Analytica Chimica Acta 412 (2000) 121130

Rapid determination of selenium, lead and cadmium in baby food

samples using electrothermal atomic absorption spectrometry and slurry
atomization
P. Vias, M. Pardo-Martnez, M. Hernndez-Crdoba
Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, E-30071 Murcia, Spain
Received 8 November 1999; received in revised form 22 December 1999; accepted 6 January 2000

Abstract
Procedures for the electrothermal atomic absorption spectrometric determination of selenium, cadmium and lead in different
types of baby foods using slurried samples are described. Suspensions prepared in a medium containing 0.1% (w/v) Triton
X-100, 30% (v/v) concentrated hydrogen peroxide, 1% (v/v) concentrated nitric acid and a matrix modifier (0.5% (w/v) nickel
for selenium, 0.2% (w/v) nickel plus 1% (w/v) ammonium dihydrogenphosphate for cadmium and 1% (w/v) ammonium
dihydrogenphosphate for lead) were introduced directly into the furnace. The graphite furnace conditions were optimized for
each element. Denterium background correction was used. Calibration with aqueous standard solutions was used for selenium
and lead determinations, while the standard additions method was used for cadmium determination. The 3 detection limits
were 5.2, 3.4 and 0.4 ng g1 for selenium, lead and cadmium, respectively. The reliability of the procedures was established by
comparing the results obtained with those found for five fish-based baby foods using a previous microwave-oven mineralization
stage and by analyzing six biological certified reference materials. The lead concentration was below the detection limit in
all the baby foods tested. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Electrothermal atomic absorption spectrometry; Slurry sampling; Selenium; Lead; Cadmium; Baby foods

1. Introduction
Multi-element surveys of baby foods [1] have been
published because of growing interest in trace element concentrations in infant foods and the need to
establish limits for infant exposure to such elements
from the diet. Lead and cadmium are toxic elements,
and the European Commission has proposed a regulation which sets maximum limits for these metals in
Corresponding author. Tel.: +34-968-367406;
fax: +34-968-364148.
E-mail address: hcordoba@fcu.um.es (M. Hernandez-Cordoba)

certain foods [2]. The most important sources of lead

exposure are industrial emission, soils, car exhaust
gases and contaminated food. Vegetables with a relatively large leaf area, such as spinach and cabbage
can contain high levels when grown near lead sources.
Cadmium ions are easily absorbed by vegetables and,
in animal-based food, are principally distributed in
the liver and kidneys. The highest cadmium concentrations are found in rice, wheat, oyster, mussels and
the kidney cortex of animals [3]. Selenium is an interesting trace element because it has an important
antioxidant function, but if intake is excessive harmful
effects appear, the difference between the necessary

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 5 8 - 3

122

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

intake and that producing toxicity being small. It

has been pointed out that concentrations in the range
28 g g1 in foods are harmful [3]. The selenium
content of foods can vary considerably. Some levels
quoted as typical are <0.01 g g1 for fruits and
vegetables, 0.010.7 g g1 for cereals and ca. 0.03
and 0.5 g g1 for meat and seafood, respectively
[3].
Owing to the toxic nature of lead and cadmium and
to the essentiality/toxicity dichotomy shown by selenium, there are numerous procedures for their determination in biological samples and, especially, in foods.
However, the results are sometimes incongruous as
a consequence of inaccuracy due to sampling errors,
contamination, losses during handling, pretreatment,
decomposition and other procedural steps. Inaccurate
results for the determination of selenium particularly
can be attributed to the decomposition or digestion
stages involved, because the chemical action of the
reagents used in such steps, the resistance of selenium
compounds to oxidation and the volatility of the selenium species present or formed all play important
roles. It follows that methods for determining trace elements should involve minimal sample handling and
have detection limits sufficiently low to permit the
determination of metal and selenium burdens easily
and reliably. Considering these requirements, electrothermal atomic absorption spectrometry (ETAAS)
would seem a good choice because it allows solid and
semi-solid samples to be analyzed with minimal manipulation.
The introduction of solid samples in ETAAS has
practical advantages over time-consuming conventional procedures based on total dissolution of the
samples [4]. Several ETAAS procedures using the
slurry-based approach have been reported for selenium [513] and cadmium and lead [5,7,1437]
determination in foods. However, in such cases, the
direct introduction of suspensions can pose practical
problems because of high background values or the
build-up of carbonaceous residues, which results in
non-reproducible sample deposition, variable rates
of atomization and even partial obstruction of the
light beam. Although these difficulties can be overcome by including an air-ashing step in the heating
cycle [14,25], this is not a viable possibility for all
commercially available electrothermal atomizers and
involves a risk of decreasing the useful lifetime of the

expensive pyrolytic furnace material. Recently, it has

been demonstrated that the addition of both hydrogen
peroxide and nitric acid to samples rich in organic
carbon can alleviate these problems. An additional
advantage is that no damage is done to the graphite
tube atomizer [3133,37,38].
In this paper we report the results obtained
when this hydogen peroxide-based methodology was
applied to determining selenium, lead and cadmium
in baby foods, which were directly introduced into
the atomizer as slurries. All the baby foods analyzed
are in the form of pure destined for infants. The
reliability of the procedure was checked by analyzing
several standard reference materials and by dissolving the samples using a closed system. Our results
indicate that this easy-to-perform approach is both
useful and practicable for these frequently required
determinations.

2. Experimental
2.1. Instrumentation
A PerkinElmer model 1100B atomic absorption
spectrometer equipped with deuterium-arc background correction and an HGA-400 (PerkinElmer)
graphite furnace atomizer were used. Pyrolytic
graphite platforms (Part number B012-1092) inserted
into the pyrolytically coated graphite tubes were obtained from PerkinElmer. Measurements were performed at 283.3 and 228.8 nm for lead and cadmium,
respectively, and hollow cathode lamps were operated
at 7 mA. For selenium, measurements were performed
using a bandwidth of 2.0 nm at 196.0 nm using an
electrodeless discharge lamp operated at 300 mA
from an external power supply (PerkinElmer System
2). For comparison purposes, some measurements
were also made with an ATI-Unicam (Cambridge,
UK) 939QZ atomic absorption spectrometer equipped
with a GF90 electrothermal atomizer. Pyrolytic platforms (reference 9423 393 95191) were obtained
from ATI-Unicam. This instrument is equipped with
both a deuterium-arc based corrector and a Zeeman
correction device, which facilitates comparison between both correction modes. Argon was used as the
inert gas, the flow rate being 300 ml min1 during all

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

the stages except for atomization, when the flow was

stopped. Background-corrected integrated absorbance
was used in all cases as the analytical signal.
Manual homogenization glass vessels (Potter,
10 ml) equipped with PTFE plungers were used. A
Branson ultrasonic bath of 14 W constant power was
also used.
To decrease the risk of cadmium contamination,
the use of glassware was reduced to a minimum and
plastic (polypropylene) vessels of the type commonly
used to collect clinical samples were used for preparing and storing the solutions or suspensions. Pipette
tips were also of polypropylene. All the glassware and
plasticware was nitric acid-washed and rinsed with ultrapure water.
Mineralization of the samples for comparison purposes was carried out in closed PTFE cups using a
MLS-1200 MEGA microwave oven (Milestone) and
a MDR-1000/6 Rotor (Radiometer).
2.2. Reagents
High quality water, obtained using a Milli-Q system (Millipore), was used exclusively. Selenium, lead
and cadmium standard solutions (1000 g ml1 ) were
obtained from Panreac (Spain) and diluted as necessary to obtain working standards. Concentrated (65%,
w/v) nitric acid (Merck), 30% (w/v) hydrogen peroxide (Fluka), Triton X-100 (Merck), 30% (w/v) aqueous
emulsion of antifoam A concentrate (100% active silicone polymer, Sigma), ammonium dihydrogenphosphate (Fluka) and nickel nitrate hexahydrate (Fluka)
were also used.
2.3. Reference materials and samples
To assess the reliability of the procedures, six reference materials were used. Total diet (HDP) sample was
supplied by the Agricultural Research Centre of Finland (ARC/CL), dogfish muscle and liver (DORM-2)
was obtained from the National Research Council of
Canada (NRCC), whereas oyster tissue (SRM 1566a),
bovine liver (SRM 1577b), rice flour (SRM 1568a) and
citrus leaves (SRM 1572) were supplied by the National Institute of Standards and Technology (NIST),
USA. Samples of pured baby food were purchased
from several local supermarkets.

123

2.4. Procedures
The baby food samples were shaken manually
before analysis. The suspensions were prepared by
weighing the samples (the weight depended on the
analyte and its concentration in the sample) directly
into a Potter maceration tube and diluted with 5 ml
of a solution containing 0.1 (w/v) Triton X-100, 1%
(v/v) nitric acid, 30% (v/v) hydrogen peroxide, the
corresponding matrix modifier and one drop of silicone antifoam. The modifier solution was 1% (w/v)
ammonium dihydrogenphosphate for lead atomization, 0.2% (w/v) nickel(II) plus 1% (w/v) ammonium
dihydrogenphosphate for cadmium and 0.5% (w/v)
nickel(II) for selenium atomization. The slurries were
homogenized by repeated movements of the plunger.
It was verified that about 20 slow movements of the
plunger, which took about 5 min, were sufficient to
obtain a quasi-stable suspension. The suspensions
were also sonicated for a few minutes to ensure the
absence of lumps, and sampled while they were being
continuously stirred with a magnetic stirrer. Aliquots
of 20 l were injected into the furnace. The heating
programme given in Table 1 (where the quoted temperatures are the values set on the HGA-400 power
supply) was run and the background-corrected peak
areas due to the analyte were obtained. Calibration

Table 1
Furnace heating programmes
Step

Parameter

Dry

( C)

T
Ramp (s)
Hold (s)

180
10
60

180
10
60

180
10
60

Ash

T ( C)
Ramp (s)
Hold (s)

1300
1
30

1000
1
30

1000
1
30

Cool

T ( C)
Ramp (s)
Hold (s)

20
1
15

NO

NO

Atomizena

T ( C)
Ramp (s)
Hold (s)

2200
0
4

1300
0
5

1700
0
3

Clean

T ( C)
Ramp (s)
Hold (s)

2650
1
3

2650
1
3

2650
1
3

Selenium

Cadmium

Flow of argon stopped during atomization.

Lead

124

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

was performed using aqueous standards for lead and

selenium and by the standard additions method for
cadmium. Certified reference samples were treated in
the same way.
To confirm the reliability of the suspension-based
procedures, the samples were mineralized by microwave oven treatment. For this, fractions (12 g)
were weighed into PTFE cups and 3 ml of concentrated nitric acid and 0.5 ml of concentrated hydrogen
peroxide were added. The programme used in the
microwave oven was that recommended by the supplier in the case of food samples with a water content
higher than 60%. The programme consisted of 1 min
at 250 W, 1 min with no power applied, 6 min at
250 W, 6 min at 400 W, 4.5 min at 600 W and 1.5 min
at 250 W. After this treatment, the solutions were
maintained in the closed cups for 10 min before being diluted with water to 10 ml in volumetric flasks.
These solutions were also analyzed by ETAAS using
the same experimental conditions indicated for the
suspensions.
3. Results and discussion
As indicated elsewhere [38], when suspensions are
prepared from baby food samples, agglomerates form

which hinder reproducible sampling by micropipette.

To overcome this, 0.1% Triton X-100 was incorporated in the suspension medium. Foaming was prevented by adding one drop of a silicone antifoaming
agent.
3.1. Furnace heating programmes
Preliminary experiments were devoted to checking
whether the heating cycles could be simplified by using fast-programme methodology [39]. However, even
when high proportions of hydrogen peroxide were
incorporated in the suspension medium, excessively
high background values caused overcompensation effects, and so a conventional heating programme was
optimized. As the ashing stage could not be omitted,
all the subsequent experiments were carried out in the
presence of phosphate or a nickel salt as modifier. The
optimal drying temperature and the holding time were
studied so that the samples were completely dry before ashing. The optimal values were 180 C with a
10 s ramp to avoid sputtering and a 60 s holding time.
Fig. 1A shows the variation of both the analytical signal and the background obtained from a baby food
suspension when the ashing temperature was varied.
For selenium, the ashing temperature was varied in the

Fig. 1. Influence of the ashing (A) and atomization (B) temperatures for a 15% (w/v) baby food suspension containing 0.1% (w/v) Triton
X-100, 30% (v/v) hydrogen peroxide, 1% (v/v) nitric acid and the matrix modifier (see Experimental). (A) Effect of the ashing temperature
on the analytical signals (solid lines) and the backgrounds (dotted lines). (B) Effect of the atomization temperature on the analytical signals
(solid lines) and the atomization time required (points with drop lines).

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

9001600 C range, the analytical signal obtained in

the atomization stage being constant up to 1300 C this
temperature also providing a low background signal.
For cadmium, and lead, the signals obtained during
the atomization stage decreased when temperatures in
excess of 1100 and 1200 C respectively, were used for
the ashing stage. Taking into account the low background signals and to avoid the risk of premature
analyte losses a temperature of 1000 C was finally
selected for these analytes.
Atomization temperatures were selected bearing in
mind both sensitivity and the need for well-shaped
atomization profiles in isothermal conditions. Fig. 1B
shows the variation in the analytical signals and the
time necessary to obtain a fully developed atomization
profile. The analytical signal of selenium decreased
above 2200 C and so this temperature, which required
an atomization time of 3 s, was selected. In addition,
as the peak appeared in non-isothermal conditions, a
cool step was included in the heating programme in
order to delay the signal. This step consisted of a 1 s
ramp at 20 C with a 15 s holding time. In the case of
lead, 1700 C was selected as the optimal atomization
temperature. The presence of chloride in the samples
made it necessary to use a lower temperature (1300 C)
for cadmium to avoid losses during atomization. A
cleaning stage was also included in the heating cycle.
Each of these experiments was repeated using aqueous solutions of the analytes and suspensions prepared
from two certified reference materials (total diet and
bovine liver). Similar results were obtained for the
three analytes. For comparison, a number of experiments were also performed using Zeeman correction
instead of deuterium-based correction. However, no
practical advantages were noted.
3.2. Optimization of the concentrations of the
chemical agents
All the above experiments were carried out in the
presence of chemical modifiers, their concentrations
being optimized once the heating programmes had
been selected. The concentration of ammonium dihydrogenphosphate used for lead determination was
varied in the 0.13.0% (w/v) range, 1% being considered as optimal since higher percentages did not
produce additional delays of the atomization profiles
but increased background signals. For selenium, the

125

optimal concentration of nickel salt was studied for

an aqueous solution containing 50 ng ml1 selenium,
a 10% (w/v) baby food suspension and a 3% (w/v)
total diet reference material suspension. The results
are summarized in Fig. 2. As can be observed, 0.2%
nickel was sufficient to stabilize aqueous selenium,
whereas higher concentrations were required to stabilize selenium in the suspensions. However, the time
required to achieve complete atomization of aqueous
selenium with high nickel percentages was very long.
Consequently, 0.5% nickel was chosen as a compromise between high atomization efficiency and acceptable atomization time. For cadmium, the simultaneous presence of phosphate and nickel was necessary to
avoid volatilization losses. Fig. 2 shows that cadmium
stabilization could be achieved using 0.2% nickel and
1% phosphate.
The presence of hydrogen peroxide had a beneficial
effect on the elimination of carbonaceous residues in
the atomizer [3133,37,38]. When the concentration
of this chemical was varied in the 050% (v/v) range,
no significant changes in the analytical signals or in
the background were noted, although the relative standard deviation (RSD) improved. This benefical effect
was especially noticeable for cadmium. Thus, in the
absence of the oxidant, the RSD for 10 consecutive injections of a 20% suspension of baby food was 8.8%,
while with the incorporation of 30% (v/v) concentrated hydrogen peroxide to the suspension medium
decreased this value to 3.6%.
To ascertain whether the presence of the oxidant affected the slopes of the calibration graphs, the slopes
for aqueous calibration and standard additions were
compared for the atomization of selenium from a 20%
baby food suspension. As can be seen in Fig. 3, the
slopes obtained in the absence and presence of different amounts of hydrogen peroxide did not differ significantly. A similar finding was obtained for lead and
cadmium. Consequently, a 30% concentrated hydrogen peroxide medium was selected.
The addition of nitric acid to the suspension was
necessary to obtain good sensitivity. The concentration
of the acid was varied between 0.2 and 3% (v/v) for
aqueous standards and baby food suspensions. A 1%
value was selected since a higher acid concentration
provided no advantages but had the disadvantage of
shortening the useful lifetime of the graphite atomizer
tube.

126

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

Fig. 2. Effect of the variation of the modifier concentration on the atomization of selenium and cadmium. The samples were aqueous
solutions containing selenium (50 ng ml1 ) or cadmium (1 ng ml1 ); 10% (w/v) baby food suspension and 3% (w/v) total diet suspension.
Solutions contained 0.1% (w/v) Triton X-100, 30% (v/v) hydrogen peroxide and 1% (v/v) nitric acid. For cadmium, the curves with filled
symbols represent the variation of the signal with nickel concentration (in the presence of 1% (w/v) dihydrogenphosphate) and the curves
with hollow symbols represent the variation with the phosphate concentration (in the presence of 0.2% (w/v) nickel).

When the concentrations of all the chemicals had

been optimized, the heating programmes were reconsidered and no changes over the values previously

selected were considered necessary. Fig. 4 shows the

selenium and cadmium atomization profiles obtained
under the final experimental conditions.
3.3. Influence of the suspension concentration and
study of matrix effect

Fig. 3. Influence of the hydrogen peroxide concentration on the

atomization of selenium from an aqueous standard and a baby
food suspension. The solutions contained 0.1% (w/v) Triton X-100,
0.5% (w/v) nickel and 1% (v/v) nitric acid.

The influence of the suspension concentration on

the analytical signals for selenium and cadmium is
shown in Fig. 5. A response that increased linearly
with increasing concentration was obtained up to a
20% (w/v) concentration. Higher proportions of sample in the suspension are not recommended because
of the difficulty in reproducible sampling. To check
possible interferences due to the matrix, the slopes of
aqueous calibration and standard additions calibration
graphs were compared. The bar graphs also included
in Fig. 5 show the results obtained. Each graph was
constructed from four points and each point represents
the mean of three replicate observations. As can be
seen, the slopes for selenium standards and standard
additions for different concentrations of the baby food
up to 20% (w/v) were similar. Minor differences can

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

127

Fig. 4. Atomization profiles for selenium (A) and cadmium (B) from a 10% (w/v) baby food suspension. Solutions contained 0.1% (w/v)
Triton X-100, 30% (v/v) hydrogen peroxide, 1% (v/v) nitric acid and the modifier (0.5% (w/v) nickel for selenium and 0.2% (w/v) nickel
plus 1% (w/v) ammonium dihydrogenphosphate for cadmium). The broken lines show the background signals.

be attributed both to matrix effects and difficulties

in sampling of more concentrated suspensions. For
cadmium, the higher the percentage of sample in the
suspension, the greater the slope. In addition, the time
necessary to obtain fully developed atomization profiles changed. Additional experiments suggested that
these effects were caused by the chloride present in
the samples, since this modifies the atomization characteristics of cadmium. This fact hinders direct calibration against aqueous standard solutions and makes
it necessary to use the standard additions method for
the determination. Data for lead are not given because

no lead was found in any of the samples analyzed.

To corroborate the observations presented above,
the study was extended to different baby food samples; Table 2 shows the results obtained. The slopes
of the best-fit regression lines for standard additions
to the different suspensions varied by 6.1% (RSD) for
selenium, 11.5% for cadmium and 5.1% for lead. To
summarize, for selenium and lead, aqueous calibration
can be carried out for suspensions with concentrations
of sample not exceeding 20% (w/v), while standard
additions calibration must be used for the determination of cadmium.

Table 2
Slopes of standard additions calibration graphs for different baby food samples
Sample (11%, w/v)

Aqueous standards
Sole with white sause
Hake with rice
Angler fish with vegetables
Sole with vegetables
Sole with potatoes
a

Meanstandard deviation (n=4).

Slopea (s ml ng1 )
Selenium

Cadmium

Lead

0.0024400.00005
0.0022890.00005
0.0025930.00002
0.0025500.00001
0.0022180.00008
0.0023400.00004

0.083030.0023
0.066310.0049
0.062260.0013
0.063990.0029
0.062910.0044
0.068810.0006

0.0032970.00004
0.0038380.00012
0.0035770.00015
0.0032970.00012
0.0035410.00017
0.0034570.00024

128

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

Fig. 5. Influence of the suspension concentration using a hake baby food sample for selenium and a sole baby food for cadmium. The bar
graphs present the slopes of aqueous calibration and standard additions calibration graphs.

3.4. Repeatability and calibration graphs

Table 3 shows the characteristics of the calibration
graphs for lead, cadmium and selenium. The detection
limits were calculated for 10 successive injections of

the blank and using the 3 criterion for the maximum

recommended suspension concentration (20%, w/v).
The repeatability was calculated using the RSD for 10
successive injections of a 20% baby food suspension.
3.5. Results and accuracy

Table 3
Analytical characteristics of the calibration graphs
Parameter

Selenium

Cadmium

Lead

Linearity (ng ml1 )

Characteristic mass (pg)
Detection limit (ng g1 )
RSD (%)a

0250
20
5.2
3.9

05
1
0.4
2.5

060
17
3.4
2.7

Ten determinations of 20% baby food suspension.

Table 4 summarizes the results obtained for the different baby food samples analyzed using the proposed
procedure as well as a procedure based on a mineralization step employing a closed system. Students
t-test revealed no difference between the results obtained using either procedure (level of significance
0.05). Cadmium was found at very low concentrations

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

129

Table 4
Selenium and cadmium contents found in baby food samples
Sample

Sole with vegetables

Angler fish with vegetables
Sole with potatoes
Sole with white sauce
Hake with rice
a

Selenium (ng g1 )

Cadmium (ng g1 )

Suspensiona

Mineralizationa

Suspensiona

Mineralizationa

36.26.5
36.02.9
21.51.3
51.61.9
72.85.3

32.33.3
32.81.7
20.83.5
48.82.5
70.52.7

7.160.075
8.080.20
9.920.067
6.720.18
5.020.25

6.600.55
7.380.13
10.320.34
5.960.10
5.000.11

Meanstandard deviation, n=4.

Table 5
Results for the certified reference materials
Sample

Total diet (HDP)

Dogfish muscle and liver (DORM-2)
Oyster tissue (SRM 1566a)
Bovine liver (SRM 1577b)
Rice flour (SRM 1568a)
Citrus leaves (SRM 1572)
a

Selenium (ng g1 )

Cadmium (ng g1 )

Lead (ng g1 )

Suspensiona

Certified

Suspensiona

Certified

Suspensiona

Certified

18718
1360110
2290100
69050
40235

18117
140090
2210240
73060
38040
Not certified

213
438
4150380
50030
222
3010

47.24
61.85
3638
13011
13.12
12636217

438
657
37114
1294
<10 (Not certified)
133002400

19.40.6
37.62.5
4475126
48730
20.70.6
37.75.8

Meanstandard deviation, n=4.

while lead was not detected in any sample. The reliability of the method was further corroborated by using several certified reference materials. Application
of both the Student t-test and the MannWhitney rank
sum test concluded that there was good agreement
(level of significance 0.05) between the reference values and the results obtained (Table 5).

4. Conclusion
Selenium, cadmium and lead can be determined reliably in different types of baby food by direct injection into the atomizer of suspensions prepared from the
samples. No previous sample mineralization is necessary, the experimental procedure being simple, which
reduces the risks of contamination and loss through
volatilization. The addition of both hydrogen peroxide
and nitric acid to the samples considerably reduces the
deposition of carbonaceous residues, which improves
reproducibility. Common deuterium background correction is appropriate, which allows the procedures to
be applied in most laboratories with no need for more

sophisticated correction systems. The procedures are

rapid and accurate, and can be considered useful for
the routine determination of selenium, cadmium and
lead in the quality control of baby food samples.

Acknowledgements
The authors are grateful to the Spanish DGICYT
(Project PB96-1100) and Comunidad Autnoma de la
Regin de Murcia (Fundacin Sneca, CARM, Project
PB/7FS/97) for financial support. M. Pardo-Martnez
also acknowledges a fellowship from Consejera de
Cultura, CARM.
References
[1] Ministry of Agriculture, Fisheries and Food, Multielemental
analysis of infant foods, Food Surveillance Paper No.62,
London, May 1995. http://www.maff.gov.uk/food/infsheet/
1995/no62.
[2] European Commission, Draft Commission Regulation Setting
Maximum Limits for Certain Contaminants in Foodstuffs,
Doc. III/5125/95 Rev. 3, 1997.

130

P. Vias et al. / Analytica Chimica Acta 412 (2000) 121130

[3] H.G. Seiler, A. Sigel, H. Sigel, Handbook on Metals in

Clinical and Analytical Chemistry, Marcel Dekker, New York,
1994.
[4] C. Bendicho, M.T.C. de Loos-Vollebregt, J. Anal. At.
Spectrom. 6 (1991) 353.
[5] D. Wagley, G. Schmiedel, E. Mainka, H.J. Ache, At.
Spectrosc. 10 (1989) 106.
[6] C. Bendicho, A. Sancho, At. Spectrosc. 14 (1993) 187.
[7] C. Cabrera, M.L. Lorenzo, M.C. Lpez, J. Assoc. Off. Anal.
Chem. Int. 78 (1995) 1061.
[8] Y. Tan, J.S. Blais, W.D. Marshall, Analyst 121 (1996) 1419.
[9] I. Lpez-Garca, P. Vias, N. Campillo, M. HernndezCrdoba, J. Agric. Food Chem. 44 (1996) 836.
[10] Y. Tan, W.D. Marshall, Analyst 122 (1997) 13.
[11] T. Kubota, T. Okutani, Anal. Chim. Acta 351 (1997) 319.
[12] G.S.B. Januzzi, F.J. Krug, M.A.Z. Arruda, J. Anal. At.
Spectrom. 12 (1997) 375.
[13] A.J. Aller, L.C. Robles, J. Anal. At. Spectrom. 13 (1998) 469.
[14] S.C. Stephen, D. Littlejohn, J.M. Ottaway, Analyst 110 (1985)
1147.
[15] K.O. Olayinka, S.J. Haswell, R. Grzeskowiak, J. Anal. At.
Spectrom. 1 (1986) 297.
[16] M. Hoenig, P. Van Hoeyweghen, Anal. Chem. 58 (1986) 2614.
[17] F. Fagioli, S. Landi, C. Locatelli, C. Bighi, At. Spectrosc. 7
(1986) 49.
[18] L. Ebdon, A. Lechotycki, Microchem. J. 34 (1986) 340.
[19] S.C. Stephen, J.M. Ottaway, D. Littlejohn, Fresenius Z. Anal.
Chem. 328 (1987) 346.
[20] L. Ebdon, A. Lechotycki, Microchem. J. 36 (1987) 207.
[21] N.J. Miller-Ihli, J. Anal. At. Spectrom. 3 (1988) 73.
[22] S. Lynch, D. Littlejohn, J. Anal. At. Spectrom. 4 (1989) 157.

[23] Z. Yu, C. Vandecasteele, B. Desmet, R. Dams, Mikrochim.

Acta 1 (1990) 41.
[24] S. Lynch, D. Littlejohn, Talanta 37 (1990) 825.
[25] L. Ebdon, A.S. Fisher, H.G.M. Parry, A.A. Brown, J. Anal.
At. Spectrom. 5 (1990) 321.
[26] M. Hernndez-Crdoba, I. Lpez-Garca, Talanta 38 (1991)
1247.
[27] F. Dolinsek, J. Stupar, V. Vrscaj, J. Anal. At. Spectrom. 6
(1991) 653.
[28] M. Hoenig, A. Cilissen, Spectrochim. Acta 48B (1993) 1003.
[29] R. Dobrowolski, J. Mierzwa, Fresenius J. Anal. Chem. 346
(1993) 1058.
[30] P. Bermejo-Barrera, M. Aboal-Somoza, R.M. Soto-Ferreiro,
R. Domnguez-Gonzlez, Analyst 118 (1993) 665.
[31] P. Vias, N. Campillo, I. Lpez-Garca, M. HernndezCrdoba, Analyst 119 (1994) 1119.
[32] P. Vias, N. Campillo, I. Lpez-Garca, M. HernndezCrdoba, Fresenius J. Anal. Chem. 349 (1994) 306.
[33] P. Vias, N. Campillo, I. Lpez-Garca, M. HernndezCrdoba, Food Chem. 50 (1994) 317.
[34] Y. Tan, W.D. Marshall, J.S. Blais, Analyst 121 (1996) 483.
[35] J. Shiowatana, A. Siripinyanond, At. Spectrosc. 17 (1996)
122.
[36] N.N. Meeravali, S.J. Kumar, J. Anal. At. Spectrom. 13 (1998)
647.
[37] P. Vias, M. Pardo-Martnez, N. Campillo, M. HernndezCrdoba, J. Assoc. Off. Anal. Chem. Int. 82 (1999)
368.
[38] P. Vias, M. Pardo-Martnez, M. Hernndez-Crdoba, J. Anal.
At. Spectrom. 14 (1999) 1215.
[39] D.J. Halls, J. Anal. At. Spectrom. 10 (1995) 169.