Return to Document Menu

Determination of Lead and Cadmium in Food Products by Graphite Furnace Atomic Absorption Spectroscopy C. Blake and B. Bourqui Nestlè Research Centre, Quality and Safety Assurance Department P. O. Box 44, 1000 Lausanne 26, Switzerland INTRODUCTION Overall exposure to lead and cadmium is a public health concern. The lead content in food products has been gradually reduced due to the phasing out of lead-soldered cans as well as the use of unleaded gasoline (11). However, cadmium levels in the environment appear to be increasing. Burgatsacaze et al. (12) recently reviewed the role of cadmium in the food chain. Various international organizations, e.g., Codex Alimentarius, the European Union (E.U.) (1,2), are debating and reviewing the maximum allowable concentration of lead and cadmium in raw materials and food products. Future norms will set the limits of metals concentration, particularly for lead, which will be rather low. This is expected to be an important factor in international trading, i.e., grain exporters must increasingly be able to certify that the grain shipments are in compliance with regulatory requirements for toxic metals content (9). The technique most commonly used by Nestle laboratories for the determination of lead and cadmium is graphite furnace atomic absorption spectroscopy (GFAAS). The current Laboratory Instructions (LI) were published in 1989 and have been implemented in many regional laboratories. These LI are similar to methods published in the German food analysis handbook (3) and by ISO (4–7). Impending EU legislation for toxic metals determination will set strict method performance criteria (1,2), based on the criteria

AS

Atomic Spectroscopy Vol. 19(6), November/December 1998

ABSTRACT Two sample wet ashing techniques for mineralization of food products and raw materials were evaluated using high pressure ashing and microwave digestion with Teflon vessels, fitted with quartz inserts. Similar accuracy and precision for the determination of lead and cadmium were obtained when analyzing a range of certified food reference materials by graphite furnace atomic absorption spectroscopy (GFAAS). The high pressure asher method is preferred due to the higher sample throughput, selection of 14 or 21 tubes, depending on the type of heating block used. The limits of quantification for lead and cadmium by GFAAS with Zeeman correction were improved using end-capped graphite tubes and an electrodeless discharge lamp in place of a hollow cathode lamp. A single matrix modifier (magnesium nitrate and ammonium dihydrogen phosphate) was found to be suitable for the determination of both lead and cadmium. The limits of detection and repeatability for Pb and Cd are close to the requirements currently being proposed by the European TC 275 working group for heavy metals methodology.

described in ISO 3535–1993. Thus, the current LI will need to be updated to meet future norms. In the past few years, a number of instrumental developments have contributed to providing more reliable results and higher detection limits for trace determination of

207

lead and cadmium by GFAAS. These include (a) improved electrodeless discharge and hollow cathode lamps for increased light output; (b) transversely heated graphite tubes with end caps for higher sensitivity; and (c) improved wet ashing sample preparation techniques, e.g., microwave digestion and high pressure ashing. These aspects have been evaluated in the current study. EXPERIMENTAL Instrumentation Sample preparation HPA-S High Pressure Asher™ system (Perkin-Elmer/Paar), equipped with stainless steel heating blocks for 14 or 21 tubes and Suprasil mineralization tubes (15 mL). Microwave digestion system, MLS 1200 Mega (Milestone), with a temperature control system. The high pressure (HPV 80) vessels are equipped with QS 50 graduated quartz liners with caps. The mineralization tubes were decontaminated before use in a nitric acid vapor decontamination system (Trabold Ltd, Berne, Switzerland). Atomic absorption instrumentation A Perkin-Elmer Model 4100 ZL atomic absorption spectrometer was used, equipped with transversely heated graphite furnace and Zeeman background correction, AS-70 autosampler, closed-circuit cooling system, fume extraction system, and System 2 electrodeless discharge lamp power supply .

Transversely heated, pyrolyzed graphite tubes with integrated platform, and transversely heated pyrolyzed graphite tubes with end caps were used. Reagents All solutions were prepared in polypropylene volumetric flasks, using ultra-pure water, prepared with a Barnstead Nanopure system. Nitric acid: Suprapure (Merck). Hydrogen peroxide: Analytical grade (30%), (Merck).

High Pressure Ashing Samples of 300 mg each were weighed into decontaminated 15mL Suprasil tubes. Two mL of concentrated double-distilled nitric acid was added. The tubes were sealed with Teflon® tape, capped, and then wet-ashed in the HPA-S. A typical temperature program used is shown in Figure 1. The acid solution was diluted to 10 mL with water. Further dilutions were made with 10% (m/v) nitric acid solution, if required to be within the calibration range.

Lead and cadmium stock solutions, 1 g/L (Spex). Cadmium and lead working solutions were prepared by dilution of the cadmium and lead stock solutions with 10% (v/v) nitric acid. Matrix modifier: Various mixtures of ammonium dihydrogen phosphate (NH4H2PO4) and magnesium nitrate [Mg (NO3)x.6H20] in 10 % nitric acid solution. Reagents were of Suprapure quality (Merck). Reference Materials A range of reference materials with certified lead and cadmium content was used for method evaluation. These products were obtained from IRMM, NIST, IAEA, and NRCC. Nestec reference materials (cereals with milk) of known lead and cadmium content were also used.

Fig. 1. HPA-S temperature program.

Sample Preparation All sample weighings were carried out in a class 100 laminar flow cabinet (Skan Model EVZ 180). This cabinet was installed in a clean-air room (class 1000 air quality) under positive air pressure with a filtered air inlet. The samples were wet ashed using the following two methods: Fig. 2. Program for Milestone MLS.

208

Microwave Digestion Samples of 300 mg each were weighed into a Milestone QS5 quartz insert with 3 mL concentrated double-distilled nitric acid. The insert was then introduced into a Milestone HPV 80 Teflon vessel. One mL hydrogen perioxide (30 %) and 4 mL water were added inside the Teflon vessel, but at the exterior of the insert. The microwave vessel was closed with its cap. Six vessels were placed into the rotor and heated in the microwave oven according to the temperature program shown in Figure 2.

Vol. 19(6), Nov./Dec. 1998 The rotor was removed from the microwave oven and allowed to cool to room temperature. The vessels were carefully opened in a fume cupboard. The quartz inserts were removed with the Tefloncoated tweezers provided. The inner wall of the quartz insert was rinsed with de-ionized water and made up to the 10-mL mark with water. Further dilutions were made as required with 10% (m/v) nitric acid solution. All subsequent dilutions were prepared in polypropylene volumetric flasks. GFAAS Determination of Lead and Cadmium The GFAAS operating parameters for the determination of lead and cadmium are listed in Tables I and II. Transversely heated pyrolyzed graphite tubes were used for the lead and cadmium determinations. The GFAAS was calibrated by the external standards method with a zero blank and five standard concentrations. All analyses were performed by triplicate firings. During an analytical series, a mid-range QC standard solution was injected every 10 analytical sample solutions to verify the calibration slope. RESULTS AND DISCUSSION Graphite Furnace Method Development Optimization of matrix modifier The current LI for lead requires the use of ammonium dihydrogen phosphate as the matrix modifier. The LI for cadmium requires a palladium matrix modifier. The main purpose of these modifiers is to stabilize the element during the graphite furnace cycle and to permit increases in the charring and atomization temperatures. This allows a better separation of the element from interferences. However, since two different modifiers are used with the current LI, a separate graphite tube is required for each element. In order

to simplify these procedures, a mixed modifier has been evaluated for the determination of both elements. The mixed matrix modifier (magnesium nitrate and ammonium dihydrogen phosphate) has been reported in several publications in different ratios. In the present study, several concentrations of the mixed matrix modifier were evaluated:

1. Magnesium nitrate 0.06% and ammonium dihydrogen phosphate 0.5%. 2. Magnesium nitrate 0.6% and ammonium dihydrogen phosphate 0.5%. 3. Magnesium nitrate 0.10% and ammonium dihydrogen phosphate 1.3% (9). 4. Magnesium nitrate 0.20% and ammonium dihydrogen phosphate 2.0% (8).

TABLE I. Instrumental Parameters for the Determination of Pb Furnace Time/Temperature Program Parameter Dry 1 Dry 2 Char 2 Atomize Clean o Temp ( C) 100 130 750 1600 2300 Ramp (sec)

2

20

10

0

5

Hold (sec)

20

60

25

5

3

250

250

250

0

250

Argon gas flow

(mL/min) Wavelength: Lamp: Slit Width: Read time: Signal measurement: Graphite tube:

Peak area Pyrolytic graphite, end-capped

Calibration Standards

I.D.

Calibration Blank Std 1 Std 2 Std 3 Std 4 Std 5 Reslope Standard

283.3.nm Electrodelss discharge 0.7 nm 5s

STD 1 STD 2 STD 3 STD 4 STD 5 STD 3

Concn. (µg/L)

1.0 2.5 5.0 7.5 10.0

Pipette speed: 100% Injection temperature: 20oC

Volume (µL)

Diluent volume (µL)

Modifier volume (µL)

20 2 5 10 15 20

18 15 10 5 0

5 5 5 5 5 5

15

5

5

Calibration type: Linear Note: The calibration range may be increased to 20 µg/L; for higher concentrations, a non-linear calibration curve was obtained.

209

The best results obtained in terms of peak profile was with a mixture of 0.6% magnesium nitrate and 0.5% ammonium dihydrogen phosphate; although the non-specific background was somewhat higher than with modifier 1. Modifier 1 also gave good results. Modifiers 3 and 4 were found to give very high non-specific backgrounds and were not further evaluated.

With respect to the GFAAS temperature programs, the final charring and atomization temperatures adopted are listed for lead and cadmium in Tables I and II. For lead, the atomization temperature was fairly critical and the peak shape changed dramatically in the range from 1400oC to 1600oC. This temperature needs to be optimized carefully.

TABLE II. Instrumental Parameters for the Determination of Cd Furnace Time/Temperature Program Parameter Dry1 Dry2 Char Atomize Clean Temp (oC) 100 130 600 1400 2300 Ramp (s) 2 20 15 0 5 Hold (s) 20 60 20 5 3 Argon gas flow (mL/min) 250 250 250 0 250 Wavelength: Lamp: Slit width: Read Time: Signal measurement: Graphite tube: Calibration Standards

Injection temp. 20oC Pipette speed 100%

228.8 nm Hollow cathode 0.5 nm 5s Peak area Pyrolytic graphite, end-capped

I.D.

Calibration blank Std 1 STD 1 Std 2 STD 2 Std 3 STD 3 Std 4 STD 4 Std 5 STD 5 Reslope Standard STD 4

Concn. (µg/L)

Volume (µL)

Diluent volume (µL)

0.5 1.0 2.0 3.0 5.0

20 2 4 8 12 20

18 16 12 8 0

3.0

12

8

Modifier volume (µL) 5 5 5 5 5 5 5

Influence of end-capped graphite tubes and lamps The main difficulty in determining lead by GFAAS is to obtain sufficient sensitivity. The use of end-capped graphite tubes over the standard graphite tubes resulted in a significant increase in signal (by a factor of 1.5). In addition, the use of an electrodeless discharge lamp instead of a hollow cathode lamp also resulted in an increase in signal due to increased light intensity. The combined effect of the end-capped tube and an electrodeless discharge lamp (EDL) over a standard graphite tube and hollow cathode lamp (HCL) resulted in an increase in sensitivity by about a factor of 2. Figure 3 illustrates the difference in calibration slope for lead. However, the major improvement was in the improved repeatability of measurements at low lead concentrations below 2.0 ng/mL. For cadmium, end-capped tubes were used with a hollow cathode lamp light source. Some further improvement in sensitivity may be obtained with an electrodeless discharge lamp in place of the hollow cathode light source. Evaluation of High Pressure Ashing and Microwave Digestion Techniques A range of different techniques has been described in the literature (10) for the sample preparation of foods and raw materials prior to GFAAS determination of lead and cadmium. For this study, the method performance of two sample preparation techniques for different reference materials with certified lead and cadmium content was evaluated: 1. High pressure asher (HPA-S) with new 15-mL Suprasil tubes. 2. Microwave digestion (MDS) with Teflon vessels fitted with graduated quartz inserts.

Calibration type: Linear

210

Vol. 19(6), Nov./Dec. 1998 The limits of quantification for each element based on the analysis of certified reference materials using the stated equipment were: Cadmium = 10 µg/kg Lead

= 15 µg/kg

CONCLUSION Two sample wet ashing techniques, high pressure ashing and microwave digestion with Teflon vessels fitted with quartz inserts, were evaluated.

Fig. 3. Comparison of different GFAAS conditions on calibration line.

HPA-S System For the HPA-S technique, good recoveries of lead and cadmium were obtained for various reference materials (Tables III and IV) and two infant cereals (Nestec reference products, MET) (Tables V and VI). The repeatability of the results obtained was also good and within or close to the range of the certified values. Microwave Digestion System (MDS) The main disadvantages of this system are that a higher volume of nitric acid is required (5–6 mL) and that the final volumes of the analytical solutions are often high (from 25 or 50 mL). This leads to lower sensitivity owing to the large dilution factor. Thus, the use of Milestone QS 50 graduated quartz inserts, which fit inside the Teflon vessels, was evaluated. Three mL of nitric acid was used, most of which was consumed during the wet ashing step. The final volume of the analytical solution, after dilution with water, was 10 mL. Thus a significant increase in sensitivity was obtained due to the decrease in total volume.

The accuracy of the results obtained by MDS in the present study was, in general, similar to that of the HPA-S technique, with occasional values being slightly below the certified reference values for lead (Tables III–VI). The open quartz-tube system of the MDS may lead to losses if the digestion unit is not allowed to cool adequately after completion of wet ashing. The repeatability of the results was similar for both methods (see Tables III–VI and VII). An overview of the relative standard deviations (%RSD) for lead and cadmium is shown in Table VII. The RSD was below 25% for concentrations 100 µg/kg. This is quite acceptable for the trace determination of lead and cadmium. Limits of Detection and Quantification An important aspect of the method performance evaluation is the calculation of the limits of detection and the limit of quantification. The limits of detection based on the repeated analysis of blank solutions were calculated to be: Cadmium = 3 µg/kg Lead = 5 µg/kg

211

Similar accuracy was obtained with the two methods for both lead and cadmium when analyzing a range of certified food reference materials. The HPA-S method is preferred due to the higher sample throughput (14 or 21 tubes, depending on the heating block used). The limit of quantification for lead by GFAAS resulted in an improvement by a factor of 2 with respect to the current LI by using end-capped graphite tubes and an electrodeless discharge lamp in place of a hollow cathode lamp. The limit of quantification of cadmium was also improved by the use of the end-capped tubes. Further improvements in sensitivity may be obtained for cadmium by using an electrodeless discharge lamp as thelight source. A single matrix modifier (magnesium nitrate and ammonium dihydrogen phosphate) is suitable for the determination of both lead and cadmium. The limits of detection and the repeatability for lead and cadmium are close to the values currently being proposed by the European TC 275 working group for heavy metals methodology.

TABLE III Lead: Comparison of Results for Samples Prepared by HPA-S and Microwave Digestion Product Reference HPA-S MLS Lead Lead Lead (µg/kg) (µg/kg) (µg/kg) Bovine Muscle NIST 8414 380 ± 240 326 ± 13 444 ± 2 n=6 n=6 Brown Bread BCR 191 187 ± 14 199 ± 39 194 ± 10 n=6 n=6 Corn Bran NIST 8344 140 ± 34 136 ± 2 146 ± 5 n=9 n=6 Dogfish Muscle NRCC, DOLT-2 220 ± 2 243 ± 6 167± 3 n=6 n=9 Milk Powder IAEA, A11 54 ± 25 57 ± 14 55 ± 5 n=6 n=3 Non-fat Milk Powder NIST 1549 19 ± 0.3 21.5 ± 0.6 15.0 ± 0.6 n=6 n=9 Skim Milk Powder BCR 150 1000 ± 40 1022 ± 30 910 ± 13 n=3 n=6 Whole Meal Flour BCR 189 379 ± 12 387 ± 20 388 ± 58 n=6 n=6 Whole Egg Powder NIST 8415 61 ± 12 66 ± 9 55 ± 5 n=6 n=6 TABLE V Lead: Comparison of Results for Infant Cereals Prepared by HPA-S and Microwave Digestion Product Reference HPA-S MLS Lead Lead Lead (µg/kg) (µg/kg) (µg/kg) Infant Cereal Met – 5 277 ± 52 251 ± 7 256 ± 13 n=6 n = 10 Infant Ceral Met – 6

826 ± 52

791 ± 21 n=6

858 ± 32 n=3

TABLE IV Cadmium: Comparison of Results for Samples Prepared by HPA-S and Microwave Digestion Product Reference HPA-S MLS Lead Lead Lead (µg/kg) (µg/kg) (µg/kg) Bovine Muscle NIST 8414 13 ± 11 14 ± 2 11 ± 1 n=6 n=6 Brown Bread BCR 191 28.4 ± 14 28 ± 2 30 ± 8 n=6 n=6 Corn Bran NIST 8344 12 ± 5 9±1 11 ± 1 n=9 n=6 Dogfish Muscle NRCC, DOLT-2 20,800 ± 500 20,247 ± 938 27,700± 560 n=6 n=9 Skim Milk Powderr BCR 150 22 ± 14 20 ± 1 21 ± 3 n=6 n=6 Total Diet NIST 1548 28 ± 4 28 ± 3 24 ± 3 n=6 n=6 Whey Powder IAEA, 155 16 ± 3.5 16 ± 3 18± 3 n=3 n=6 Whole Meal Flour BCR 189 71 ± 3 74 ± 2 58 ± 11 n=6 n=3 TABLE VI Cadmium: Comparison of Results for Infant Cereals Prepared by HPA-S and Microwave Digestion Product Reference HPA-S MLS Lead Lead Lead (µg/kg) (µg/kg) (µg/kg) Infant Cereal Met – 5

127 ± 22

133 ± 4 n=6

TABLE VII Range of Repeatability Values (%RSD) from the Various Reference Materials Range of Pb Pb Cd Cd Pb or Cd RSD (%) RSD (%) RSD (%) RSD (%) (µg/kg) HPA-S MLS HPA-S MLS 10 –99 3 – 25 4–9 3 – 19 9 - 25 100 – 1000+ 1.5 – 20

212

127 ± 2 n=6

0.5 – 15

2

5

Vol. 19(6), Nov./Dec. 1998 REFERENCES 1. European Commission, “Draft: Commission regulation setting maximum limits for certain containments in foodstuffs, amending commission regulation (EC) 194/97 of 31 January 1997 setting maximum limits for certain contaminants in foods-maximum limits for lead and cadmium in foodstuffs.” European commission III/5125/95 Rev. 3 (March 1997). 2. European Community, “Commission decision 90/515/EEC of 26 September, 1990.” Off. J. Eur. Commun. 33 (L286), (1990). 3. LMBG, “Bestimmung von Spurenelementen in Lebensmitteln. Teil 3: Bestimmung von Blei, Cadmium, Chrom und Molybdän mit der Atomabsorptionspektrometrie (AAS) im Graphitrohr.” Amtliche Sammlung von Untersuchungsverfahren nach Paragraph 35 LMBG, (August, 19/3, 1993).

4. ISO, “Fruits, vegetables and derived products. Determination of lead content. Flameless atomic absorption spectrometric method,” ISO 6633 (1984). 5. ISO, “Fruits, vegetables and derived products. Determination of cadmium content. Flameless atomic absorption spectrometric method.” ISO 6561 (1983). 6. ISO, “Starch and derived products. Heavy metal content. Part 3. Determination of lead by atomic absorption spectrophotometry with electrothermal atomization.” ISO Norm 11212–3 (1997). 7. ISO, “Starch and derived products. Heavy metal content. Part 4. Determination of cadmium by atomic absorption spectrophotometry with electrothermal atomization.” ISO Norm 11212–4 (1997). 8. G. Ellen and J.W. Van Loon, Food Addit. Contam. 7 (2), 265 (1990).

213

9. E.J. Gawalko,T.W. Nowicki, J. Babb, and R. Tkachuk, J. AOAC Int. 80 (2), 379 (1997). 10. C.J. Blake, “Analysis of lead and cadmium in foods and raw materials – a literature review.” R&D Note No. QS-RN 970055 (1997). 11. P.M. Bolger et al., Food Addit. Contam. 15 (1), 53 (1996). 12. V. Burgatsacaze, L. Craste, and P. Guerre, Revue de medicine Veterinaire 147 (10), 671 (1996).