Vous êtes sur la page 1sur 6

Microchemical Journal 79 (2005) 91 96


Determination of cadmium, lead, iron, nickel and chromium in

selected food matrices by plasma spectrometric techniques
Francesco Cubadda*, Andrea Raggi
National Centre for Food Quality and Risk Assessment, Istituto Superiore di Sanita`, Viale Regina Elena 299, 00161 Rome, Italy

Received 20 September 2004

Available online 5 November 2004


In the frame of a study aimed at investigating the transfer of metal contaminants through the food chain and the effects of food processing,
five elements, namely Cd, Pb, Fe, Ni and Cr, were accurately determined in (i) durum wheat grain and derived products, (ii) wheat-based
reference materials, (iii) drinking water, used both as an ingredient and for technological purposes in the investigated industrial process.
Microwave closed vessel digestion was selected as the dissolution technique for solid samples, whereas water samples were acidified with
ultrapure nitric acid and analysed directly. As several analytes had to be quantified at trace or ultratrace levels, inductively coupled plasma
mass spectrometry (ICP-MS) was resorted to for analytical determinations. Overall, this straightforward analytical approach enabled to detect
the often small changes in element concentration associated with the different technological steps of processing. Nevertheless, detection of
heavily interfered elements, especially Cr, as well as analyte quantification at ultratrace-level level in water, posed analytical challenges that
required suited analytical solutions.
Changes in the sample introduction system and complementary use of inductively coupled plasma atomic emission spectrometry (ICP-
AES) straightforwardly overcame the difficulties in determining the analytes in the selected food matrixes. The benefits of ultrasonic
nebulization in reducing the effects of problematic spectral interferences were demonstrated. Overall, a robust and high-throughput analytical
method was outlined.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Spectral interferences; Ultrasonic nebulization; Inductively coupled plasma mass spectrometry; Inductively coupled plasma atomic emission
spectrometry; Wheat; Water

1. Introduction In the frame of a research activity on metal contaminants

in the durum wheat food chain, a study on the effects of
The studies on the transfer of chemical contaminants processing on five selected elements, namely Cd, Cr, Fe, Ni
through the food chain provide useful information for the and Pb, was carried out by our group [1]. Cd and Pb are
development of surveillance programs aimed at ensuring the well-known toxic elements regulated by the European
safety of the food supply and minimising human exposure legislation on food contaminants [2], while Cr, Fe and Ni
to toxic agents. Knowledge of the routes enabling the were included in the study as markers of metal release from
carrying on of environmental contaminants such as cad- equipment during processing. The above-mentioned ele-
mium and lead in the food chain from staples to consumer is ments have been determined in cereal matrixes (durum
an essential component of an effective system for the wheat grain and derived products) and in the drinking water
prevention and control of food risks. For this purpose, the used both as an ingredient and for technological purposes in
identification of any possible source of food contamination the investigated industrial process.
during processing is also required. The need of quantifying several analytes at trace or ultra-
trace levels required a highly sensitive multielemental
* Corresponding author. Tel.: +39 06 4990 2713; fax: +39 6 49487101. technique and inductively coupled plasma mass spectrometry
E-mail address: francesco.cubadda@iss.it (F. Cubadda). (ICP-MS) was resorted to for analytical determinations.
0026-265X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
92 F. Cubadda, A. Raggi / Microchemical Journal 79 (2005) 9196

However, even though the sensitivity of ICP-MS was ade- manually broken in small pieces before being submitted to
quate to trace the often small changes in element concen- grinding. Two aliquots of each sample were ground until a
tration associated with the different technological steps of fine powder was obtained (15 min). The first aliquot was
processing, the detection of heavily interfered elements, discarded, while the second one was submitted to closed
especially Cr, as well as analyte quantification at ultratrace- vessel microwave digestion. The same dissolution technique
level level in water, posed analytical challenges that required was used for semolina. Each sample was digested in
suited analytical solutions. Various authors addressed the duplicate. For sample digestion, a Milestone MLS-1200
analytical issues connected with ICP-MS determination of Mega MW oven (FKV, Bergamo, Italy) was used with the
low-level and/or spectrally interfered elements in cereal following irradiation program: 1 min at 250 W, 2 min at 0 W,
matrixes [38] and water [911]. Time-consuming analyte 5 min at 250 W, 6 min at 400 W and 6 min at 650 W. About
separation/preconcentration procedures [3,7,9], the employ- 0.35 g of fresh matter were placed in the digestion vessels
ment of dcoldT plasma conditions [3], or the use of more with 3 and 1 ml of ultrapure concentrated HNO3 and H2O2,
complex and expensive sector field instruments [5,8,10,11] respectively. All samples were made up to 30 ml in
have been proposed in most of the studies carried out so far. polypropylene disposable tubes with high purity deionized
In the present work, the use of ultrasonic nebulization water. Water samples were analysed as such following
(UN) as a viable alternative to conventional nebulization not acidification with ultrapure HNO3 (1% v/v).
only for increasing sample introduction efficiency and thus
improving analytical sensitivity, but also for the alleviation of 2.3. Analyses
problematic spectral interferences, was investigated. It is
shown that changes in the sample introduction system and ICP-MS measurements were performed by a quadruple
complementary use of another fast and simple atomic Sciex Elan 6000 spectrometer (Perkin-Elmer, Norwalk, CT,
detector, namely inductively coupled plasma atomic emission USA), equipped with an ASX-500 autosampler model 510
spectrometry (ICP-AES), straightforwardly overcome the and an ADX-500 autodilutor (both from CETAC Technolo-
difficulties in determining the analytes in the selected gies, Omaha, NE, USA). Two sample introduction systems
matrixes. were employed in this study: a cross-flow nebulizer with a
Scott type spray chamber and a U-5000AT+ ultrasonic
nebulizer (CETAC Technologies). The U-5000AT+ features
2. Experimental a desolvation system consisting of a heated tube and a Peltier
condenser. Temperature settings were 142 8C for the heater
2.1. Samples and reagents and 3 8C for the cooler.
Calibration was performed with external standards or the
Samples of durum wheat grain, semolina (durum wheat method of standard addition, depending on the matrix.
flour), pasta and water were collected at an industrial plant Rhodium was selected as the internal standard at a
for milling and pasta making as described in a separate concentration of 2 Ag l 1 with UN and 20 Ag l 1 with
paper [1]. Two reference materialsRM 8436 durum wheat pneumatic nebulization (PN). Details on the instrumentation
flour, by NIST (Gaithersburg, MD), and BCR CRM 189 and the operating conditions are summarized in Table 1.
wholemeal flour, by IRMM (Geel, Belgium)were
included in the analytical batches to assess the accuracy of
the analytical procedure.
Table 1
Standard solutions were prepared by diluting commer- Instrumental operating conditions for ICP-MS
cially available mono-element stock solutions (1 g l 1)
Spectrometer Sciex Elan 6000 (Perkin-Elmer,
(Merck, Darmstadt, Germany) with high purity deionized Norwalk, CT, USA)
water (Milli-Q, Millipore, Molsheim, France). Calibration Plasma
standards were 1% v/v (water analyses) or 2% (other RF generator Frequency: 40 MHz, power
matrixes) in ultrapure concentrated HNO3 (Carlo Erba output 10501150 W
Ar flow rate (l min 1) Plasma: 16, auxiliary: 0.9,
Reagenti, Milan, Italy). The reagents used in sample diges-
nebulizer: 0.901.00
tion were ultrapure concentrated HNO3 and H2O2 (Merck). Solution uptake rate 1.8 ml min 1
In the study of spectral interferences, standard solutions of Interface Ni sampler and skimmer cones,
Mo and Ca and ultrapure Na2CO3 (Merck) were used. i.d. 1.1 and 0.9 mm, respectively
Data acquisition Peak hopping; dwell time 200 ms,
sweeps/reading 20, number of
2.2. Sample preparation
replicates 3
Internal standard Rh
Grain and pasta were homogenized for subsampling by an Analytical masses 114
Cd, 53Cr, 57Fe, 60Ni,
206, 207, 208
automatic mortar grinder (RM100, Retsch, Haan, Germany). Pb
The device bears a mortar and a pestle made of agate, which Masses for interference C, 43Ca, 98Mo
avoids the risk of sample contamination [12]. Pasta was
F. Cubadda, A. Raggi / Microchemical Journal 79 (2005) 9196 93

Table 2
Elemental equations for interference correction in ICP-MS measurements
Isotope Abundance Elemental equationa Interferences
114 114 118 98 114
Cd 28.7 M0.02683 Sn0.001 Mo Sn,98Mo16O, 96Mo18O
53 53
Cr 9.5 M0.001 13Cb 40
57 57
Fe 2.1 M0.04 43Ca 40
Ca16O1H, 40Ca17O
60 60
Ni 26.2 M0.004 43Ca 44
Ca16O, 43Ca16O1H
Correction factors for polyatomic interferences determined daily and entered into the instrument software with four significant figures.
Correction factor obtained with UN as sample introduction system.

Table 2 shows the elemental equations used for correction of were calculated on the basis of the 3j criterion from the
spectral interferences. standard deviation (j) of 10 determinations of the method
ICP-AES measurements were performed by an Optima blank carried out during the same analytical run. In the case
4300 DV spectrometer (Perkin-Elmer), using a U-5000AT+ of water samples, the method blank was a 1% v/v ultrapure
ultrasonic nebulizer for sample introduction. Calibration HNO3 solution. Repeatability was calculated as the relative
was performed with external standards and yttrium was standard deviation of 10 measurement of a sample with
selected as the internal standard at a concentration of 20 Ag concentration values in the central region of the analytical
l 1. Details on the instrumentation and the operating range carried out during the same analytical run.
conditions are summarized in Table 3.
Each sample was analysed in duplicate. Duplicate
analyses of selected samples were performed twice in 3. Results and discussion
different days and accordance of results was verified.
Samples of the two reference materials were analysed in 3.1. Analysis of cereal matrixes
each analytical run to check the accuracy of measurements.
Digestion blanks were run together with digested samples The analytical masses selected for ICP-MS measure-
belonging to the same analytical batch and their signal was ments are shown in Table 1. In the case of Cd and Pb, the
subtracted to that of samples when calculating analyte most abundant isotopes were chosen. The Pb isotopes are
concentrations. free of isobaric overlaps, while 114Sn interference on 114Cd
was of minor importance due to the low Sn levels in the
2.4. Figures of merit investigated matrixes (b5 ng g 1 dry wt. [14]) and was
corrected by an elemental equation based on relative natural
The uncertainty of analytical measurements was deter- abundances. For Fe, Ni and Cr, the major isotopes were
mined according to established guidelines [13]. Sampling avoided because of the severe spectral interferences from
and subsampling procedures did not significantly contribute plasma background and matrix elements. On the basis of our
to the combined standard uncertainty given the precautions previous experience with cereal samples [12,1417], the
taken during these steps (samples collected along the only significant interferences from matrix-induced polya-
processing line at fixed time intervals, homogenization tomic ions on 57Fe, 53Cr and 60Ni are represented by Ca
obtained by grinding the samples into a fine powder before oxides/hydroxides (57Fe, 60Ni) and the argoncarbon ion
subsampling). The expanded uncertainty U was calculated (53Cr). As regards the other analytes, the interference of Mo
by multiplying the combined standard uncertainty by a oxides on 114Cd was considered in this study as being
coverage factor k (k=2). possibly significant at low Cd concentrations. All these
Estimates of method detection limits (LODs) and short- interferences were characterized and quantified with the aim
term precision (repeatability) are reported in Table 4. LODs

Table 4
Table 3 Method detection limits (ng l 1) and repeatability (%)
Instrumental operating conditions for ICP-AES
Element Wheat based matrixes Water
Spectrometer Optima 4300 DV (Perkin-Elmer, Norwalk,
CT, USA) LOD Repeatability LOD Repeatability
a b
Plasma Cd 6.1 1.7 0.7 2.1
RF generator Frequency: 40 MHz, power: 1300 W Cr 24b 2.1 8.2b 1.9
Gas flow rate (l min 1) Plasma: 15, auxiliary: 0.2, nebulizer: 0.8 Fe 1.4103a 1.6 3.1102c 1.6
Solution uptake rate 1 ml min 1 Ni 99a 2.3 6.7b 2.5
Optical viewing Axial Pbd 16a 1.6 1.5b 1.7
Resolution High a
ICP-MS with PN for sample introduction.
Data acquisition Peak area; number of replicates 3 b
ICP-MS with UN for sample introduction.
Spectral lines (nm) Fe: 238.2, 239.6, 259.9 c
ICP-AES (259.9 nm).
Internal standard (nm) Y: 324.2 d 208
94 F. Cubadda, A. Raggi / Microchemical Journal 79 (2005) 9196

to minimizing their influence by careful selection of 340,000 cps/100 pg g 1 for 208Pb and the signal-to-noise
operating conditions and sample introduction systems. ratio was improved at least by a factor of 3. A LOQ of 15 ng
The effect of the residual signal from interfering species l 1 was achieved, which enabled quantification of Pb in all
was corrected by means of elemental equations (Table 2). In the samples.
the equations of Table 2, the signal of a particular Ni and Fe measurements were successfully carried out in
interference (e.g., 40Ar13C) is related to the signal of the the conventional mode. At the levels reported in Table 5 and
parent ion (13C) via a correction factor (CF, in this case: given the ranges of the concentration ratios Ca/Fe and Ca/Ni
Ar13C/13C). The CFs were determined daily by aspirating observed in the samples (1020 and 14002700, respec-
an interference check solution containing the particular tively, estimated on the basis of previous Ca determinations
interferent (i.e. Mo, Ca, C). These solutions were run [14]), no serious analytical problems were encountered and
regularly during analytical measurements to check for the spectral overlaps of Ca oxides/hydroxides could be
proper correction (i.e. dzeroT concentration for the affected easily corrected. The magnitude of these interferences was
analyte). If deviations were observed, the CF was adjusted in the range 14% and 29% of the measured Ni and Fe
until a zero output was obtained after reprocessing. It is concentrations, respectively.
worth noticing that this correction method requires a linear The quantification of Cr was carried out by means of
dependence of the interference signal on the parent ion UN-ICP-MS. In fact, the severity of 40Ar13C interference
signal. In all cases, a linear dependence in the region of does not allow the reliable determination of this analyte at
analytical interest was observed in this study, with the levels shown in Table 5 with conventional nebulization.
regression lines always characterized by a R 2z0.995. As we reported in a previous study [16], the use of UN (i)
Another important remark on the CFs of Table 2 concerns improves the signal intensity and the signal-to-noise ratio
the operating conditions selected for their determination. for 53Cr, thus overcoming a major limitation of this isotope
Throughout the study, the RF power and nebulizer gas flow (i.e. the limited counting statistics which affects the
rate were adjusted in order to keep the formation of oxides precision of measurements), (ii) enhances signal stability
(measured as the signal ratio CeO/Ce) below 2%. This due to the alleviation of the spectral interferences arising
obviously influenced the values of the CFs, especially those from the plasma background, (iii) reduces the interference of
of the oxygen-containing polyatomic ions. Ar13C at very low levels. We found that the net signal of
For the quantification of Cd in grain and derived this polyatomic species was less intense when UN was used
products, ICP-MS with conventional (i.e. pneumatic) instead of PN, whereas that of 53Cr showed a major (i.e.
nebulization performed more than satisfactorily. Table 5, about 10-fold) enhancement with the operating conditions
which shows the results obtained for one of the matrixes used in this study. Thus, the severity of the 40Ar13C/53Cr
analysed in this study, gives an idea of the range of overlap was definitely reduced and this allowed to easily
concentration values encountered. The interference from correct the residual bias from 40Ar13C with the equation
MoO turned out to be of minor importance. In fact, the shown in Table 4. This interference always accounted for
signal ratio MoO/Mo was found to be about 0.1% under the less than 7% of the measured Cr concentrations.
instrumental conditions used and the overall interference of In Table 6, the accuracy of the analytical procedure as
this species was in the range 13% of the measured Cd tested by means of the two reference materials can
concentrations. be appraised. The good agreement with the certified or best
The same analytical approach was used for Pb determi- estimated values indicates effective recovery of analytes
nation. However, in this case several samples showed after digestion and subsequent accurate detection. These
concentration values close to or below the limit of results were obtained by calibration with external standards.
quantification of the method (LOQ), estimated on the basis With the aim of assessing the possible existence of more
of the 10j criterion. Therefore, for these samples, analyses severe matrix effects in the unknown samples, 10 digestates
were repeated using UN for sample introduction in order to were analysed with method of standard addition. However,
increase signal intensities and improve the detection no significant differences were observed in comparison with
capability. With UN, the net sensitivity was typically about the results obtained by the other quantification method.

Table 5
Concentration values found in semolina (ng g dry wt.)
Element Sample 1 Sample 2 Sample 3
Mean U Mean U Mean U
Cd 26.4 1.0 15.9 0.9 18.0 0.8
Cr 7.1 0.7 16.7 0.9 10.0 0.6
Fe 12.9103 0.5103 13.7103 0.5103 11.4103 0.4
Ni 93 6 98 5 215 13
Pb 1.9 0.2 12.8 0.6 6.2 0.4
F. Cubadda, A. Raggi / Microchemical Journal 79 (2005) 9196 95

Table 6
Concentration values found in the reference materials (ng g dry wt., n=5)
Element RM 8436 (durum wheat flour) CRM 189 (wholemeal flour)
Found Best estimated Found Certified
Mean U Mean CIa Mean U Mean CIa
Cd 103.0 8.9 110 50 72.1 7.9 71.3 3.0
Cr 24.9 6.0 23 9 69.3 6.1 [5773]b
Fe 41.5103 2.3103 41.5103 4.0103 69.2103 2.5103 68.3103 1.9103
Ni 173 22 170 80 361 24 [380]c
Pb 23.5 4.4 23 6 385 30 379 12
Uncertainty as half-width of the 95% confidence interval of the mean.
Indicative range.
Indicative value.

As regards the precision of measurements, Table 4 shows The results obtained for two water samples collected
that a repeatability relative standard deviation V2.3% was within short time intervals at the same point of the
achieved for all analytes. Repeated analyses of selected investigated manufacturing plant are reported in Table 7.
samples in different days showed a satisfactory accordance Notwithstanding the low concentrations of the analytes and
of results, with a disagreement between duplicates always the severity of specific interferences, quantification of all the
b10%. elements was successfully achieved, with the exception of
The main problem encountered in the study on cereal iron. Fe determination was remarkably improved by the use
matrixes was blank and sample contamination that, despite of UN-ICP-MS, and satisfactory calibration lines were
all precautions, caused rejection of the results for a certain reliably obtained in the range 020 Ag l 1, which is not
number of digestates. About 6% and 4% of Pb and Cr feasible with conventional nebulization. The Ca-related
measurements, respectively, had to be discarded, partic- interference was reduced by a factor of 23, but was still
ularly in the low concentration range. Precautions included so severe to enable only a semiquantitative measurement of
sample treatment in clean room conditions, use of ultra- the Fe concentrations. Thus, ICP-AES was resorted to and
pure reagents and new non-contaminating disposable the same ultrasonic nebulizer used above was employed for
plasticware, two cleaning irradiation cycles before dis- sample introduction. Excellent four point calibration lines
solution of each batch of samples in order to effectively were obtained in the range 010 Ag l 1 (the analytical
decontaminate the digestion vessels. This finding clearly region of interest) and satisfactory measurements were
shows that determinations at very low levels are often performed, with a mean precision of about 2% for the
hampered rather by unforeseeable contamination events unknown samples.
and blank values, than by the detection power of As no reference material was available to assess to
instruments. accuracy of water determinations, the samples were spiked
with the analytes at the levels reported in Table 7 and the
3.2. Water analysis resulting recoveries were calculated. These latter turned out
to be in the range 96104% for all the elements and
Water samples were analysed by UN-ICP-MS with indicated that reliable detection of the analytes was
external calibration. UN was resorted to in order to improve performed at the concentration levels found in the samples.
the detection limits of all the analytes and reduce the effects The short-term precision of measurements was also satis-
of specific spectral interferences, particularly those of the factory, with repeatability relative standard deviations
oxygen-containing polyatomic ions affecting Fe and Ni V2.5% for all analytes (Table 4).
determination. Generally, solvent-related interferences are
reduced with UN because of water removal by the
desolvation unit included in the nebulization device.
Table 7
Minimization of the bias arising from Ca oxides/hydroxides Concentration values found in water (Ag l 1)
is mandatory because of the severity of these interferences Element Sample 1 Sample 2 Spike recovery (%)a
in water analysis. In fact, Ca levels in drinking as well as
Mean U Mean U
mineral water can be of several tenths of mg 1 1, whereas
other elements, including Fe and Ni, are usually present as Cd 0.0113 0.0009 0.0105 0.0010 96
Cr 0.334 0.015 0.321 0.014 97
traces. In the water samples analysed in this study, Ca Feb 4.47 0.17 3.91 0.19 97
concentrations of about 40 mg l 1 were measured and the Ni 0.173 0.011 0.199 0.013 98
concentration ratios Ca/Fe and Ca/Ni were, respectively, Pb 0.312 0.013 0.223 0.010 104
about 500 and 60 times higher than those in the cereal a
Spike concentrations (Ag l 1): Cd, 0.1; Cr, Pb, Ni, 0.2; Fe, 5.
samples. Determined by ICP-AES.
96 F. Cubadda, A. Raggi / Microchemical Journal 79 (2005) 9196

4. Conclusions References

The results obtained in this work show that judicious [1] F. Cubadda, A. Raggi, E. Marconi, Microchem. J. 79 (2005) 97 102.
choice of the sample introduction system can often [2] European Commission, Commission regulation (EC) No 466/2001 of
8 March 2001 setting maximum levels for certain contaminants in
straightforwardly overcome the difficulties in determining foodstuffs. Official Journal of the European Communities, 16.3 (2001)
ultratrace and spectrally interfered elements in water and L77/113.
food matrixes. The benefits of UN in enhancing sensitivity [3] C.J. Park, J.K. Suh, J. Anal. At. Spectrom. 12 (1997) 573.
and reducing the effects of spectral interferences from [4] S. Wu, X. Feng, A. Wittmeier, J. Anal. At. Spectrom. 12 (1997) 797.
different polyatomic species (oxides, hydroxides, argides) [5] I. Rodushkin, Fresenius J. Anal. Chem. 362 (1998) 541.
[6] X. Feng, S. Wu, A. Wharmby, A. Wittmeier, J. Anal. At. Spectrom. 14
were demonstrated. Water removal by the desolvator (1999) 939.
condenser unit of the ultrasonic nebulizer evidently accounts [7] A. Kawasaki, H. Watanabe-Oda, in: G. Holland, S.D. Tanner (Eds.),
for the reduction of solvent-related isobaric interferences Plasma Source Mass SpectrometryNew Developments and Appli-
such as oxides and hydroxides. However, the observed cations, Spec. Publ., vol. 241, The Royal Society of Chemistry,
alleviation of the 40Ar13C interference on 53Cr must also be Cambridge, UK, 1999, pp. 173 181.
[8] S. DIlio, M. Alessandrelli, R. Cresti, G. Forte, S. Caroli, Microchem.
linked to a partial reduction of the aerosol carbon load J. 73 (2002) 195.
during the desolvation step. In fact, with the operating [9] W. Zhu, E.W.B. de Leer, M. Kennedy, G.J.F.R. Alaerts, Fresenius J.
conditions used the net signal of this polyatomic species was Anal. Chem. 360 (1998) 74.
less intense with UN than with conventional PN and this [10] B. Klaue, D.B. Blum, Anal. Chem. 71 (1999) 1408.
demonstrates that carbon removal following UN occurred to [11] J. Riondato, F. Vanhaecke, L. Moens, R. Dams, J. Anal. At. Spectrom.
15 (2000) 341.
some extent. [12] F. Cubadda, M. Baldini, M. Carcea, L.A. Pasqui, A. Raggi, P.
UN in conjunction with ICP-AES was selected for the Stacchini, Food Addit. Contam. 18 (2001) 778.
determination of iron in drinking water containing several [13] EURACHEM/CITAC, in: S.L.R. Ellison, M. Rosslein, A. Williams
milligrams per liter of Ca. This should be regarded as the (Eds.), EURACHEM/CITAC Guide CG 4, 2nd edition, 2000, p. 120.
[14] F. Cubadda, E. Marconi, A. Raggi, F. Zanasi, R. Cubadda,
best analytical solution for the reliable quantification of
Proceedings of the 2nd International Workshop dDurum Wheat and
Fe at the micrograms per liter level in such a Ca-rich Pasta Quality: recent Achievements and New Trends, Rome, 2003,
matrix. pp. 41 46.
Overall, a robust analytical approach for the determina- [15] F. Cubadda, A. Raggi, A. Testoni, F. Zanasi, J. AOAC Int. 85 (2002)
tion of Cd, Pb, Fe, Ni and Cr in drinking water, cereal-based 113.
products and similar matrixes was outlined. Furthermore, [16] F. Cubadda, S. Giovannangeli, F. Iosi, A. Raggi, P. Stacchini, Food
Chem. 81 (2003) 463.
high-throughput is guaranteed due to the adoption of [17] F. Cubadda, A. Raggi, F. Zanasi, M. Carcea, Food Addit. Contam. 20
microwave closed vessel digestion as rapid sample treatment (2003) 353.
procedure for solid matrixes.