Graphite furnace atomic absorption spectrometry as a routine method for the quantification of beryllium in blood and serum
© Stephan et al 2008
Received: 20 March 2008
Accepted: 02 July 2008
Published: 02 July 2008
Skip to main content
© Stephan et al 2008
Received: 20 March 2008
Accepted: 02 July 2008
Published: 02 July 2008
A routine method for the quantification of beryllium in biological fluids is essential for the development of a chelation therapy for Chronic Beryllium Disease (CBD). We describe a procedure for the direct determination of beryllium in undigested micro quantities of human blood and serum using graphite furnace atomic absorption spectrometry. Blood and serum samples are prepared respectively by a simple 8-fold and 5-fold dilution with a Nash Reagent. Three experimental setups are compared: using no modifier, using magnesium nitrate and using palladium/citric acid as chemical modifiers.
In serum, both modifiers did not improve the method sensitivity, the optimal pyrolysis and atomization temperatures are 1000°C and 2900°C, respectively. In blood, 6 μg of magnesium nitrate was found to improve the method sensitivity. The optimal pyrolysis and atomization temperatures were 800°C and 2800°C respectively.
In serum, the method detection limit was 2 ng l-1, the characteristic mass was 0.22 (± 0.07) pg and the accuracy ranged from 95 to 100%. In blood, the detection limit was 7 ng l-1, the characteristic mass was 0.20 (± 0.02) pg and the accuracy ranged from 99 to 101%.
Beryllium is the 35th most abundant element in the earth's crust, with an average of 6 mg kg-1 . It has unique physical and chemical properties that improves the characteristics of alloys producing greater tensile strength, high electrical and thermal conductivity, along with good corrosion and fatigue resistance . Beryllium and its metal alloys have been widely used for electrical equipment, electronic instrumentation, structural components for aircraft, missiles, satellites and nuclear reactors [3, 4].
Beryllium and its compounds are very toxic. They can cause chronic beryllium disease (CBD), a lung disorder initiated by an electrostatic interaction with the MHC class II human leukocyte antigen (HLA) . Recent molecular epidemiological studies found a significant correlation between the risk of developing CBD and the predicted surface electrostatic potential of the HLA-DP alleles, suggesting that individuals who carry the most negatively charged alleles are at greater risk of beryllium sensitization and CBD . Because of these findings, increased research efforts are being targeted towards the development of a CBD treatment by chelation therapy [7, 8]. For that purpose, a routine method is required to analyse beryllium in micro samples of biological fluids with high sensitivity, accuracy, and low detection limits.
The commonly used methods for the determination of beryllium in tissues and urine include fluorometric [9, 10], gas chromatographic [11–14], atomic emission [15–17], inductively coupled plasma mass spectrometry (ICP-MS) [18–21] and atomic absorption [22–25]. Fluorometric methods are sensitive (Be detection limit of 5 ~ 10 μg L-1)  but require extensive sample preparation, derivatization and relatively large sample volumes. Gas chromatographic methods have sufficient sensitivity and selectivity when coupled with mass spectrometry (Be detection limit of 30 ~ 40 μg L-1) , but they require a laborious, multi-step derivatization. Atomic emission, either arc emission or inductively coupled plasma emission are sensitive and accurate (Be detection limit of 0.2 ~ 1 μg L-1) , but often require large sample volumes and multi-steps preparation often including digestion. Inductively coupled plasma mass spectrometry (ICP-MS) recently became the methode of choice for biomonitoring trace elements in biological matrices (blood, plasma, urine, etc). It is a multi-elementary method that offer outstanding sensitivity, accuracy and detection limit in the low ng L-1 with a reported beryllium detection limit in the range of 5 to 20 ng L-1 depending on the matrix . As for the atomic absorption methods, graphite furnace atomic spectrometry (GF-AAS) with a suitable chemical modifier (such as, lutetium , ammonium 12-molybdophosphate , palladium  and magnesium nitrate ) has been widely used for the analysis of beryllium in urine, environmental samples and tissue digests with detection limits in the range of 2 to 20 ng L-1. Limited work is published on the determination of beryllium in human serum, blood or biological fluids. The objective of this study is to develop a simple procedure for the direct determination of beryllium in undigested micro quantities of human blood and serum using GF-AAS as a routine analytical method. The analyses were performed by simple 5-fold and 8-fold dilution of the serum and blood respectively with a Nash reagent (NR) containing nitric acid, ammonium hydroxide, Triton X-100, antifoam B and EDTA [21, 29]. We also compared three different experimental setups: using no modifier, using magnesium nitrate and using palladium/citric acid (reduced palladium) as chemical modifiers in order to study their effect on beryllium quantification in such complex matrices. We selected magnesium nitrate, known to improve beryllium sensitivity by a factor of 1.5 to 2 in matrices rich in halide ions (e.g. urine and environmental samples [24, 27]) and reduced palladium, one of the commonly-used modifiers in GF-AAS analysis known to stabilize the atomization process regardless of the matrix complexity were evaluated [30–32]. Optimal pyrolysis and atomization conditions, detection limits and characteristic mass (the mass of element generating 0.0044 absorbance) were determined in order to develop a rapid and precise method for beryllium analysis in human blood and serum. The accuracy of the analytical method is tested with a control sample (Seronorm trace elements whole blood (STEWB) Level 2).
Beryllium contents in blood and serum of ten individuals. Averages of triplicate measurements with standard deviations are given.
Blood (μg L-1)
Serum (μg L-1)
0.71 ± 0.09
0.49 ± 0.08
0.64 ± 0.09
0.43 ± 0.09
0.74 ± 0.06
0.45 ± 0.08
0.67 ± 0.08
0.40 ± 0.08
0.57 ± 0.06
0.46 ± 0.09
0.68 ± 0.04
0.40 ± 0.05
0.64 ± 0.08
0.41 ± 0.04
0.67 ± 0.09
0.42 ± 0.04
0.48 ± 0.06
0.43 ± 0.03
0.55 ± 0.07
0.42 ± 0.06
0.63 ± 0.08
0.43 ± 0.03
The method detection limit was determined as the concentration corresponding to three times the standard deviation of 8 replicates of the lowest standard (0.05 μg L-1). We found a beryllium detection limit of 7 ng L-1 in blood. The beryllium characteristic mass was found to be 0.20 (± 0.02) pg, of the same order of magnitude as that reported by the manufacturer: 0.5 pg determined in 0.1% nitric acid matrix. It was calculated as the mass of beryllium in 8-fold diluted blood that yields an absorbance equal to 0.0044 (1% absorption) in the peak height mode, with a new pyrolytic coated graphite tube and under the optimal furnace temperature program . This similarity in the beryllium characteristic mass reassures us that our experimental conditions have eliminated signal suppression for the blood matrix. The accuracy of the method was verified by analysing the STEWB level 2 control sample with a beryllium concentration of 5.9 (± 0.5) μg L-1. Accuracy values of four different control samples prepared by 10 or 20-fold dilution of STEWB level 2 with NR were found to be 100 (± 1) %. For the actual 10 samples we tested, neither sex nor smoking habit had a statistically significant influence on the concentration of beryllium in blood (T-test, p > 0.05), but for the two smokers, data points were higher. A larger dataset would be necessary to properly explore exposure questions and may yield interesting differences.
We found that beryllium concentrations in the serum of ten non-exposed individuals, analysed under optimal temperature, using aqueous beryllium standards without any chemical modification, varied from 0.40 to 0.49 μg L-1 with an average of 0.43 ± 0.03 μg L-1 (± SD) (Table 1). The beryllium detection limit was found to be 2 ng L-1 determined as the concentration corresponding to three times the standard deviation of 12 replicates of the lowest calibration standard (0.05 μg L-1). The beryllium characteristic mass in 5-fold diluted serum was found to be 0.22 (± 0.07) pg, slightly lower than that reported by the manufacturer: 0.5 pg determined in 0.1% nitric acid matrix . This also confirms that no signal suppression was caused by the serum matrix under our experimental conditions. The accuracy of the method was verified using the same control material (STEWB level 2). We obtained an accuracy of 97.5 (± 2.5)% on four different control samples. As observed earlier for blood, no significant influence of beryllium concentration in serum is attributed to sex or smoking (T-test, p > 0.05).
We found a significant difference between the beryllium concentration in blood and serum (p < 0.05). We measured higher beryllium concentration in blood than in serum. On average, we noted a 30 ± 10% increase in the blood beryllium concentration over serum. These findings suggest that roughly two thirds of the beryllium concentration in the blood stream is found in the serum and only one third is attached to the blood clot. Despite the complexity of serum and blood matrices, no signal suppression was noticed following our optimal conditions and our beryllium detection limits (2 to 7 ng L-1) using the Varian AA280Z Zeeman atomic absorption spectrometer are among the lowest reported to date.  and , reported a detection limit of 370 ng L-1 and 4.3 ng L-1 respectively for beryllium in urine by graphite furnace atomic absorption spectrometry. In a similar work, , and , reported a detection limit of 50 ng L-1 for beryllium in urine. In water,  found 2.3 ng L-1 while , found 10 ng L-1. The proposed method is ideally suited to evaluate occupational exposure and other factors contributing to beryllium risks.
Instrument conditions and furnace programme for the determination of beryllium (argon flow rate was set at 0.3 L min-1, except during atomization)
All reagents are of analytical grade, unless otherwise stated. Antifoam B silicone emulsion (J.T. Baker, NJ, USA), ammonium hydroxide (certified A.C.S. Plus, Fisher scientific, NJ, USA), beryllium standard solution (Specpure, Alfa Aesar, MA, USA), palladium matrix modifier (Sigma-Aldrich, USA), ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA) (Fluka Chemika, Switzerland), nitric acid (trace metal grade, Fisher Scientific, Ontario, Canada), Triton X-100 (Acros, NJ, USA), citric acid and magnesium nitrate were purchased from Fisher-scientific (Ottawa, Ont, Canada). Seronorm trace elements whole blood (STEWB) Level 2 (Ref # 201605; Lot # 0503109) was purchased from SERO (Billingstad, Norway).
Blood and serum samples were diluted with a Nash Reagent (NR) prepared weekly containing 5% (v/v) nitric acid, 5% (v/v) of ammonium hydroxide, 0.2% (v/v) Triton X-100, 0.2% (v/v) antifoam B and 0.5% (w/v) of EDTA. A 50 μg L-1 working beryllium(II) solution was prepared by dilution of the beryllium stock solution (1000 μg L-1) in 2% (v/v) HNO3. Beryllium standard solution was prepared daily by dilution of the working beryllium(II) solution with the Nash Reagent to give a final concentration of 0.5 μg L-1. A series of palladium working solutions containing 300, 600, 1000, 1200, 1500 and 2000 mg L-1 were prepared by dilution of the stock solution (10 000 mg L-1) in 2% (v/v) HNO3 and 2% (w/v) citric acid aqueous solutions. Another series containing 300, 600, 1000, 1200, 1500 and 2000 mg L-1 magnesium nitrate was prepared by dissolving the appropriate mass in 2% (v/v) HNO3 aqueous solution. A 10-μL volume of the chemical modifier or Nash Reagent (when no modifier is used) was co-injected with 20-μL of the sample into the furnace.
Blood and serum samples of 10 individuals (8 females and 2 males) were collected respectively in BD Vacutaine (sodium Heparin) and BD Vacutaine (SST) respectively (BD Franklin lakes (NJ, USA)). All subjects live in the region of Montreal, Quebec, Canada with no previous history of occupational exposure to beryllium, from questionnaire-based interviews. Blood was diluted 8-fold and serum 5-fold with the Nash Reagent. Triplicates of blood and serum samples were refrigerated at 4°C until analysis. All samples were analysed on the day of collection.
We thank the Natural Sciences and Engineering Research Council of Canada and the Institut de Recherche Robert-Sauvé en Santé et en Sécurité du Travail for financial support.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.