Elsevier

Food Chemistry

Volume 83, Issue 3, November 2003, Pages 457-462
Food Chemistry

Analytical, Nutritional and Chemical Methods
Direct determination of iron and selenium in bovine milk by graphite furnace atomic absorption spectrometry

https://doi.org/10.1016/S0308-8146(03)00224-3Get rights and content

Abstract

Milk is a complex sample containing high contents of organic compounds and its analysis generally involves digestion procedures that can be affected by losses and contamination. In the work here described it was developed a procedure for determination of Fe and Se in bovine milk. Samples were diluted using a mixture of water-soluble tertiary amines (10% v/v CFA-C). The tertiary amines had a favorable effect on the action of the autosampler and consequently on repeatability. Using this strategy the direct analysis of milk without any digestion procedure by graphite furnace atomic absorption spectrometry is feasible. Pyrolytic graphite tubes were used and all measurements were based on peak area of transient signals with background signal correction based on Zeeman effect. Palladium was used as chemical modifier for Se. The pyrolysis and atomization temperatures were 1300 and 2300 °C for Fe and 1500 and 2400 °C for Se, respectively. The procedure was applied for 13 bovine milk samples from different sources. The quantification was based on the standard additions method. The lifetime of the graphite tube was 300 and 250 heating cycles for Fe and Se, respectively.

Introduction

Milk is a close to ideal food essential for newborns due to its composition and availability. In addition to its macronutrients, i.e. protides, glucides, and lipids, milk also contains micronutrients, i.e. vitamins and elements, that are absolutely essential during the first months of a baby's life since it is the only source of nutrients. According to Coni, Alimonti, Bocca, La Torre, Pizzuti, and Caroli (1996), this is particularly true for micronutrients that are not stored by the fetus during its growth inside the uterus. These authors pointed out that essential elements could be divided into two groups: those with reserves usually sufficient to protect the baby from potential deficiencies during the first 4–6 months, and those which require immediate reintegration right from birth so as to reach the best rate of growth. Iron is a typical element of the first group and Se is representative of the second one. This clearly demonstrated the importance of the determination of these trace elements in milk. Human and cow's milk are relatively low in Fe ranging from 0.2 to 0.8 mg l−1, but it is highly bioavailable mainly in human milk (Casey, Smith, & Zhang, 1995). On the other hand, Se concentration in milk is directly affected by levels in the food chain and, hence, reflects the food habits and the geochemical environment. Selenium concentration in cow's milk can vary as many as 2–1270 μg l−1 depending on the availability of this element in the food and geographical area. Nowadays there is an increasing knowledge about the role of Se in physiological processes, however despite its clear beneficial effects the Se intake is declining over the years in Europe (Rayman, 2002).

The determination of trace inorganic constituents in milk is not a trivial task because of the complexity of the emulsion. According to Jensen (1995), bovine milk contains around 3.4% of proteins, 2.8% of casein, 3.7% of fat and 4.6% of lactose; additionally, the elements are present as different compounds that can affect both the sample preparation and the measurement strategies. For example, about one-third of the Fe in human milk is associated with the low molecular weight aqueous fraction, one-third with the milk fat, and of the remainder, about 10% is found with the casein fraction (Jensen, 1995).

Most proposed procedures involve a step of digestion to eliminate the organic matrix. Owing to the complexity of this matrix, the procedure can involve several steps and contamination can become a serious obstacle for obtaining accurate data. A critical study of different procedures, such as dry ashing, hot-plate procedure, high-pressure ashing, open and closed vessels microwave-assisted, that could be applied for the determination of Zn in milk was presented by Krushevska, Barnes, Amarasiriwaradena, Foner, and Martines (1992). Despite the usefulness of these procedures, some of them are time consuming and it is attractive to investigate dilute-and-shoot procedures that could be properly applied for trace analysis of milk.

Direct procedures for milk analysis were proposed using different spectroanalytical techniques. Each one presents some difficulties. Quináia and Nóbrega (2000) demonstrated that the presence of fat compounds affects the performance of the autosampler in graphite furnace atomic absorption spectrometry (GF-AAS) and the pneumatic nebulization in the inductively coupled plasmas coupled to optical emission spectrometry (ICP-OES—Coni, Stacchini, Caroli, & Falconieri, 1990) or to mass spectrometry (ICP-MS—Stürup & Büchert, 1996). The formation of carbon residues causes a gradual deterioration of the graphite tube in GF-AAS, affects excitation conditions in ICP-OES and isobaric interferences in ICP-MS. In GF-AAS milk samples were diluted with a solution containing ethanol, nitric acid and hydrogen peroxide to remove the organic matter during the heating cycle (Viñas, Campillo, López-Garcı́a, & Hernández-Córdoba, 1997). This procedure was successfully applied for the determination of Al, Cr, Mn, and Mo.

The addition of surfactant agents could also be employed to attenuate at least partially the undesirable effects. Wagley, Schmiedel, Mainka, and Ache (1989) have studied the direct determination of Se in milk introducing a suspension prepared in 3% v/v Triton X-100 and containing Rh+Mg as chemical modifier. More recently, this same surfactant was also used for determining nine elements, including Se, in different milk slurries by GF-AAS (Garcia, Lorenzo, Cabrera, Lopez, & Sanchez, 1999). Foster and Sumar (1995), in a review about Se determination in milk and infant formulae, emphasized that most AAS works involved hydride generation, that is critically dependent on a harsh sample decomposition treatment to convert all organoselenium forms to Se(IV).

The use of a chemical modifier for Se determination by GF-AAS is imperative due to its volatility, but the choice was not trivial in earlier works. In spite of the use of Ni as chemical modifier (Wagley et al., 1989), the main critical aspect that should be pointed out is that this modifier did not stabilize at the same degree different oxidation states and compounds of Se (Welz & Sperling, 1999). Despite of the mentioned 1800 °C pyrolysis temperature applied when the mixture Rh+Mg was used (Wagley et al., 1989), the use of Pd should be recommended owing to its general application in GF-AAS as an universal modifier and the reliability to prevent losses of all Se compounds up to a pyrolysis temperature of at least 1000 °C. The relatively high levels of total phosphate, i.e. around 150 mg l−1, caused spectral interferences at the 196.0 nm resonance line and the use of a Zeeman background corrector is strongly recommended.

There is also a lack of information in the literature related to the direct determination of Fe in milk by GF-AAS. Bermejo, Dominguez, and Bermejo (1997), demonstrated the feasibility of a high-performance nebulizer for introduction of milk suspension in a flame atomic absorption spectrometer (FAAS) and determination of Fe. We had observed in preliminary experiments that the use of a conventional concentric nebulizer for non-diluted milk samples was not possible due to gradual clogging of the nebulizer and when the milk sample was diluted to overcome this effect the sensitivity of FAAS was not enough.

In a previous work it was proposed a dilute-and-shoot procedure for direct analysis of milk by ICP-MS (Nóbrega, Gélinas, Krushevska, & Barnes, 1997). The milk samples were diluted in a mixture of water-soluble tertiary amines and the resulting suspensions were introduced by pneumatic nebulization using a concentric nebulizer. Despite the good results obtained, the accuracy attained for Al, Cr, Fe, and Se was not suitable probably due to the occurrence of isobaric interferences.

In the work here described a procedure was developed for dilution of milk samples with a mixture of water-soluble tertiary amines and the suspension was introduced into the graphite tube by autosampler action. Iron and Se were determined in bovine milk using conventional heating programmes without any modifier or with Pd modifier, respectively.

Section snippets

Apparatus

A Varian AA-800 atomic absorption spectrometer (Mulgrave, Victoria, Australia) was connected to a GTA-100 graphite furnace atomizer equipped with an automatic sampler, also from Varian. Pyrolytic graphite tubes (Part number 63-100011-00) were used. Background correction was based on the Zeeman effect with a transversal electromagnetic field. An ultra hollow cathode lamp and a hollow cathode lamp were employed for Se and Fe, respectively. Argon was used as purge gas at a flow-rate of 3.0 l min−1

Effect of dilution factor and the use of water-soluble tertiary amines

As already mentioned, the preliminary experiments focused on the possibility of milk introduction in the graphite tube without any dilution. This strategy was unsuitable due to poor repeatability caused by both carbon residues in graphite tube and fat residues in the autosampler capillary tube. The dilution of milk samples in water using different dilution factors was not effective for overcoming these effects. Taking into account these results, the effect of milk samples dilution with a 10%

Acknowledgments

The authors are grateful to CNPq-PADCT (Process 62.0585/94-3) and FAPESP (Processes 98/10814-3) by providing research funds. P.C.A would like to thank FAPESP by the fellowship (Process 02530-5/98). J.A.N. is thankful to CNPq by individual research support.

References (15)

There are more references available in the full text version of this article.

Cited by (73)

  • Spot test for fast determination of hydrogen peroxide as a milk adulterant by smartphone-based digital image colorimetry

    2020, Microchemical Journal
    Citation Excerpt :

    Furthermore, hypochlorite would react with hydrogen peroxide, forming chloride, oxygen, and water. Interference from iron(III) is not expected by considering the concentrations in bovine milk, which are typically lower than 1.0 mg L−1 [28] and the detection limit of the procedure (1.7 mg L−1 H2O2 would yield 5.6 mg L−1 Fe(III)). In this way, the typical iron(III) concentrations in milk would not yield an analytical response different from the blank.

View all citing articles on Scopus
View full text