used for boiler feeds and in power stations. The results are given in Table 4. The sensitivity and accuracy was assessed by recovery experiments. The results obtained were good except for copper in the presence of high concentrations of Zn, due to the intermetallic effect between copper and zinc. The recoveries of 95% as well as the relative standard deviations values (≤14%) demonstrated the reliability of using the proposed system for simultaneous heavy metal

Table 4 Simultaneous determination of Cu, Cd, Pb and Zn in water samples used for boiler feeds and in power stations Metal

Cu Cd Pb Zn

Water samples (␮g l−1 )

Conc. added (␮g l−1 )

Conc. found (␮g l−1 )

Recovery (%)

1

2

1

2

1

2

1

2

9.52 16.67 9.54 14.29

2.38 – – 114.0

50 50 100 100

75 75 150 150

54.4 67.8 103.9 105.2

83.3 76.5 160.6 178.7

91 102 95 92

108 98 107 83


monitoring in water samples. One run takes about 5 min, without the wash time, resulting in a sample throughput frequency of 8–12 samples per hour. References [1] Y. Nagaosa, T. Mizuyuki, Anal. Chim. Acta 311 (1995) 225. [2] Y.A. Zolotov, E.I. Morosanova, S.V. Zhalavannaya, S.S. Dyukarev, Anal. Chim. Acta 308 (1995) 386. [3] O.W. Lau, S.Y. Ho, Anal. Chim. Acta 280 (1993) 269. [4] R.M. Liu, D.J. Liu, A.L. Sun, Talanta 40 (1993) 511. [5] M.D. Luque de Castro, M. Valcarcel, Analyst 109 (1984) 413. [6] A. Rios, M.D. Luque de Castro, M. Valcarcel, Analyst 110 (1985) 277. [7] R.N. Khandekar, R.G. Dhaneshwar, M.M. Ralrecha, L.R. Zarapkar, Fresenius’ J. Anal. Chem. 307 (1981) 365. [8] E.S. Pilkington, C. Weeks, A.M. Bond, Anal. Chem. 48 (1976) 1665. [9] H. Bultstein, A.M. Bond, Anal. Chem. 48 (1976) 759.

207

[10] L. Vos, Z. Komy, G. Reggers, E. Roekens, R. Van Grieken, Anal. Chim. Acta 184 (1986) 271. [11] J.H. Aldstadt, H.D. Dewald, Anal. Chem. 64 (1992) 3174. [12] M.G. Paneli, A.N. Voulgaropoulus, Fresenius’ J. Anal. Chem. (1994) 348. [13] T.C. Pan, C.S. Horng, S.R. Lin, T.A. Lin, C.W. Huang, Biol. Trace Element Res. 14 (3) (1993) 233. [14] Z. Khoulif, C. Jombon, M. Chatelut, O. Vittori, Electroanalysis 5 (1993) 339. ◦ [15] J. Ružiˇcka, E.H. Hansen, Flow Injection Analysis, 2nd Edition, Wiley, New York, 1988. [16] M. Valcárcel, M.D. Luque de Castro, Flow Injection Analysis. Principles and Applications, Horwood, Chichester, 1987. [17] J.L. Burguera, Flow-Injection Atomic Spectrometry, Marcel Dekker, New York, 1989. [18] J.F. van Staden, M.C. Matoetoe, Anal.Chim. Acta 376 (1998) 325. [19] J.W. Mellor, Comprehensive Treatise on Inorganic And Theoretical Chemistry, Vol XIV, Longmans & Green, London, 1935.