JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 205 Determination of Calcium Potassium Magnesium Iron Copper and Zinc in Maternal Milk by Inductively Coupled Plasma Atomic Emission Spectrometry* Nereida Carrion Ana Itriago Miguel Murillo Elias Eljuri and Albert0 Fernandez Centro de Quimica Analitica Escuela de Quimica Facultad de Ciencias Universidad Central de Venezuela P. 0. Box 4 71 02 Caracas 704 7 -A Venezuela A simple and rapid method was developed for the determination of Ca P Mg Fe Cu and Zn in maternal milk samples by inductively coupled plasma atomic emission spectrometry. Milk samples were emulsified with ethoxy nonylphenol to 0.03% m/v. The emulsified samples were diluted 10-fold with nitric acid and to 1 O/O v/v. Aqueous solutions with the same amount of emulsifier and acid were used as calibration standards.The accuracy of the proposed method was assessed using the National Institute of Standards and Technology Standard Reference Material 1549 Non-Fat Milk Powder. No significant difference at the 95% confidence level was observed. The precision of the method when applied to real samples was in the range 0.3-2% expressed as %RSD with no internal standard. Maternal milk samples from 50 women residing in Caracas Venezuela collected at three different lactation stages (on the third seventh and twenty-first days postpartum) were analysed. Keywords lnductively coupled plasma atomic emission spectrometry; human milk; trace element determination Human milk has long been regarded as the optimal source of essential nutrients for the young infant if the maternal diet is nutritionally adequate and a sufficient amount is consumed.Accurate data on the concentration of trace elements in human milk throughout lactation are important both for formulating nutritional requirements and for obtaining baseline levels leading to an understanding of the physiology of milk secretion.’ For this reason sensitive rapid and precise methods for the determination of trace elements in human milk are necessary. Inductively coupled plasma mass spectrometry (ICP-MS) and ICP atomic emission spectrometry (AES) have been used in the past for the determination of trace elements in Both techniques have analytical characteristics that are appli- cable in the routine determination of elements in milk i.e.low detection limits long linear working ranges and the capability of simultaneous multi-element determinations. Different sample treatments such as dry and wet digestion and decomposition with microwave heating in open and closed vessels have been used. Their efficiency is based on the time required and the completeness of the decomposition. Emmett2 reported trace metal determination in liquid and powdered milk by ICP-MS. Samples were digested by a wet acid pro- cedure and the results obtained compared well with the reference values. The decomposition method was lengthy. Dry ashing3-’ and wet acid digestio1-1~9~ procedures were also used for the determination of trace metals in liquid and powdered milk by ICP-AES. Human milk is a complex colloidal system that can be difficult to dissolve completely.Krushevska et a1.* carried out a comparison of several dry ashing and wet digestion pro- cedures for the determination of Zn in milk samples using ICP-AES. They reported that the wet dissolution procedure with a hot-plate is time-consuming and prone to contamination due to the large amounts of reagent required. The use of microwave systems increased the speed of sample dissolution compared with classical digestion but did not completely eliminate the requirement for chemical reagents. High-pressure digestion with a high-temperature programme destroys the organic carbon almost completely with only HN03. Zinc losses were observed at high temperatures when a dry ashing pro- * Presented at the XXVITI Colloquium Spectroscopicurn Inter- nationale (CSI) York UK June 29-July 4 1993.cedure was used without an ashing aid especially in the presence of chlorides. Direct dilution of sample is simple it can be automated and is less time-consuming than alternative procedures such those mentioned above. Durrant and Ward’ determined 18 elements in milk samples by ICP-MS. The samples were simply diluted to 2% v/v in aqueous acidic solution but the accuracy of the method is not very clear. Emmett2 in an attempt to directly analyse milk samples diluted with water reported a poor accuracy that was attributed to the fatty nature of the milk. Coni et aL5 also tried direct aspiration of the liquid milk samples into an ICP torch. They obtained an analytical signal lower (as much as one third) than those obtained by aspiration of the digested milk sample.This was associated with the rather large average droplets size characteristic of the un-treated milk and the incomplete atomization of components during residence in the plasma. They suggested that this method of determination is impracticable due to the difference in viscosity and surface tension values between real samples and aqueous calibration solutions. It has been reported1@13 when using flame atomic absorption spectrometry that the addition of surfactants to aqueous sample solutions improves the analyte sensitivity. It reduces the average droplet size in aerosols produced by pneumatic nebulizers due to a depression of the surface tension. Nevertheless Bertagnolli et a1.14 have reported that the use of surfactants in ICP does not improve sensitivity. They have mentioned that only the sample transport process is better thereby allowing a more stable plasma when using high sample uptake rates. One of the useful uses of surfactants in atomic spectrometry has been the emulsification of oil samples in water for the direct introduction of emulsions in both and ICP1s22 atomizers.The finely divided particles are uniformly dispersed in the water phase and the sample behaves like an aqueous solution. In this way the use of emulsions without prior destruction of the organic matrix has become an alternative method for sample introduction in atomic spectrometry. The aim of this investigation was to develop a direct rapid and simple method for the determination of Ca P Mg Fe Cu and Zn in human milk by ICP-AES by direct aspiration of emulsified samples.The accuracy was assessed by analys- ing the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1549 Non-Fat Milk Powder.206 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Experimental Apparatus A Jobin-Yvon Model JY24 inductively coupled plasma spec- trometer was used. Experimental parameters are presented in Table 1. Reagents and Materials All reagents were of the highest purity available. Milli-Q de- ionized water (Millipore) was used throughout the work. Nitric acid was Merck Suprapure grade. Ethoxy nonylphenol was obtained from Etoxil of Venezuela. Standard solutions were prepared from 1000 pg ml-' stock solutions (BDH) of each element.A 100 pg ml-' phosphorus standard stock solution was obtained by dissolving K2HP04 (BDH) in high-purity water. Cleaning of the plastic- and glass-ware All laboratory glass-ware and polystyrene plastic-ware con- tainers were acid-washed overnight in 3% v/v nitric acid and rinsed repeatedly with de-ionized water before drying at 60 "C. Once dried all plastic ware was heat-sealed in clean plastic bags to prevent further contamination. Sam p 1 e s A total of 50 lactating mothers provided milk samples for this investigation. All were residents of the city of Caracas (Venezuela). Maternal milk samples were collected at three Table 1 Instrumental parameters ICP atomic emission spectrometer Nebulizer Spray chamber Forward power/W Sample delivery Coolant flow rate/l min-' Nebulizer flow rate/l min-' Sample flow rate/l min-' Observation height Analytical lines/nm Jobin Yvon Model JY-24 Meinhard C-type Scott - type 1000 Peristaltic pump (Sorvall type 49061) 12 0.7 0.5 15 mm above load coil Ca I1 393.66 Cu I1 324.754 Fe I1 239.562 Mg I1 279.553 P 1213.618 Zn I1 213.856 Table 2 Analysis of SRM materials by direct ICP-AES methods Element Concentration Certified Found Zn P g g-' 46.1 f 2.2 45.0 f 0.7 c u Pi2 g-' 0.7 f 0.1 ND* Fe Pg g-' 1.78 _+ 0.02 1.72 f 0.12 Ca YO m/m 1.30+0.05 1.29 f 0.03 Mg YO m/m 0.120 f 0.003 0.1 18 k 0.005 P % m/m 1.06 f 0.02 1.05 k 0.03 * ND =Not detected.different lactation stages (on the third seventh and twenty- first days postpartum). Samples were collected in the early morning into previously prepared acid-washed 20 ml con- tainers from INDEMEDICA.Ethoxy nonylphenol aqueous solution (25% m/v) was added as an emulsifying agent (final concentration 0.30% m/v). The samples were stored at - 15 "C prior to analysis. Procedure The maternal emulsified milk samples were diluted 10-fold with nitric acid to 1% v/v (final concentration 0.030% m/v ethoxy nonylphenol and 1 % v/v nitric acid). Aqueous standard solutions for calibration purposes were prepared using identical acid and ethoxy nonylphenol concentrations as for the samples. The milk reference material (SRM 1549 Non-Fat Milk Powder) was reconstituted by suspending 2 g of the dried material in 25 ml of 0.03% ethoxy nonylphenol solution and then homogenized by ultrasonic agitation. The iron concen- tration was determined in this solution.For the determination of Zn however this solution was diluted 8-fold and 40-fold for Ca Mg and P. No agglomeration of solid particles was observed under these conditions. Results and Discussion Optimization of Analytical Conditions Under ideal conditions freezing is the best long-term preser- vation method. However during freezing some irreversible processes may take place because of denaturation of proteins and re-distribution of elements due to rupture of cell walls by ice crystals. An oily layer on the top of the liquid milk is produced which cannot be re-dispersed by mechanical pro- cedures such as stirring shaking or ultrasonic treatment. Ethoxy nonylphenol is a non-ionic surfactant that is com- pletely soluble in water.It is not affected by the presence of Ca Mg and Fe salts.23 This surfactant has been successfully used by Murillo et al. to emulsify lubricating oils2' and crude samples for the determination of metals by ICP-AES. The concentration of surfactant was evaluated to obtain the most stable milk emulsion. For this the chosen criterion was the re-dispersion of the oil layer into the whole milk sample so as to produce a homogeneous and stable emulsion. The homogenization of the emulsion was visually evaluated. The ethoxy nonylphenol concentration was varied from 0 to 0.53 % m/v. The optimum emulsifier concentration selected was 0.3%; concentration greater than 0.3% m/v did not improve the emulsion stability. At this optimum concentration stable emulsions are formed which can be kept under refriger- ation for a long time.Reconstruction of such samples is carried out by manual shaking. The emulsified sample was diluted 10-fold with nitric acid prior to nebulization (final concen- tration 0.030% m/v ethoxy nonylphenol and 1% v/v nitric acid). Reproducibility of the analyte signal was checked using samples that had not been frozen. In order to obtain representative portions of the sample it was better to add the emulsifier before sampling. By adding ethoxy nonylphenol to the samples several limitations associ- ated with milk fat separation during sample storage under frozen conditions could be avoided. Also ethoxy nonyphenol enabled direct sample introduction without blockage of the Table 3 Concentration ranges of elements in maternal milk collected 3 7 and 21 days postpartum ~ Concentration/pg ml- ' Day Ca Mg Ft c u Zn P 3 107-388 20.8-47.6 0.22-0.74 0.30-0.78 2.6-1 1.6 35-169 136-278 7 192-393 20.2-39 0.22-0.57 0.32-0.72 2.0-5.75 21 144-330 19.8-32.5 0.22- 0.5 5 0.32-0.65 1.01-4.3 105-234JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 207 nebulizer.The introduction of emulsions has improved trans- port and atomization efficiency of milk sample in ICP-AES. This may be associated with the reduction of the average size of the droplets.24 The technique also overcomes the problems associated with viscosity and surface tension differences between real samples and aqueous calibration solution^.^ Detection Limit The detection limits were calculated based on the Kayser definition25 with a signal-to-noise ratio of 3.The detection limits obtained were 0.05 0.0089 0.0079 0.023 0.023 and 0.055 pg ml-' for Zn Cu Fe Ca Mg and P respectively. Analysis of an SRM by the Proposed Method To assess the accuracy of the proposed method the NIST SRM 1549 Non-Fat Milk powder was analysed. The results obtained are presented in Table 2. The significance testz6 indicated no significant difference at the 95% confidence level between the observed and certified values for the elements determined. Analysis of Maternal Milk Samples The proposed method was applied to the determination of Ca Cu Fe Mg P and Zn in 50 maternal milk samples. The concentration ranges found are presented in Table 3. The precision was found to be satisfactory (between 0.3 and 2%).It can be seen that element concentration is higher in the early period of lactation namely colostrum (third day) and in transitory milk (seventh day postpartum). Similar results have been found by other worker^.^^.^^ Conclusions The method described enables the rapid and direct determi- nation of Ca Mg Fe Cu Zn and P in whole maternal milk and similar materials by ICP-AES. The addition of ethoxy nonylphenol to milk samples bypasses several limitations associated with milk fat separation during storage of frozen samples. This surfactant also facilitates direct sample introduction without nebulizer blockage. Aqueous standard solutions prepared under identical acid and surfactant concentrations as the samples were adequate for calibration purposes and complete recovery was obtained.For real samples reproducibility of the proposed method (as %RSD) varied from 0.3 to 2%. Good accuracy was obtained with no internal standard. This work was supported in part by Consejo de Desarrollo Cientifico y Humanistic0 of Universidad Central de Venezuela (Research Grant 03.12.2136.89). The authors gratefully acknowledge the financial support that has made this work possible. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 References Barnes L. A. Mauer A. M. Anderson A. S. Dallman P. R. Forbes E. B. Nichols B. L. Roy C. Smith N. J. Walker W. A. and Winick M. J. Pediatr. (St. Louis) 1978 2 591. Emmett S. E. J. Anal. At. Spectrom. 1988 3 1145. Feely R. M. Eitenmiller R. R. Jones J. B. and Barnhart H. J. Pediatr. Gastroenterol.Nutr. 1983 2 262. Feely R. M. Eitenmiller R. R. Jones J. B. and Barnhart H. J. Clin. Nutr. 1983 37 443. Coni E. Stacchini A. Caroli S. and Falconieri P. J. Anal. At. Spectrom. 1990 5 581. Suzuki K. T. Tamagawa H. Hirano S. Kobayashi E. Takahashi K. and Shimojo N. Biol. Trace Elem. Res. 1991 28 109. Li J. Yoshinaga J. Suzuki T. Abe M. and Morita M. J. Nutr. Sci. Vitaminol. 1990 36 65. Krushevska A. Barnes R. M. Amarasiriwaradena C. J. Foner H. and Martines L. J. Anal. At. Spectrom. 1992 7 851. Durrant S. F. and Ward N. I. J. Micronutrient Anal. 1989,5 111. Kodama M. and Miyagawa S. Anal. Chem. 1980 52 2358. Yan Z-y. and Zhang W. J. Anal. At. Spectrom. 1989 4 797. Mora J. Canals A. and Hernandis V. J. Anal. At. Spectrom. 1991 6 139. Ruiz A. I. Canals A.and Hernandis V. J. Anal. At. Spectrom. 1993 8 109. Bertagnolli J. A. Neylan D. and Hammargren D. D. At. Spectrosc. 1993 14 1. Berenguer V. and Hernhndez J. Quim. Anal. 1977 31 81. Berenguer V. Guinon J. L. and De la Guardia M. Anal. Chem. 1979 294,416. Polo-Diez L. Hernandez-Mendez J. and Pedraz-Penalva F. Analyst 1980 105 37. De la Guardia M. Salvador A. and Berenguer V. Analusis 1980 8 488. Lord C. Anal. Chem. 1991 63 1594. Borszeki J. Knapp G. Halmos P. and Bartha L. Mikrochim Acta 1992 108 157. Murillo M. Gonzalez A. Ramirez A. and Guillkn N. At. Spectrosc. in the press. Murillo M. and Chirinos J. paper presented at the XXVIII Colloquium Spectroscopicurn Internationale (CSI) York UK June 29-July 4 1993. Martinez A. A. Introduccidn a la Quimica de SuperJcies y Coloides Alhambra S.A. Madrid 1977. Becher P. Emulsions Theory and Practice Reinhold New York 2nd edn. 1965 ch. 6 and 7. Liteanu C. and Rica I. Statistical Theory and Methodology of Trace Analysis Wiley New York 1980 p. 255. Miller J. C. and Miller J. N. Statisticsfor Analytical Chemistry Wiley New York 1985 pp. 52 and 82. Casey C. E. Hambidge K. M. and Neville M. Am. J. Clin. Nutr. 1985 41 1193. Casey C. E. Neville M. and Hambidge K. M. Am. J. Clin. Nutr. 1989 49 773. Paper 3/03888B Received July 6 1993 Accepted December 14 1993