736 VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 RESIDUES AND TRACE ELEMENTS
Determination of Cadmium, Aluminium, and Copper in Beer and Products Used in Its Manufacture by Electrothermal Atomic Absorption Spectrometry PILAR VIÑAS, NEREA AGUINAGA, IGNACIO LÓPEZ-GARCÍA, and MANUEL HERNÁNDEZ-CÓRDOBA1 University of Murcia, Department of Analytical Chemistry, Faculty of Chemistry, E-30071 Murcia, Spain
Procedures were developed for determining cadmium, aluminium, and copper in beer and the products used in its manufacture by electrothermal atomic absorption spectrometry. Beer samples were injected into the furnace and solid samples were introduced as suspensions after preparation in a medium containing hydrogen peroxide, nitric acid, and ammonium dihydrogen phosphate for cadmium atomization. Calibration was performed with aqueous standards, and characteristic masses and detection limits were, respectively, 1 and 0.3 pg for cadmium, 18 and 5.4 pg for aluminium, and 5.6 and 6.8 pg for copper. Different samples of beer, wort, brewer’s yeast, malt, raw grain, and hops were analyzed by the proposed procedures. Cadmium was found in low concentrations (0.001–0.08 mg/g and 0–1.3 ng/mL); copper (3–13 mg/g and 25–137 ng/mL) and aluminium (0.6–9 mg/g and 0.1–2 mg/mL) were found at higher levels. The reliability of the procedure was confirmed by comparing the results obtained with others based on microwave oven sample digestion, and by analyzing several certified reference materials.
eer is manufactured from malted barley, hops, yeast, and water (1). Although barley is the most important raw material, it is the hops that clearly differentiate beer from other drinks. Hops also have a great capacity to prevent infectious processes. Brewer’s yeast has a high nutritional value because it contains high levels of important nutrients, including vitamins, amino acids, and minerals. Recent studies have demonstrated that a moderate intake of beer has beneficial effects on health. The minerals it contains (0.3–0.4%) are principally potassium, phosphate, calcium, magnesium, iron, rubidium, manganese, chloride, sulfur, and silicon (2). Other elements present in smaller amounts are zinc, copper, chromium, aluminium, and selenium. Beer and the products used in its manufacture are subject to certification
B
Received October 11, 2001. Accepted by JS December 29, 2001. 1 Author to whom correspondence should be addressed; e-mail:
[email protected].
to ensure that chemical specifications are met for trace metals, because the contamination of cereals by industrial emissions has the potential to produce toxicological problems. Electrothermal atomic absorption spectrometry (ETAAS) is an excellent technique for determining low concentrations of metals in foods. The samples are introduced into the atomizer in the form of solutions obtained from the mineralized samples. Digestion methods require dry ashing or acid treatment, which may result in contamination or loss of metals. Modern digestion methods (microwave in Teflon vessels under pressure) yield good results; however, this instrumentation is expensive and is not available in quality control laboratories. Another possibility is direct introduction of liquid samples without mineralization or preparation of suspensions from solid samples (3). The introduction of suspensions into the atomizer has practical advantages over time-consuming conventional procedures that are based on total dissolution of the samples, and more exact results are obtained because reagent contamination is avoided. Although atomic absorption spectrometry (AAS) with flame atomization (FAAS) or ETAAS has been used to determine aluminium, copper, and cadmium in beer and the products used in its manufacture (4–17), those procedures involve previous mineralization of the sample. This study describes the determination of the essential metals aluminium and copper and the toxic metal cadmium by using ETAAS. Beer is manufactured from natural materials including water, barley, and yeast, all of which are potential sources of aluminium and copper. Furthermore, beer may be sold in aluminium or stainless steel cans and kegs, and corrosion processes may increase the risk of contamination by these metals or other toxic metals such as cadmium. Liquid samples of different types of beer, nonfermented beer, and wort were introduced into the atomizer. Solid samples such as yeast, malt, raw grain, and hops were introduced in the form of suspensions, thus avoiding a prior mineralization step. To avoid accumulation of carbonaceous residues and high background values produced by calcination of organic matter, suspensions were prepared in the presence of hydrogen peroxide and nitric acid to provide an oxidizing environment during atomization (18–20). The procedure was checked by analyzing several standard reference materials (SRMs) and by dissolving the samples in a closed system.
VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 737 Table 1. Furnace heating programs Element Step Dry
Pyrolysis
Cold
a
Atomization
Clean
a
Parameter
Cd (platform atomization)
Al (platform atomization)
Cu (wall atomization)
T (°C)
200
120
120
Ramp (s)
15
1
3
Hold (s)
35
50
35
T (°C)
500
1300
Ramp (s)
10
1
Hold (s)
30
T (°C)
20
Ramp (s)
1
Hold (s)
15
T (°C)
—
30 —
—
1500
2400
2000
Ramp (s)
0
0
0
Hold (s)
3
4
3
2600
2600
2600
Ramp (s)
T (°C)
0
0
0
Hold (s)
3
3
3
The flow of argon was stopped during the atomization step.
Experimental
Instrumentation (a) Atomic absorption spectrometers.—(1) Perkin-Elmer Model 1100B equipped with deuterium-arc background correction and an HGA-400 (Perkin-Elmer, Norwalk, CT) graphite furnace atomizer. Pyrolytic graphite coated tubes (Part No. B013-5653) and pyrolytic graphite platforms (Part No. B012-1092) inserted into pyrolytic graphite-coated tubes were obtained from Perkin-Elmer. (2) ATI-Unicam (Unicam Atomic Absorption, Cambridge, UK) 939QZ equipped with GF90 electrothermal atomizer was also used. 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. Measurements were made at 324.8, 309.3, and 228.8 nm for Cu, Al, and Cd, respectively, by using hollow cathode lamps (Photron) operated at 15 mA for Cu and Al, and 10 mA for Cd with a bandwidth of 0.7 nm for all the elements. Argon was used as the inert gas, with a flow rate of 300 mL/min in all stages except atomization, when the flow was stopped. Background-corrected integrated absorbance was used in all cases as the analytical signal. Finally, as the Zeeman corrector was not necessary, analyses were made with the Perkin-Elmer instrument. (b) Manual homogenization vessels (potters).—10 mL; equipped with Teflon plungers (Afora, Barcelona, Spain). Plastic (polypropylene) vessels of the type commonly used to collect clinical samples were used to store solutions or suspen-
sions. These were previously washed with 20% (v/v) nitric acid and rinsed with ultrapure water. Pipet tips were also of polypropylene. A Branson (Danbury, CT) ultrasonic bath of 55 KHz and 14 W constant power was used.
Figure 1. Effect of both calcination and atomization temperatures on analytical signals (solid lines) and backgrounds (dotted lines) for 0.4 ng/mL aqueous cadmium solution and beer sample in absence (A, C) and presence (B, D) of 0.2% (m/v) ammonium dihydrogen phosphate.
738 VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002
Figure 2. Effect of both calcination and atomization temperatures on analytical signals (solid lines) and backgrounds (dotted lines) for atomization of Al from 0.5% (m/v) yeast suspension using wall and platform atomization.
(c) Microwave oven.—Mineralization of the samples for comparison purposes was performed in closed Teflon cups by using an MLS-1200 MEGA microwave oven (Milestone, Bergamo, Italy) and an MDR-1000/6 Rotor (Radiometer, Copenhagen, Denmark).
Reagents
phate (for Cd atomization) to 10 mL beer. Suspensions of solid samples were prepared by weighing samples (0.2–0.3 g) directly into the potter, and then adding 50 µL concentrated (65%, m/v) nitric acid, 6.75 mL concentrated (30%, m/v) hydrogen peroxide (1.35 mL for atomization of Al), and 0.02 g ammonium dihydrogen phosphate (for Cd atomization) diluted to 10 mL with deionized water. The slurries were homogenized by repeated movements of the plunger to obtain good suspension and were sonicated for 5 min to ensure absence of lumps. Then, 20 µL aliquots were taken while the solution was stirred continuously with a magnetic stirrer and injected into the furnace. The heating programs shown in Table 1 were run, and the background-corrected peak areas caused by analytes were obtained. Calibration was performed by using aqueous standards. These standards were prepared by diluting the concentrated solutions with water at concentrations of 0.1–5 for Cd, 0.5–40 for Cu, and 1–50 ng/mL for Al. The certified reference samples were analyzed in the same manner. To confirm the reliability of the procedure, the samples were previously analyzed for comparison purposes. Fractions (1 g) of samples were weighed into Teflon cups, and 3 mL concentrated nitric acid and 0.5 mL concentrated hydrogen peroxide were added. The program used in the microwave oven was 1 min at 250 W, 2 min at 400 W, 4 min at 600 W, 2 min at 750 W, 1 min at 1000 W, 1 min at 500 W, and 1 min at 250 W. After this treatment, samples were maintained in the closed cups for 10 min before being diluted with deionized water in 10 mL volumetric flasks. Solutions were analyzed by ETAAS.
(a) High quality water.—Obtained with a Milli-Q system (Millipore, Bedford, MA); used exclusively. (b) Standard solutions (1000 mg/mL) of Al, Cu, and Cd.—Panreac (Barcelona, Spain), quality for AAS, 1.000 ± 0.002 g/L, and diluted as necessary to obtain working standards. (c) Nitric acid.—High-quality concentrated (65%, m/v; Merck, Darmstadt, Germany). (d) Hydrogen peroxide.—30% (m/v; Fluka, Buchs, Switzerland), puriss. >99%. (e) Ammonium dihydrogen phosphate.—Fluka, puriss. >99%.
Reference Materials and Samples (a) SRMs.—Citrus leaves (SRM 1572), Apple leaves (SRM 1515), Rice flour (SRM 1568a), Wheat flour (SRM 1567a), and Oyster tissue [SRM 1566a; National Institute of Standards and Technology (NIST); Gaithersburg, MD]. (b) Liquid samples.—Beer of different types (with and without alcohol), packed with different containers, filtered beer, beer in guard, fermenting beer, and wort. (c) Solid samples.—Brewer’s and freeze-dried yeast, malt, raw grain, and hops.
Procedures Liquid samples were prepared by adding 50 µL concentrated (65%, m/v) nitric acid, 1.35 mL concentrated (30%, m/v) hydrogen peroxide, and 0.02 g ammonium dihydrogen phos-
Figure 3. Variation of Al signal with atomization temperature for 0.5% (m/v) yeast suspension. Segments indicate standard deviation values; bar graph represents variation of RSD for each temperature.
VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 739
Figure 4. Influence of atomization temperature on analytical signals of Cu from 2 ng/mL aqueous solution, beer sample, and 2% (m/v) yeast suspension using fast program methodology.
Results and Discussion
Optimization of Furnace Heating Programs The furnace program for Cd was optimized by platform atomization. The optimal drying temperature and holding time were selected to avoid sputtering and to dry the sample totally before pyrolysis. Temperatures >300°C produced sputtering, and best results were obtained at 200°C, with 15 s slow ramp and 35 s holding time. The conditions of the pyrolysis step were selected to provide high temperature without analyte loss, and to achieve good separation between atomic signal and background. The influence of pyrolysis temperature on analytical signals is shown in Figure 1 in the absence of a matrix modifier (A) and in the presence of ammonium dihydrogen phosphate (B). In the absence of phosphate (Figure 1A), there was a slight decrease in the analytical signal for the aqueous solution, which was considerably more pronounced for the beer sample when higher temperatures were used. This effect was probably due to the presence of chloride anions in the beer, which led to the rapid volatilization of Cd. The background signals rapidly decreased with the temperature. In the presence of ammonium dihydrogen phosphate (Figure 1B), Cd was stabilized at 500°C, which was selected as the optimal temperature. The atomization temperature providing maximal signals was 1500°C for both samples in the presence of phosphate (Figure 1D). However, because the atomization profile appeared very quickly in nonisothermal conditions, a cool step was included in the heating program to retard the signals. This step consisted of a 1 s ramp at 20°C with 15 s holding time. A cleaning stage was also included in the heating cycle. Finally, the selected temperatures were checked for atomization of solid samples. A 5% (m/v) brewer’s yeast suspension was prepared, and the heating program was restudied. It was found that the same optimized temperatures were valid for solid samples.
For Al determination, the heating program was optimized by using both wall and platform atomization modes. Figure 2 shows the variation of atomic signals and background values when pyrolysis and atomization temperatures were varied. The optimal value for pyrolysis temperature was 1300°C because higher temperatures led to Al losses or the appearance of double peaks. When atomization temperature were varied, maximum analytical signals were obtained at 2200 and 2400°C, using wall and platform atomization, respectively. Comparison of both atomization systems shows that the best area and height for the peaks were obtained by platform atomization, with the peaks appearing in isothermal conditions. Wall atomization led to very high and narrow peaks, which were resolved during the first second of atomization. Figure 3 compares both atomization modes and repeatability of results obtained. The figure shows variation of the Al signal with atomization temperature; the segments are the standard deviation values for 3 sucessive injections of 0.5% (m/v) yeast suspension by using both atomization modes. The bar graph shows the relative standard deviation (RSD) values obtained, with mean values of 7.5 and 3.9% for wall and platform atomization, respectively, demonstrating the greater repeatability obtained with platform atomization. Moreover, the linear calibration range was wider with this technique. For Cu, wall atomization was selected. Because the samples contained high amounts of Cu and very concentrated suspensions were not needed, the use of the fast-program methodology (21) was tried. The drying and pyrolysis temperatures were substituted by a modified drying step, at 120°C with a ramp of 3 s to avoid sputtering. The atomization temperature was varied, and the analytical signal increased to 2000°C for all samples (Figure 4). At higher temperatures (2200°C), the signal again increased, but the peak became distorted and was not reproducible. The repeatability of both methodologies was also compared to check the possible accumulation of carbonaceous residues inside the atomizer when the fast program was used. Ten successive aliquots of the 2% (m/v) yeast suspen-
Figure 5. Effect of ammonium dihydrogen phosphate concentration on Cd signals for aqueous solution containing 0.4 ng/mL Cd and beer sample.
740 VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002
Figure 6. Influence of suspension concentration on atomic signals of Cd, Cu, and Al from brewer’s yeast (A), and slopes of aqueous calibration and standard additions calibration graphs for Cd in yeast suspensions (B).
sion were injected, and RSD values of 2.8 and 11.7% were obtained for the rapid and the conventional programs, respectively. Consequently, fast program methodology was selected. It was also necessary to include a cleaning step at 2600°C. Table 1 shows the programs finally recommended for all the analytes.
Optimization of Chemical Agent Concentrations A matrix modifier was added to avoid volatilization losses and to retard the analytical signal with respect to the background. The matrix modifier selected for atomization of Cd was ammonium dihydrogen phosphate, the concentration of which was varied between 0 and 0.5% (m/v). Figure 5 shows that the presence of phosphate had a pronounced effect on the signal for the beer sample, which increased as phosphate concentration increased to 0.2% (m/v), and then remained constant. For the
aqueous solution, the signal did not increase in the presence of phosphate. Consequently, this 0.2% (m/v) was selected for stabilizing Cd in beer. Cd was determined by the 2 systems most commonly used for background correction. A comparison study was performed by injecting consecutive aliquots of a 10% (m/v) yeast suspension into the atomizer and using both the deuterium and Zeeman correction systems. Similar results concerning the shape of analytical peaks and corrected signals were obtained, and the deuterium device was selected. For atomization of Cu, the addition of ammonium nitrate has been recommended as matrix modifier (22); therefore, the concentration of this chemical was varied between 0 and 0.5% (m/v). As the analytical signal was not modified in the presence of ammonium nitrate and the separation from the background was not improved, Cu was atomized in the absence of modifier. Aluminium was also atomized without a matrix modifier. The presence of hydrogen peroxide has an important effect on the elimination of carbonaceous residues in the tube when materials of a high organic content are introduced into the atomizer (18–20). This effect is especially important for Cd because the low content of the metal in the samples requires highly concentrated suspensions. Consequently, the concentration of this oxidant was varied between 0 and 30% (v/v). The results obtained for Cd and Cu showed that an increase in the oxidant concentration did not produce a variation in the atomic signals, whereas background values slightly decreased for all the samples. A 30% (v/v) hydrogen peroxide concentration was selected to decrease the formation of carbonaceous residues inside the atomizer caused by the organic matter of the solid samples. For beer samples, a 5% (v/v) concentration was selected to avoid sensitivity losses by dilution. For atomization of Al, the addition of hydrogen peroxide considerably increased both atomic signal and blank values. Thus, 5% (v/v) was selected for all samples to improve the detection limit and the repeatability of Al determination. Nitric acid was added to the suspension to obtain good sensitivity and to increase the fraction of analytes extracted into the supernatant, thereby improving reproducibility of the procedure. When the concentration of the acid was varied in the 0–3% (v/v) range for a yeast suspension, the percentage of all metals extracted into the supernatant increased as the nitric
Table 2. Slopes of standard additions for beer manufacturing products Sample (0.5%, m/v)
Cd slopea, s mL/ng
Cu slopea, s mL/ng
Al slopea, s mL/ng
Aqueous standards
0.0867 (0–2 ng/mL)
0.0077 (0–30 ng/mL)
0.0049 (0–40 ng/mL)
Beer
0.0911 (0.4–2 ng/mL)
0.0080 (25–38 ng/mL)
0.0048 (22–40 ng/mL)
Brewer’s yeast
0.0865 (0.1–2 ng/mL)
0.0068 (17–30 ng/mL)
0.0053 (12–40 ng/mL)
Hops
0.0793 (0.7–2 ng/mL)
0.0075 (12–30 ng/mL)
0.0045 (29–40 ng/mL)
Malt
0.0708 (0.1–2 ng/mL)
0.0081 (5–30 ng/mL)
0.0047 (8–40 ng/mL)
Wort
0.0795 (0–2 ng/mL)
0.0075 (26–38 ng/mL)
0.0054 (2–40 ng/mL)
Raw grain
0.0854 (0.2–2 ng/mL)
0.0075 (8–30 ng/mL)
0.0055 (9–40 ng/mL)
a
Values in parentheses indicate concentration range for standards and samples for which slopes are calculated.
VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 741 Table 3. Characteristics of the recommended procedures Parameter
Cd
Cu
Al
0.1–5
0.5–40
1–50
Characteristic mass, pg
1
5.6
Detection limit, pg
0.3
6.8
5.4
RSD, %
2.3
1.9
4.9
Calibration range, ng/mL
18
acid concentration increased. Finally, 0.5% (v/v) nitric acid was selected.
Influence of Suspension Concentration and Study of Matrix Effect The influence of suspension concentration on Cd signal was studied for a brewer’s yeast suspension in the 0–10% (m/v) range (Figure 6A). Linearity was obtained up to a 5% (m/v) concentration; however, this limit must be considered as a guideline, as it depends on the concentration and na-
ture of Cd in the sample. To test the absence of a matrix effect, we compared the slopes of the aqueous calibration and standard additions calibration graphs obtained for the yeast suspension at different concentrations. The bar graph in Figure 6B shows the results obtained. Each graph was constructed from 4 points (0, 0.6, 1.2, and 1.8 ng/mL), and each point was measured 3 times. Although, the slopes were similar up to a 5% (m/v) concentration, they were lower for higher suspension concentrations due to loss of linearity or appearance of interferences by the matrix. Similar studies were performed for Cu and Al, with linearity observed up to 1% (m/v) for Cu and 2% (m/v) for Al (Figure 6A), a limit which again must be considered as a guideline. This study was extended to beer samples and to the products used in its manufacture to ascertain whether matrix effects existed. Table 2 shows the results obtained for the 3 metals, which indicate that slopes of best-fit regression lines for standard additions to the different suspensions varied by 8.1% (RSD) for Cd, 5.6% for Cu, and 7.6% for Al (n = 7). Consequently, no matrix effect exists and direct calibration against aqueous standard solutions can be made for the 3 elements for suspension concentrations within the linearity range.
Table 4. Metal contents in beer manufacturing products Cd Sample
Suspensiona
Cu Mineralizationa
Suspensiona
µg/g
Solid samples
Al Mineralizationa
Suspensiona
µg/g
Mineralizationa µg/g
Brewer’s yeast
0.001 ± 0.002
0.001 ± 0.002
2.9 ± 0.3
2.6 ± 0.3
1.2 ± 0.3
1.0 ± 0.4
Yeast (freeze-dried)
0.083 ± 0.005
0.075 ± 0.005
5.7 ± 0.2
6.1 ± 0.2
0.59 ± 0.03
0.51 ± 0.07
Hops
0.056 ± 0.005
0.050 ± 0.005
13.6 ± 0.3
13.1 ± 0.4
1.0 ± 0.3
1.3 ± 0.2
Malt
0.001 ± 0.002
0.001 ± 0.002
6.2 ± 0.2
5.8 ± 0.3
8.6 ± 1.2
9.4 ± 1.0
Raw grain
0.003 ± 0.004
0.007 ± 0.003
6.2 ± 0.2
5.7 ± 0.3
4.4 ± 0.2
4.7 ± 0.2
Liquid samples
ng/mL
µg/mL
ng/mL
Must
Not detected
Not detected
137 ± 3
131 ± 4
0.11 ± 0.05
0.15 ± 0.04
Beer in guard
Not detected
Not detected
32 ± 3
34 ± 2
0.09 ± 0.04
0.12 ± 0.03
Nonfermented beer
Not detected
Not detected
84 ± 2
80 ± 3
0.23 ± 0.04
0.20 ± 0.05
Filtered beer
0.69 ± 0.002
0.70 ± 0.01
31 ± 3
29 ± 2
0.41 ± 0.1
0.46 ± 0.1
Bottled beer
0.35 ± 0.01
0.33 ± 0.01
50 ± 3
53 ± 2
0.52 ± 0.1
0.56 ± 0.1
Canned beer
1.0 ± 0.003
Draught beer
0.75 ± 0.006
55 ± 1
52 ± 2
0.76 ± 0.06
0.71 ± 0.08
0.84 ± 0.01
1.1 ± 0.009
41 ± 3
42 ± 3
0.96 ± 0.1
1.0 ± 0.1
Bottled beer without alcohol
0.35 ± 0.002
0.37 ± 0.008
49 ± 3
46 ± 2
0.30 ± 0.1
0.25 ± 0.1
Canned beer without alcohol
0.54 ± 0.005
0.51 ± 0.01
25 ± 2
27 ± 3
0.32 ± 0.04
0.35 ± 0.05
Wheat beer
0.47 ± 0.007
0.45 ± 0.01
59 ± 1
57 ± 3
0.52 ± 0.1
0.59 ± 0.1
Hemp beer
0.084 ± 0.002
0.088 ± 0.002
42 ± 1
40 ± 2
1.3 ± 0.3
1.2 ± 0.1
Black beer
0.34 ± 0.004
0.36 ± 0.008
28 ± 3
31 ± 2
2.1 ± 0.1
2.2 ± 0.1
Lemon beer
0.58 ± 0.004
0.56 ± 0.01
32 ± 2
35 ± 2
0.70 ± 0.1
0.78 ± 0.1
Cherry beer
0.89 ± 0.002
0.91 ± 0.09
51 ± 4
55 ± 3
0.81 ± 0.1
0.86 ± 0.1
1.3 ± 0.002
1.2 ± 0.01
44 ± 2
41 ± 1
1.3 ± 0.1
1.2 ± 0.1
Homemade beer a
Mean ± standard deviation (n = 3).
742 VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 Table 5. Results for certified reference materials Cd, µg/g Sample
Suspensiona
Citrus leaves (SRM 1572) Apple leaves (SRM 1515)
Cu, µg/g
Al, µg/g
Certified
Suspensiona
0.038 ± 0.008
0.03 ± 0.01
17.2 ± 0.5
—
Not certified
5.6 ± 0.1
5.64 ± 0.24
289 ± 11
286 ± 9
Rice flour (SRM 1568a)
0.023 ± 0.002
0.022 ± 0.002
2.7 ± 0.2
2.4 ± 0.3
4.0 ± 0.8
4.4 ± 1.0
Wheat flour (SRM 1567a)
0.028 ± 0.002
0.026 ± 0.002
2.5 ± 0.2
2.1 ± 0.2
5.1 ± 0.9
5.7 ± 1.3
3.93 ± 0.01
4.15 ± 0.38
70 ± 4.4
66.3 ± 4.3
200 ± 15
202.5 ± 12.5
Oyster tissue (SRM 1566a) a
Certified 16.5 ± 1
Suspensiona 92 ± 4
Certified 92 ± 15
Mean ± standard deviation (n = 3).
Calibration Graphs and Repeatability Table 3 shows the characteristics of the calibration graphs. The detection limits were calculated for 10 successive injections of the blank and using the 3σ criterium. The values obtained for the blanks were 0.08, 0.63, and 18.4 ng/mL for Cd, Cu, and Al, respectively. The repeatability was calculated by using the RSD for 10 successive injections of a beer sample containing 0.35, 50, and 520 ng/mL Cd, Cu, and Al, respectively.
Results and Accuracy Table 4 summarizes results obtained for the different beer samples and products used in its manufacture by the proposed procedure as well as a reference method based on mineralization and involving a closed system. Cd was found at very low concentrations, whereas Cu and Al appeared at higher concentrations, with metal contents increasing as the manufacturing process progressed. The statistical study was performed separately for solid and liquid samples. The paired t-test revealed no significant difference between results obtained by using either procedure at 0.05 level of significance. Statistical values were t = 0.913 (p = 0.413) for Cd in solid samples; t = –0.402 (p = 0.695) for Cd in liquid samples; t = 1.537 (p = 0.199) for Cu in solid samples; t = 0.572 (p = 0.576) for Cu in liquid samples; t = –1.277 (p = 0.271) for Al in solid samples; and t = –0.816 (p = 0.428) for Al in liquid samples. The reliability of the method was further corroborated by using several certified reference materials. Table 5 shows results obtained for metal contents by the suspension procedures together with the certified values. Very good agreement existed, demonstrating satisfactory analytical performance of the proposed procedures. Again, the statistical study using the Wilcoxon signed rank test revealed no significant difference between the results obtained by the suspension procedure and the certified values (level of significance 0.05). Values were W = –2.0, T+ = 4.0, T– = –6.0 (p = 0.875) for Cd; W = –13.0, T+ = 1.0, T– = –14.0 (p = 0.125) for Cu; and W = 9.0, T+ = 12.0, T– = –3.0 (p = 0.313) for Al.
Conclusions Acid mineralization of the products used in brewing is a difficult process that may lead to volatile analyte loss or contamination by the reagents added except when a microwave oven under pressure is used. Thus, the procedures developed, which involve direct introduction of solid products into the atomizer in the form of suspensions, have considerable advantages: total analysis time is reduced because there is no need for prior sample dissolution, and results are more accurate because analyte losses through volatilization are reduced. Preparation of suspensions of beer products in the presence of both hydrogen peroxide and nitric acid considerably reduces deposition of carbonaceous residues inside the atomizer. The procedures are rapid and accurate, and can be applied to the quality control of beer manufacturing products. Acknowledgments We are grateful to the Spanish Instituto Nacional de Investigacion y Tecnologia Agraria y Alimentaria (INIA; Madrid, Spain; Project CAL01-025) and Direccion General de Investigacion Cientifica y Tecnica (DGICYT; Madrid, Spain; Project BQU2000-0218) for financial support. References (1) Belitz, H.D., & Grosch, W. (1988) Química de los alimentos, Acribia, Zaragoza, Spain (2) USDA Nutrient Database for Standard Reference (2000) National Agricultural Library, Department of Agriculture, Washington, DC (3) Bendicho, C., & De Loos-Vollebregt, M.T.C. (1991) J. Anal. Atom. Spectrom. 6, 353–374 (4) Borriello, R., & Sciaudone, G. (1980) At. Spectrosc. 1, 131–132 (5) Donhauser, S., Wagner, D., & Jacob, F. (1987) Monatsschr. Brauwiss. 40, 247–256 (6) Ybañez, N., Navarro, A., & Montoro, R. (1989) J. Inst. Brew. 95, 257–262 (7) Wagner, H.P., & McGarrity, M.J. (1989) J. Am. Soc. Brew. Chem. 47, 68–72 (8) Li, H., & Lian, J. (1991) Fenxi Ceshi Tongbao 10, 51–55
VIÑAS ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 3, 2002 743 (9) Thalacker, R., Stelz, A., & Taschan, H. (1996) Brauwelt 136, 1813–1816 (10) Bulinski, R., Wyszograodzka-Koma, L., & Marzec, Z. (1996) Bromatol. Chem. Toksykol. 29, 167–172 (11) Matusiewicz, H., & Kopras, M. (1997) J. Anal. At. Spectrom. 12, 1287–1291 (12) Fernandes, S.M.V., Rangel, A.O.S.S., & Lima, J.L.F.C. (1998) J. AOAC Int. 81, 645–647 (13) Vela, M.M., Toma, R.B., Reiboldt, W., & Pierri, A. (1998) Food. Chem. 63, 235–239 (14) Bag, H., Lale, M., & Turker, A.R. (1999) Fresenius J. Anal. Chem. 363, 224–230 (15) Svendsen, R., & Lund, W. (2000) Analyst 125, 1933–1937 (16) Matusiewicz, H., & Mikolajczak, M. (2001) J. Anal. At. Spectrom. 16, 652–657
(17) Wyrzykowska, B., Szymczyk, K., Ichichashi, H., Falandysz, J., Skwarzec, B., & Yamasaki, S. (2001) J. Agric. Food. Chem. 49, 3425–3431 (18) Viñas, P., Pardo-Martínez, M., Campillo, N., & Hernández-Córdoba, M. (1999) J. AOAC Int. 82, 368–373 (19) Viñas, P., Pardo-Martínez, M., & Hernández-Córdoba, M. (1999) J. Anal. At. Spectrom. 14, 1215–1219 (20) Viñas, P., Pardo-Martínez, M., & Hernández-Córdoba, M. (2000) Anal. Chim. Acta 412, 121–130 (21) Halls, D.J. (1989) J. Anal. At. Spectrom. 4, 149–152 (22) Welz, B., & Sperling, M. (1999) Atomic Absorption Spectrometry, 3rd Ed., Wiley, Weinheim, Germany
