Anal Bioanal Chem (2002) 373 : 310–313 DOI 10.1007/s00216-002-1328-5
TECHNICAL NOTE
Nadica Todorovska · Irina Karadjova · Trajče Stafilov
ETAAS determination of nickel in serum and urine
Received: 13 November 2001 / Revised: 18 March 2002 / Accepted: 10 April 2002 / Published online: 23 May 2002 © Springer-Verlag 2002
Abstract Methods for the direct determination of Ni in human blood serum and urine by electrothermal atomic absorption spectrometry (ETAAS) are described. Hydrogen peroxide was proposed as matrix modifier, assisting thermal decomposition of proteins during the ashing step. A pyrolysis temperature of 1,200 °C was found to be optimal while 2,100 °C and 2,200 °C were found to be optimal atomizing temperatures for Ni in serum and urine respectively. Calibration was performed by using a calibration curve prepared with aqueous standard solutions of Ni (glycine must be used as modifier for Ni in aqueous solutions). The limits of detection, defined as the blank values plus 3 times the standard deviation of the blank values, were 0.2 µg/L for both serum and urine samples. Relative standard deviations for serum samples with concentrations of Ni in the range 0.5–2 µg/L were 10–15% and for urine samples with Ni concentrations in the range 0.5–2.5 µg/L were 8–10%. Keywords Nickel · Serum · Urine · Electrothermal atomic absorption spectrometry · Determination
Introduction The determination of Ni has been the subject of numerous investigations in view of the industrial and environmental importance of this element on the one hand, and of its biological relevance on the other, since Ni is one of the essential trace elements in the human body.
N. Todorovska Institute of Preventive Medical Care and Toxicology, Military Health Institution Center, 1000 Skopje, Macedonia I. Karadjova Faculty of Chemistry, University of Sofia, 1126 Sofia, Bulgaria T. Stafilov (✉) Institute of Chemistry, Faculty of Science, Sts. Cyril and Methodius University, POB 162, 1000 Skopje, Macedonia e-mail:
[email protected]
Electrothermal atomic absorption spectrometry (ETAAS) is one of the most frequently used techniques among the instrumental methods applied to Ni determination in blood serum and urine. Two main approaches can be distinguished in determination of Ni in blood serum and urine samples: preliminary separation and preconcentration of Ni [1, 2, 3, 4, 5, 6], or direct sample introduction into the graphite tube [7, 8, 9, 10, 11]. Separation procedures are time consuming and may introduce contamination and serious systematic errors. Although direct spectrometric determination is therefore preferable, matrix interferences were seen as limiting factors in this case. In many papers some of these problems are overcome by using prior sample dilution with water [12], Triton X-100 [13, 14], nitric acid [15, 16] or by a mixture of nitric acid and Triton X-100 [10, 17, 18]. Chemical modifiers (NH4H2PO4 [7, 10] or Pd [2, 5, 6]) have also been recommended for Ni ETAAS determination in blood serum and urine samples. The purpose of the present study was to define optimal temperature programs and suitable modifiers for interference-free direct ETAAS determination of Ni in serum and urine. In the procedure developed, hydrogen peroxide was used as the matrix modifier and glycine as the element modifier. Calibration against aqueous standard calibration curves is recommended. Analytical quality assurance was carried out by analyzing standard reference materials. The proposed and validated method was applied to the determination of Ni in the serum and urine of normal individuals, electroplating workers and patients on dialysis (serum samples only).
Materials and methods Instrumentation. The atomic absorption spectrometer Varian Spectra AA 640Z Zeeman AAS, equipped with a GTA100 graphite furnace (Varian, USA) and PSD-100 autosampler (Varian, USA), was used. Argon was applied as protective gas and 10 µL serum or urine was injected into the graphite furnace. Operating conditions were as summarized in Table 1. Reagents and samples. The standard solutions were prepared by dissolving a Merck stock solution containing 1 g/L Ni in nitrate
311 Table 1 Instrumental parameters for determination of Ni by ETAAS Parameter Wavelength Slit Lamp current Background correction Sample volume
232.0 nm 0.2 nm 5.0 mA Zeeman 10 µL
Step
Time s
Temperature °C Serum
Drying 1 2 3
Results and discussion Argon flow L/min
Urine
85 95 120
85 95 120
5 40 10
3 3 3
1,200 1,200 1,200
1,200 1,200 1,200
5 2 2
3 3 0
Atomization 7 2,100 8 2,100
2,200 2,200
2 2
0 0
Cleaning 9
2,200
2
3
Pyrolysis 4 5 6
2,100
Procedures. The serum or urine samples of 10 µL were introduced directly into the graphite furnace with an equal volume of hydrogen peroxide (30%, 10 µL as matrix modifier). The calibration curves (2–10 µg/L Ni) were prepared by using aqueous standard solutions of Ni and 1% aqueous solution of glycine introduced as matrix modifier (10 µL).
form. Hydrogen peroxide was additionally purified by ion exchange, produced in the Laboratory for High Purity Substances, University of Sofia, Bulgaria. Glycine (H2NCH2COOH) solution 1% (w/v) was prepared by dissolving glycine p.a. in doubly distilled water. All solutions were prepared fresh daily. All disposable devices were rigorously cleaned shortly before use by immersion in hot concentrated nitric acid and rinsing with doubly distilled water. Serum and urine samples were obtained from 20 volunteers presumed to be healthy, from 15 occupationally exposed electroplating workers, and in addition to this, serum samples were obtained from 27 dialyzing patients and were transferred to plastic tubes. The serum samples were collected with plastic IV kanula No. 24 (TIK, Slovenia) with injection valve. Urine samples were taken as spot samples. The samples were kept frozen until analysis.
ETAAS measurement of Ni in serum The well known and widely used procedures for Ni determination in serum need prior protein precipitation, using nitric acid [15, 16, 17, 18, 19], Triton X-100 [17, 18] or trichloracetic acid [1] and ETAAS measurement of Ni in protein-free supernatant. Experiments performed showed that direct ETAAS measurement of Ni in serum is almost impossible, due to very high values of background absorption. In the present study hydrogen peroxide was proposed as matrix modifier, assisting thermal decomposition of proteins during the ashing step. Optimal instrumental parameters were defined according to pyrolysis and atomization curves constructed for serum samples spiked with 5 µg/L Ni with hydrogen peroxide as modifier. Ashing temperatures higher than 1,100 °C and atomization temperatures lower than 2,300 °C should be used (Fig. 1). It should also be noted that even an ashing temperature of 1,200 °C is not enough high to eliminate background absorption without hydrogen peroxide as modifier. It is probable that H2O2 evolves free oxygen, thus ensuring virtually residue-free ashing of the serum. At the same time, a relatively low atomization temperature perFig. 1 a Ashing curves for Ni in serum, Tat 2,200 °C. Square 5 µg/L Ni aqueous standard, circle 5 µg/L Ni aqueous standard with glycine as modifier, diamond serum, triangle serum with H2O2 as modifier. b Atomization curves for Ni in serum, Tpr 1 200 °C. Square 5 µg/L Ni aqueous standard, circle 5 µg/L Ni aqueous standard with glycine as modifier, triangle serum, diamond serum with H2O2 as modifier
312
mits lower background absorption and more reproducible measurements. The thermal behavior of an aqueous standard solution of Ni is different. An atomization temperature of 2,200 °C is too low for complete atomization of Ni in the aqueous standard solution. Experiments performed showed that using glycine as modifier decreases the required atomization temperature for Ni (due to reducing the action of active and finely dispersed carbon formed after glycine ashing) and permits its correct ETAAS determination in aqueous solutions under the same instrumental parameters. ETAAS measurement of Ni in urine Heating programs for ashing of urine samples in order to determine Ni were developed by gradually increasing the temperature for 5 µg/L Ni spiked samples and aqueous standard solution. It was found that for spectral interference-free determination of Ni in urine ashing, temperatures of at least 1,100 °C should be used. Lower ashing temperatures lead to irreproducible signals affected by high background absorption signals. Keeping the ashing temperature fixed, the furnace was programmed for the variation of the atomization temperature. The optimal atomization temperature found was in the range 2,200– 2,300 °C (Fig. 1). Hydrogen peroxide as matrix modifier does not influence the shape of absorbance signals for Ni, but significantly decreases the value of nonspecific absorption. Calibration In order to evaluate the degree of matrix depression on atomization of Ni, several serum or urine samples were spiked with Ni in the range 2–5 µg/L, and 10 µL of these
Table 2 The ratio of the slopes of the calibration curves. bo Slope of calibration curve for aqueous standard solutions, b1 slope for calibration curve for previously spiked samples, b2 slope for calibration curve for samples spiked in the graphite furnace, sr standard deviation of the ratio of the slopes, where [sr/(b1/b0)]2= (sb1/b1)2+(sb0/b0)2 and [sr/(b2/b0)]2=(sb2/b2)2+(sb0/b0)2 Sample
b1/bo (mean±sr)
b2/bo (mean±sr)
Serum Serum in presence of H2O2 Urine Urine in presence of H2O2
0.48±0.1 0.95±0.02 0.71±0.12 0.97±0.02
0.74±0.1 0.97±0.02 0.86±0.09 0.98±0.02
Table 3 Determination of Ni in certified reference materials (five parallel determinations). The theoretical t-value for n=5 is 2.72
samples were introduced into the graphite furnace. Spiking of serum or urine samples was also performed by using the capabilities of the auto sampler, introducing into the graphite furnace 10 µL serum or urine and 10 µL aqueous standard of Ni. Matrix interferences were evaluated by the ratio of the slopes of calibration curves obtained in the presence of serum or urine and in the presence of aqueous standard solution. Results obtained are shown in Table 2 and clearly illustrate how strongly serum and urine matrices affected the degree of atomization of Ni. The depression effect is much more pronounced for previously spiked samples than for samples spiked in the graphite furnace. This means that calibration by using the capabilities of the auto sampler (injection of sample together with standard solution) is incorrect. The ratios of the slopes of the calibration graphs obtained in the presence of H2O2 as modifier are in the range 0.93–0.96, which confirms the capability of this modifier to reduce the spectral interferences and high background absorption values encountered with serum or urine matrices and at the same time to improve the degree of atomization of Ni. Therefore, with H2O2 as modifier, calibration could be performed against a calibration curve prepared with an aqueous standard solution of Ni (glycine must be used as modifier for Ni in aqueous solutions for serum samples analysis). Accuracy and precision The accuracy of the proposed procedure for ETAAS determination of Ni in serum and urine samples was checked by spike recovery experiments. Serum and urine samples were spiked with Ni in the concentration range 1–20 µg/L Ni. The recovery values obtained ranged between 95.0% and 97.0% for all spikes. The accuracy of the analytical method developed was also checked by analyzing certified reference materials for urine. Results obtained (Table 3) are in very good agreement with certified values (t-test, 95%). The limits of detection (LOD) and limits of quantification (LOQ) were evaluated on the basis of repeated analysis of blanks (10 µL of 30% H2O2). LOD and LOQ were calculated as the average Ni level in blank plus 3 times and 10 times the standard deviation of the blank, respectively. For both serum and urine samples the LOD was 0.2 µg/L and the LOQ was 0.4 µg/L. The linearity range was 0.4–20 µg/L Ni. Within-batch precision strongly depends on analyte concentration in the measuring solution and for Ni in the range 0.5–2 µg/L was 10–15% and for urine samples with a concentration of Ni in the range
Sample
Certified value µg/L
Analytical results (mean±s) µg.L–1
RSD %
tcalc, n=5
Bio-Rad LipochekUrine Metals ControlLevel 1 Lot 69031 Bio-Rad LipochekUrine Metals ControlLevel 2 Lot 69032
11.9 (7.7–16.1)
12.5±0.8
6.4
1.7
24.1 (18.1–30.1)
23.5±0.7
2.9
1.9
313
0.5–2.5 µg/L was 8–10%. The relative standard deviation for serum and urine samples with a Ni content of 2– 10 µg/L was 3–6%. Between-batch precision (calculated as the standard deviation for results obtained for parallel samples analyzed on different days) was in the range 10–15%. Application of the method The methods developed were applied to the determination of Ni in serum and urine samples obtained from 20 presumed healthy volunteers, from 15 occupationally exposed electroplating workers, and additionally serum samples were obtained from 27 patients on dialysis. In the control group were 20 healthy patients aged 19–27 years. All occupationally exposed workers were healthy and aged 35–50 years with discontinuous exposure time to Ni of 9.74±1.52 years. For the 20 non-exposed persons the value of Ni concentration in serum ranged from 0.44 to 1.52 µg/L. For the 15 occupationally exposed workers the value of Ni concentration in serum ranged from 0.51 to 1.85 µg/L. For the 27 patients on dialysis the Ni concentration in serum was in the range 1.1–7.21 µg/L. The urine samples were obtained from 10 presumed healthy volunteers, and from 14 occupationally exposed electroplating workers. Ni in urine varied from 0.7 to 1.52 µg/L for nonexposed persons and from 0.5 to 2.29 µg/L for occupationally exposed workers. Nonparametrical Mann-Whitney (U-test) was used for statistical analysis of data obtained. Significant differences were obtained (P
