ANALYST, JANUARY 1992, VOL. 117 31 Determination of Lead in Wines by Hydride Generation Atomic Absorption Spectrometry Juan Cacho, Vicente Ferreira and Cristina Nerin Department of Analytical Chemistry, Faculty of Sciences, University of Zaragoza, Zaragoza, Spain The optimization of lead hydride generation in aqueous ethanolic media and the influence on its generation of the wine components, both white and red, have been studied. These interferences were overcome by careful control of the parameters affecting hydride generation and the procedure was applied to the determination of Pb in wines. The method is fast, accurate and sensitive and can be used to quantify 24 ppb of Pb in wines. Keywords: Lead determination; atomic absorption spectrometry; wines; h ydride generation; atomization Lead is one of the elements that has received considerable attention in recent years, because of its high toxicity, and the broad range of organs and systems that are affected in man and animals through many routes. Lead is present in the atmosphere and in foodstuffs, either naturally or as a contaminant, although usually at very low concentrations.Therefore, the determination of Pb requires procedures that are sufficiently sensitive for detection at the ppb level. Atomic spectrometry is one such procedure. Lead hydride generation and its measurement by atomic absorption spec- trometry has been shown to meet the sensitivity and selectivity requirements for the determination of Pb. However, the efficiency of the lead hydride generation is critically depen- dent on the generation conditions.For this reason, much work has been carried out to optimize the acidity, and on the use of an oxidant,' the nature of the acid and oxidant'-10 and their role in the generation.3.11-12 Whereas the generation of lead hydride in water is well established, the same is not true for aqueous ethanolic media, where the generation efficiency is very low. This is a great disadvantage, as it prevents this method from being directly applied to the analysis of alcoholic beverages, and conse- quently tedious procedures must be used. The official method for determining Pb in wine involves digestion with mineral acids followed by spectrophotometric determination, 13 which is clearly more time-consuming and also less sensitive. In order to apply the lead hydride generation method to such an important beverage as wine, the generation of the gas in the presence of the characteristic components of wine was optimized, and a procedure for determining Pb in wine with this technique was developed.This paper describes the results of the study. Experimental Apparatus A Perkin-Elmer 2280 atomic absorption spectrometer and a Perkin-Elmer MHS- 10 hydride generator were used. The argon head pressure was 250 kPa. The pH measurements were made using a Crison 2002 pH meter. Reagents Aqueous solutions of Pb(N03)2, 6 g 1-1 and 6 ppm. Identical solutions of Pb(N03)2 in wine and in aqueous ethanolic-tartaric acid media (12.5% ethanol, 0.4% tartaric acid). Sodium tetrahydroborate solutions in 1% NaOH, 5 , 7 , 10, 12, 15, 17, 19, 21, 23 and 25%.Hydrochloric acid, 0.8, 1.2 and 1.4 mol dm-3. Ethanol solution in water, 12.5% with 0.4% mlv tartaric acid. Aqueous HCl solution, p H 0.86. Hydrogen peroxide solutions, 4, 5 , 6 , 7, 7.5, 8 and 10%. Study of the Influence of Different Components of Wine In the reaction flask place 5 ml of Pb solution in wine, or Pb in a 12.5% ethanol-0.4% aqueous tartaric acid solution. Add 3 ml of HCl solution and 2 ml of H202. Turn on the argon flow and inject the NaBH4 solution in 1% NaOH, to obtain a signal maximum at 283.3 nm. One parameter was varied while keeping the others fixed. The optimum values were adopted for subsequent studies. Recommended Procedure Take 30 ml of wine, adjust the pH to 0.86 by adding 1.2 mol dm-3 HCl and dilute to 50 ml with aqueous HCl solution (pH = 0.86).Transfer 7.5 ml of this solution into the hydride generator reaction vessel, add 2.5 ml of H202 (7.5%) and leave the argon gas flowing for 3 min. Then inject the 21% NaBH4 solution until a maximum signal is obtained at 283.3 nm (after approximately 10 s). Results and Discussion Ethanol, tartaric acid and SO2 were chosen as the characteris- tic components of wine. Their individual and combined influence on the generation of lead hydride was studied. Influence of Ethanol As expected, on changing from aqueous to aqueous ethanolic media, the hydride generation efficiency decreased markedly , as can be seen in Fig. 1. However, above a concentration of 5% ethanol, the efficiency varies very little, which may be owing to the low variation in pH and oxidation-reduction potentials of the substances present.This slight variation is an 0 5 10 15 20 25 Ethanol (%) Effect of the presence of ethanol on PbH4 generation Fig. 132 ANALYST, JANUARY 1992, VOL. 117 0.2 I I I I I 1 0 2 4 6 8 10 Tartaric acid/g I-' 1 I I 1 1 I 1 0 0.2 0.4 0.6 0.8 1 .o Sulfur dioxide/g I-' Fig. 2 Effect of the presence of tartaric acid and SO2 on PbH4 generation. A, Tartaric acid; and B, SO2 V J I I 1 I I I 4 5 6 7 8 9 10 Hydrogen peroxide (%) Optimization of the concentration of H202. The signal value is Fig. 3 expressed in relation to the optimum signal obtained in water advantage when preparing a calibration graph for the determi- nation of Pb in alcoholic beverages, as the ethanol content in different samples is generally similar, although not identical. Influence of Tartaric Acid The presence of tartaric acid also affects the lead hydride generation efficiency, although to a lesser extent than ethanol, as shown in Fig.2. The decrease is due to a modification of the solution pH. Influence of Sulfur Dioxide Sulfur dioxide at the concentrations normally found in wines does not interfere in lead hydride generation even if the SO2 concentration is more than twice the highest SO2 concentra- tion founded in the wines (1 g 1-1). Higher concentrations have not been studied, as they are unlikely to be encountered in real samples. Fig. 2 shows the influence of SO2. Optimization of Lead Hydride Generation in Aqueous Ethanolic Media Once the influence of ethanol and tartaric acid was known, the generation of the lead hydride in aqueous solutions containing proportions of these compounds similar to those found in wine was studied.The concentrations of the reducing agent, HCl and oxidizing agent were examined. The lead hydride absorption signals obtained in this way were compared with those obtained in water under optimum conditions. The influence of acidity reveals that the HCI concentration has a strong influence on the generation of lead hydride, and that small variations in the acid concentration produce large variations in the signal. This marked effect is not observed in waters nor in organic solvents such as dimethylformamide (DMF).14 Less HCl is required to produce the maximum signal in this medium than in water. The optimum pH was found to be 0.86.I I I 1 I 1 0 5 10 15 20 25 Tetrahydroborate (%) Fig. 4 Optimization of the concentration of NaBH4. The signal value is expressed in relation to the optimum signal obtained in water The peroxide concentration is also important, although much less so than that of HCI. Working with optimum HCl concentrations, the maximum signal is obtained at an H202 concentration of 7.5% (Fig. 3 ) , i . e . , the same concentration which gives the maximum signal in water. However, the signal varies much less with variations in H202 concentration in an aqueous ethanolic medium than in water, unlike HCI. The variation of the atomic absorption signal with the NaBH4 concentration is completely different from that found in water. Using optimum conditions for HCl and H202, the signal maximum is found when the concentration of added NaBH4 is 21% (Fig.4), whereas in water the maximum is obtained with 7%. However, even with the different optimum conditions for lead hydride generation found in water and in aqueous ethanolic media, the signals are equal for the two media if the amount of Pb is the same and in both instances the work is performed under the respective optimum conditions; Le., the interferences due to ethanol and tartaric acid are eliminated. Other experimental parameters were studied in order to optimize the working conditions. The argon flow rate was varied slightly and it was found that small variations in the argon head pressure did not affect the generation and detection conditions and that 250 kPa was the optimum argon head pressure.The two main absorption lines for Pb were tested and it was found that better blanks were obtained by working at 283.3 nm. Lead Hydride Generation in Wine It is difficult to study the influence on lead hydride generation of other compounds present in wine because of their large number, and the fact that there are no standards for most of them. Therefore, for this study, different wines were taken and the lead hydride signal variation was measured when the generation conditions varied slightly around the optimum conditions found above for aqueous ethanolic media. The characteristics of the wines are given in Table 1, and as there is a broad range, the results can be extrapolated for any type of wine. A study of the signal variation with the pH shows that the other components in wine do not alter the optimum pH, which is still 0.86 in all instances.However, they do produce substantial modifications in the dependence curves (Fig. 5). Therefore, this parameter must be measured with a pH meter to guarantee reproducibility and accuracy. The signal variation with NaBH4 and H202 concentrations is similar to that found in aqueous ethanolic media, and the maximum is obtained at the same concentration, 7.5% H202 and 21% NaBH4. This last study showed that the lead hydride signal depends on the reaction time between the oxidizing agent and the wine components. A study of the influence of reaction time showed that white and red wines behave differently, as shown in Fig. 6. White wines take about 50 s to reach the maximum signal, whereas red wines need 2.5 min.After this time, the signal isANALYST. JANUARY 1992, VOL. 117 33 0.6 I 1 Table 1 Physico-chemical characteristics of the wines selected for the study Acidity/ Ethanol mequiv 1-1 Relative Total SO2 Extract/ Sample (%) H2S04 density (ppm) gl-I 1 Cava-brut 2 White-dry 3 White-dry 4 White-dry 5 White-dry 6 White-semi 7 Rose 8 Rose 9 Rose 10 Red 11 Red 12 Red 13 Red 14 Red 10.97 11.94 12.92 12.40 13.10 12.27 12.73 11.30 10.45 13.16 13.29 13.13 13.14 12.94 85 79 77 66 72 93 69 72 81 76 69 70 84 74 0.99230 0.99175 0.99090 0.99 140 0.99095 0.99465 0.99070 0.99165 0.99670 0.99290 0.99227 0.99660 0.99435 0.99330 157 15.4 155 19.3 188 20.1 162 14.8 212 20.9 232 27.9 113 19.0 70 23.7 129 21.4 115 26.1 106 27.2 0 19.8 248 22.6 104 26.3 20 0.7 I 0.8 0.9 1 .o PH Fig.5 Variation of the signal with pH in two different types of wine the same for both white and red wines and is stable for at least a further 3 min. For this reason, in other studies and in the recommended procedure, the reaction time between the H202 and Pb in the wine was fixed at 3 min. After this time, the NaBH4 solution is added and the lead hydride generation takes place rapidly, reaching the maximum value in 10 s . Having found the optimum generation conditions, the parameters of the analytical method were studied. The following were found: Beer’s law holds for all wines up to at least 800 ppb of Pb; the calibration graph is described by the equation A = 0.005 + 0.75c, where c is expressed in ppm and the correlation coefficient, r = 0.998.The value of the blank varies considerably with the origin of the NaBH4 used (this study found extreme absorbance values of 0.015 and 0.062). For the same quality of NaBH4, the standard deviation of the measurement was 0.002 A. Under these conditions, the detection limit, for a value of K = 10, is 24 ppb. In order to test the repeatability of the method six different wines were used. Each wine was analysed following the proposed method with five replicates. The reproducibility was tested by repeating exactly these operations on five different working days. The mean relative standard deviation (RSD) found for the repeatability was 2.2% with extreme values of 1.3 and 3.2% for each of the five working sessions and the mean RSD found for the reproducibility was 2.4% with extreme values of 1.9 and 3.7%.This procedure was tested on wines with different charac- teristics and origins. The Pb content was determined both by interpolation of the calibration graph and by the standard additions method. Table 2 shows the results obtained, together with the characteristics of the wines analysed. 0.5 0.4 a C 4 0.3 0 Ll a 0.2 O.’ tlf 14 I I I 0 100 200 300 Tirne/s Fig. 6 Variation of the Pb signal with time of reaction between the oxidizing agent and white and red wines. A, Red wine; and B, white wine Table 2 Comparative study between the official method, proposed method and standard additions method (mean of three replicates) Pb content (ppb) Sample 1 Cava-brut 2 White-dry 3 White-dry 4 White-dry 5 White-dry 6 White-semi 7 Rose 8 RosC 9 RosC 10 Red 11 Red 12 Red 13 Red 14 Red * Ref.13. Proposed method 63 89 65 0 123 91 280 72 17 162 95 77 49 43 Standard additions method 69 94 68 1 116 86 275 70 14 159 95 76 51 45 Official method* 64 92 68 0 119 88 274 74 4 167 98 79 49 41 simple, fast, sensitive and accurate, and is therefore recom- mended for determining the Pb content of wine. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Fleming, H. D., and Ide, R. G., Anal. Chirn. A d a , 1976,83,67. Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 595. Vijan, P. N., and Wood, G. R., Analyst, 1976, 101,966. Jin, K., and Taga, M., Bunseki Kagaku, 1978, 27, 759. Jin, K., and Taga, M., Bunseki Kagaku, 1980, 29, 522. Vijan. P. N., and Sadana, R. S . , Talanta, 1980, 27, 321. Smith, R., At. Spectrosc., 1981, 2, 155. D’Ulivo, A., and Pappof, P., Talanta, 1985, 32, 383. Ikeda. M., Jiro, N., and Nakahara, T., Bunseki Kagaku, 1981, 30, 368. Kumamaru, T., Nakata, F., and Hara, S . , Bunseki Kagaku, 1984. 33, 624. Jin. K., and Taga, M., Anal. Chim. Acta, 1982, 143,229. Nerin, C., Olavide, S . , Cacho, J . , and Garnica, A . , Water, Air, Soil Pollut., 1989, 44, 339. Metodos Oficiales de Analisis, Ministerio de Agricultura, Madrid, 1980. Aznarez, J . , Palacios, F., Ortega, M. S., and Vidal, J . C., Analyst, 1984, 109, 123. Conclusion Lead can easily be determined in wine by generating its hydride and atomizing it in a silica tube. The procedure is Paper 0101 166 E Received March 16, 1990 Accepted July 26, 1991