ANALYST, JANUARY 1989, VOL. 114 25 Applicability of Gallium as Copper Scavenger in the Determination of Zinc in Samples of High Copper Content by Potentiometric Stripping Analysis Stavros V. Psaroudakis and Constantinos E. Efstathiou" Laboratory of Analytical Chemistry, University of Athens, 104 Solonos Street, Athens 106 80, Greece ~ ~~ The determination of zinc in the presence of copper by potentiometric stripping analysis is hindered by the formation of Cu - Zn intermetallic compounds on the working electrode during the deposition step. An excess of gallium is usually added to the sample to prevent the formation of Cu - Zn by forming much more stable Cu - Ga intermetallics. If the Cu : Zn mass ratio is higher than about 5, only part of the analytical signal of zinc is restored after addition of gallium and both the precision and accuracy of the determination are reduced.A simple extraction procedure is recommended to remove the bulk of copper from the sample. The pH of the working sample solution is adjusted to 4 (0.5 M acetate buffer) and is extracted with half of its volume of a 3% VNsolution of acetylacetone in benzene. Each extraction removes about 80% of the copper present without affecting the zinc. The determination of zinc is subsequently carried out after adding Hg" as an oxidant and gallium to mask the remaining copper. The method was applied to the determination of zinc in low-zinc, hig h-copper brass samples. Keywords: Potentiometric stripping analysis; zinc determination; intermetallic compounds; acetylacetone extraction of copper The determination of zinc and copper in the presence of each other by electrochemical stripping techniques such as anodic stripping voltammetry (ASV) and potentiometric stripping analysis (PSA) is hindered by the formation of intermetallic compounds between these two metals.These Cu - Zn intermetallics are formed during the deposition step and remain dissolved or as a separate crystalline phase on the mercury film of the working electrode. Kemula et al. 1 were the first to report on this type of interference in determinations of zinc by ASV. Shuman and Woodward,2 using data obtained by ASV and cyclic voltammetry, showed the formation of three compounds of the type CuZn, (n = 1-3), soljuble in the mercury phase, and the formation of an insoluble compound CuZn3, at high amalgam concentrations.More recently, Piccardi and Udisti3 examined the problems arising from the formation of Cu - Zn intermetallics in the determination of the apparent complexing capacity (versus copper) of sea water by ASV. Their findings indicated the formation of insoluble CuZn and soluble CuZn2 intermetallics. During the stripping step, the Cu - Zn intermetallics are oxidised at potentials close or equal to the stripping potential of copper. Therefore, depending on the acutal Cu : Zn ratio, the resulting zinc signal is decreased or completely eliminated, whereas the copper signal is increased. This is a particularly problematic situation as zinc and copper are ubiquitous metals found in a wide diversity of samples. In the determination of copper, the above error can be readily prevented by using a deposition potential less cathodic than that required for the deposition of zinc.Unfortunately, the deposition of zinc without the parallel deposition of copper is not feasible. Therefore, the analytical problem related to the determination of zinc cannot be alleviated as easily. Copeland et al.4 suggested the addition of an excess of gallium to the sample. Gallium forms more stable intermetal- lics with copper, allowing the oxidation of zinc at its normal stripping potential. This method is widely used in determina- tions of zinc by ASV, and it has been adopted by Jagner and co-workers5.6 in analogous determinations performed by PSA. Hoyer and Kryger7 applied the generalised standard additions method (GSAM)8 in PSA in order to resolve overlapping signals and to correct those signals affected by the formation of intermetallics.* To whom correspondence should be addressed. On trying to apply the gallium treatment in the determina- tion of zinc by PSA to samples containing copper in excess over zinc, we found that the analytical signal of zinc is only partially restored. Also, non-linear and/or heavily scattered zinc signal versus [Zn"] plots were obtained on trying to apply the multiple standard additions technique to the calculation of the concentration of zinc. Therefore, the sensitivity, precision and accuracy were reduced under these conditions. The limits of the scavenging action of gallium on copper in the determination of zinc in samples of high copper content by PSA are examined in this paper.Experimental Instrumentation The microcomputer-controlled PSA system, the cell and the electrodes used in this study have been described briefly elsewhere.9 The deposition and stripping steps took place under continuous delivery of oxygen-free nitrogen and con- stant mechanical stirring. A glass propeller rotated at 1300 rev min-1 was used for stirring instead of the magnetic bar used previously. Reagents Stock solutions (1000 mg 1-1) of zinc, copper and mercury are prepared by dissolving the appropriate amount of the corre- sponding nitrate (analytical-reagent grade) in distilled, de- ionised water. Spectroscopic quality (Johnson-Matthey) cop- per sulphate is used for the preparation of solutions with high Cu : Zn ratios.Gallium stock solution (10000 mg 1-1) is prepared by dissolving the appropriate amount of pure Ga203 in excess of 6 M HCl. Boiling is required to facilitate the dissolution. The excess of HC1 is expelled on a steam-bath before diluting to the appropriate volume. An acetate buffer (pH 4.0,0.50 M in total acetate) is used as an ionic medium. Procedure Measurements First step: determination of copper. The PSA system is programmed as follows (all potentials are vs. SCE) : deposi-26 ANALYST, JANUARY 1989, VOL. 114 tion potential, -0.80 V; deposition time, 60 s; final potential, 0.00 V; resting potential, 0.00 V; and stripping mode, "convective. " A 30.00-ml aliquot of the working sample solution contain- ing copper in the range 0.1-5 pg ml-1, spiked with t-Ig(N03)2 to a final HgII concentration of 10 pg ml-1, is transferred into the measurement cell and de-aerated for 10 min.Three plating - stripping cycles serve as an in situ plating procedure for the formation of the mercury film on the glassy carbon electrode. The potential is recorded during the stripping cycle and the copper E - t plateau is evaluated on the recordings graphic- allylo (Fig. 3 in reference 10, curve 111). The concentration of copper is calculated by the multiple standard additions method. Three additions of the appropriate volumes of the stock C U ( N O ~ ) ~ solution are usually made for better accuracy. Second step: determination of zinc. The working sample solution is spiked with Ga"' to a final concentration approxi- mately ten times that of Cut1 (in pg ml-1).The deposition potential is set to - 1.40 V and a measurement is obtained as before. The zinc E - t plateau, followed by the gallium E - t plateau, is measured graphically and the concentration of zinc is calculated by the multiple standard additions method. Three additions of the appropriate volumes of the stock Zn(N03)2 solution are made. For better accuracy, each standard addition should increase the zinc signal by a step within the range 25-50% of the initial signal. If the calculated concentra- tion of zinc is less than one fifth of the copper concentration, the result should be suspected to be inaccurate and this step is then repeated after extracting the working sample solution as described below. With high initial concentrations of copper, the first step should also be repeated to confirm the efficient removal of the bulk amount of copper.Extraction of copper from the working sample solution A 40-ml aliquot of the working sample solution is extracted with 20 ml of a 3% V/V solution of acetylacetone in benzene. For working solutions with higher copper concentrations, the extraction can be repeated with fresh amounts of organic phase. Each extraction removes about 80% of the copper. Five successive extractions remove almost all of the copper (>99.5%) without any noticeable decrease in the concentra- tion of zinc. Preparation of the working solution of the brass samples An accurately weighed amount of a low-zinc, high-copper brass sample (1-5% Zn, 80-95% Cu) in the range 0.8-1 g is dissolved in 20 ml of 8 M HN03.The solution is evaporated to about 5 ml and diluted to about 100 ml with water. The white hydrous SnOz precipitate is filtered off and washed thoroughly with warm 0.8 M HN03 solution. The filtrate is diluted with water to 200.0 ml and a 1.00-ml aliquot of this solution is diluted to 500.0 ml with the acetate buffer. This is the working sample solution containing zinc in the range 0.08-0.50 pg ml-1 and copper in the range &lo pg ml-1. Results and Discussion Zinc - Copper Mutual Interference Patterns The analytical signal ( E - t plateau) of zinc was measured in ZnII solutions after successive additions of Cull. Four different concentrations of Zn" were used and the resulting plots of zinc signal versus Cu : Zn molar ratio are shown in Fig.1 . All plots indicate a sharp decrease in the zinc signal which can be described by two intersecting lines. The average value of the Cu : Zn molar ratio, at the intersection points, is 0.33 k 0.04. The ratio of metal amalgam concentration to metal solution concentration, after a given deposition period (accumulation coefficient), is approximately the same for both metals.2 Therefore, the intersections at this particular ratio are further h 1 v) C N -. A .c 0 0.5 1 .o 1.5 [ C ~ ~ i I i l Z n ~ ~ l Fig. 1. Effect of increasing Cu: Zn molar ratio on the analytical signal of zinc at Zn" concentrations of (A) 0.33; (B) 0:67; (C) 1.?0; and (D) 1.33 pg ml-1. Deposition time, 60 s; deposition potential, -1.40 V 0 1 2 3 4 5 6 [ZnllI/[C~~~l Fig. 2. Effect of increasing the Zn : Cu molar ratio on the analytical signal of copper at Cull concentrations of (A) 0.13; (B) 0.33; (C) 0.67; and (D) 1.33 pg ml-1.Deposition time, 60 s; deposition potential, -1.40 V evidence that a CuZn3 intermetallic compound is formed when zinc is present in excess over copper. Crosmun et al." have reported an analogous experiment using differential-pulse ASV. An intersection point at a Cu : Zn molar ratio of 1 was noted, denoting the formation of intermetallic CuZn. These contradictory results may be attributed to the much thinner mercury film that we used (plating with an Hg" concentration of 1300 pg ml-1 was used by Crosmun e f ~1.11). Probably, the formation of CuZn and CuZn3 is favoured at low and high amalgam concentrations, respectively.This experiment was repeated using gallium instead of zinc. The resulting plots of gallium signal versus Cu : Ga molar ratio gave a single intersection point at a molar ratio of 1, denoting the formation of CuGa species. The effect of increasing Zn : Cu molar ratio on the copper signal was studied in a similar experiment. The resulting plots with four different initial concentrations of Cut1 are shown in Fig. 2. In all instances there are intersections of linear and non-linear graphs. An intersection point at a Zn : Cu molar ratio of about 0.9 appears consistently in all graphs, whereas a second intersection point, at a Zn : Cu molar ratio of about 1.9, is apparent in only two of them. These plots should be considered as evidence of the successive formation of CuZn and CuZn2 intermetallics. Obviously, each intermetallic is oxidised at a different rate by Hg", resulting in the patterns shown in Fig.2. A further increase in ZnII concentration increases the copper signal because of the enhanced deposi-ANALYST, JANUARY 1989, VOL. 114 27 0 2 4 6 8 PH Fig. 3. Effect of the pH on the analytical signal of gallium for a solution containing 0.67 pg ml-1 of Ga"'. Deposition time, 60 s; deposition potential, - 1.40 V; ionic medium, HCI - CH,COOH - NH4Cl. each 0.10 M; the pH was adjusted by adding small volumes of 5 M NaOH 0 ' ' I I I 0.1 1 10 100 [Gallllipg ml-' Fig. 4. Effect of the concentration o f gallium on the analytical signal of zinc (0.100 ug ml-1) in the presence of copper at concentrations of (A)O.lOOpgml-1; (B) l.OOpgml-';and(C)S.00p.gml-~. Deposition time, 60 s; deposition potential, -1.40 V tion of copper during the stripping of zinc (class ii interfer- enceg).The absence of purely linear relationships between the analytical signals of zinc and copper and their respective concentration ratios precludes computational corrections based on simple linear models of interferences.8 The computa- tional approach can be applied only over a relatively narrow concentration ratio range, and therefore it is of limited analytical importance. Selection of Ionic Medium The proposed ionic medium has the appropriate pH for the efficient separation of copper from the sample by extraction with acetylacetone without affecting the zinc present,l2 and it is also appropriate for the efficient deposition of zinc and gallium. Therefore, the over-all procedure is simplified, as further pH adjustments are not required.It also has the appropriate buffering capacity, as the gallium analytical signal is critically dependent on the pH of the solution (Fig. 3). In acidic solutions (e.g., 0.1 M HCl) the deposition of gallium is very poor. The relatively high concentrations of acetic acid and sodium hydroxide used for the preparation of the ionic medium make blank determinations for zinc necessary. Blanks for zinc were determined by PSA with each batch of the ionic medium. The average value of the blank ZnII concentrations was 0.010 k 0.002 pg ml-1. These blank values are much lower than the actual lower determination limit of ZnII (0.080 pg ml-1) in the present work, and therefore all analytical results can be simply corrected.If the determination limit is extended toward lower ZnII concentrations, e.g., by increasing the deposition time, the ionic medium should be purified electrolytically. Restoration of Zinc Signal With Gallium The percentage restoration of the zinc signal as a function of Gal11 concentration for three solutions containing 0.100 pg ml-1 of Zn" and different concentrations of Cu" (0.100, 1.00 and 5.00 pg ml-1) is shown in Fig. 4. In all instances the zinc signal reaches a steady value after adding about a 10-fold excess of Ga"' over Cur1 (concentrations in pg ml-1). These steady values correspond to a 100, 60 and 25% restoration of the original zinc signal (in the absence of CU" and GaII') for the three concentrations of copper examined.The observed partial restoration of the zinc signal may be attributed to saturation of the mercury film with Cu - Ga intermetallics and partial coverage of the mercury film with them, with a concomitant reduction in the hydrogen over- potential. This reduction in the hydrogen overpotential decreases the accumulation coefficient of zinc, also adversely affecting the reproducibility of the zinc signal. The decrease in the zinc signal appears to be the limiting factor preventing the application of the simple gallium treatment to samples of high copper content. Using ASV it has been reported that the strong overlap of the high gallium current peak with the small zinc current peak is the actual limiting factor.4 In PSA, the gallium E - t plateau (E+ = -0.88 V) not only does not obscure the development of the preceding zinc E - t plateau (Ei = -1.10 V), regardless of their relative widths, but also facilitates the evaluation of the latter on the recordings by the graphical procedure we have used.Removal of Bulk Amounts of Copper Large amounts of copper can be effectively removed from solutions containing minute amounts of zinc by controlled potential bulk electrolysis. This procedure is extremely time consuming and therefore it was not considered further for the present application. Removal of copper by precipitation with one of the many available copper precipitants always involves a risk of the coprecipitation of zinc. On the other hand, the excess of the precipitating reagent would also precipitate Hg", which is subsequently added to the working sample solution as an oxidant.Solvent extraction seems the most efficient method for removing large amounts of copper in the minimum time period. A wide variety of copper extractants can be found in the literature but only those which do not extract zinc can be considered for the present application. The extractability of various metals by weakly acidic organic extractants can be readily differentiated by controlling the pH of the aqueous phase. A literature search revealed that P-diketones show large differences between pHt,cu and pH,,,, values ( P H ~ , ~ corresponds to the pH of the aqueous phase where the extractability of the metal M is half of its maximum value). Acetylacetone (pentane-2,4-dione), the simplest P-diketone, is inexpensive, it can be used at high concentra- tions in a variety of organic solvents and it establishes extraction equilibria relatively rapidly.Copper is extracted by an acetylacetone solution in benzene (pH,,,, =: 3), whereas zinc is almost non-extractable when the pH of the aqueous phase is less than about 7 (pH:,,, = 8).12 The extraction of copper with acetylacetone is not as complete as might be required and more than one extraction is needed to reduce high Cu : Zn ratios effectively. Acetylacetone is an even weaker zinc extractant. Other more lipophilic P-diketones such as benzoylacetone and dibenzoylmethane can also be used instead of acetylacetone. These P-diketones extract copper almost completely without affecting zinc over a sufficiently wide pH range (pH;,,, = 2.5, PH;,~, = 6), but they are more expensive and relatively slow extractants.28 ANALYST, JANUARY 1989, VOL.114 Table 1. Determination of zinc in synthetic samples containing a fixed zinc concentration (0.100 pg ml-1) and various concentrations of copper using gallium as copper scavenger Concentration of zinc found k SD (n = 5)/ PSml ' Concentration of Without prior With prior copper/pg ml-1 extraction of copper extraction* of copper 0.100 0.100 L 0.008 0.109 k 0,008 (1) 2.50 0.142 i 0.031 0.113 t 0.01 1 (2) 10.0 t 0.107 k 0.025 (3) 50.0 i- 0.102 k 0.019 (4) 100 t 0.102 k 0.006 ( 5 ) 0. so0 0.106 k 0.012 0.099 i 0.005 (1) * The number of extractions required is shown in parentheses. t Not applicable owing to the high concentration of Ga"' needed.Table 2. Determination of zinc in low-zinc, high-copper brass samples Zn found k SD ( n = 5),% Certified content,% Without prior With prior extraction of extraction of 74 92.88 1.00 0.53 * 0.40 1 .00 k 0.08 78 81.04 4.69 4.51 t 1.05 4.79 k 0.30 89 85.50 1.50 1.29 2~ 0.13 1.39 k 0.09 Brasssample Cu Zn copper copper Results Zinc was determined in solutions containing a fixed ZnII concentration of 0.100 pg ml-1 and various CuII concentra- tions in the range 0.1-100 pg ml-1. The determination took place without (wherever possible) and with prior extraction of the working sample solution with acetylacetone. In all instances GaflI was added in a 10-fold excess over the remaining CU" prior to the PSA measurement. The results are given in Table 1.The accuracy of each result and the precisions between the results obtained with both methods were examined statistically at the 95% confidence level. Application of the Student's f-test revealed a significant error in the determination of Zn" concentration in the presence of a Cu*I concentration of 2.5 pg ml-1 without prior extraction of copper. At the same CU" concentration, application of the F-test revealed a significant difference between the precisions obtained, the determination of zinc with prior copper extrac- tion being more precise. The results for the determination of zinc in low-zinc, high-copper brass samples of known composition (Thorn Smith, Beulah, MI, USA) are shown in Table 2. Statistical examination of these results revealed a significant error in the determination of zinc in brass sample 89 without prior extraction of copper, and a significant difference in the precisions of the results obtained with samples 78 and 74, the determination of zinc with prior copper extraction always being more precise.Conclusions The determination by PSA of zinc in samples containing copper with a Cu : Zn mass ratio exceeding about 5 should not be conducted by simple addition of gallium to prevent the formation of Cu - Zn intermetallics. The high concentrations of gallium needed and the reduction in the over-all sensitivity and accuracy of the determination make it imperative to remove the bulk of the copper. The proposed scheme for the extraction of copper with acetylacetone is simple and rapid. Other separation schemes may also be applied.The Zn - Cu mutual interference patterns are indicative of the formation of CuZn, ( n = 1-3) intermetallic compounds during the deposition of these metals on the mercury film electrode. These findings obtained with PSA parallel previous results obtained with ASV. In a previous paper we reported on the strong interference of nickel and cobalt on the zinc signal due to the formation of stable Ni - Zn and Co - Zn intermetallic compounds.' Preliminary results have shown that gallium can also be used as a scavenger of nickel and cobalt. The optimisation of the determination of small amounts of zinc in the presence of large amounts of these two metals by PSA is under investigation. This work was supported in part (50%) by a research grant from the Ministry of Industry, Energy and Technology. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. References Kemula, W., Galus, Z . , and Kublik, Z., Nature (London), 1958, 182, 1228. Shuman, M. S . , and Woodward, G. P., Jr., Anal. Chem., 1976, 48, 1979. Piccardi, G . , and Udisti, R., Anal. Chim. Acta, 1987,202, 1.51. Copeland, T. R., Osteryoung, R. A . , and Skogerboe, R. K., Anal. Chem., 1974, 46, 2093. Danielson, L. G . , Jagner, D., Josefson, M., and Westerlung, S . , Anal. Chim. Acta, 1981, 127, 147. Jagner, D., Josefson, M., and Westerlund, S . , Anal. Chirn. Acta, 1981, 129, 153. Hoyer, B., and Kryger, L., Anal. Chim. Actu, 1985, 167, 11. Saxberg, B. E . H . , and Kowalski, B . R., Anal. Chem., 1979, 51, 1031. Psaroudakis, S. V., and Efstathiou, C. E., Analyst, 1987, 112, 1587. Jagner, D., Anal. Chem., 1978, 50, 1924. Crosmun, S . T., Dean, J. A . , and Stokely. J. R., Anal. Chim. Acfa, 1975, 75, 421. Stary, J., in Irving, H., Editor, "The Solvent Extraction of Metal Chelates." Pergamon Press, Oxford, 1964, p. 54. Paper 8102365 D Received June I4th, I988 Accepted September 12th, 1988