Anal. Chem. 2001, 73, 4319-4325
Direct and Simultaneous Determination of Copper, Chromium, Aluminum, and Manganese in Urine with a Multielement Graphite Furnace Atomic Absorption Spectrometer Ting-Wan Lin and Shang-Da Huang*
Department of Chemistry, National Tsing Hua University, Hsinchu, 300, Taiwan
A simple method was developed for the direct and simultaneous determination of copper, chromium, aluminum, and manganese in urine using a multielement GFAAS (Perkin-Elmer SIMAA6000). Pd was used as the chemical modifier along with a special purge gas (5% H2 in Ar). A simple calibration curve method can be used (with 1:1 dilution). A standard reference material (Seronorm Trace Elements Urine) was used to find the optimal temperature program and to confirm the accuracy of the technique. The analyzed values were within 90110% of the certified values. The relative standard deviations were 1.7, 1.5, 1.6, and 1.5% for these four elements and the detection limits were 0.08 µg L-1 for Cu, 0.05 µg L-1 for Cr, 0.06 µg L-1 for Al, and 0.06 µg L-1 for Mn. The recoveries of Cu, Cr, Al, and Mn from real urine samples were 100 ( 5%, except for Cu (80%). The found values of Cu, Cr, Al, and Mn in a real urine sample were 14.3, 0.78, 18.9, and 0.06 µg L-1, respectively. Scanning electron micrographs were used to investigate the physical form of Pd on the surface of the platform in the graphite furnace. Use of 5% H2 in Ar as the purge gas resulted in smaller and more uniformly distributed Pd particles (Pd particle diameters 0.4-0.6 µm using 5% H2 in Ar compared to 0.4-1.2 µm using pure Ar), increasing the effect of the Pd chemical modifier and promoting the efficiency of atomization. There are a lot of elements recognized as essential for human or animal life. Most of them are part of metalloenzymes and participate in biological functions, such as oxygen transport, freeradical scavenging, structural organization of macromolecules, and hormonal activity. The significance of trace elements in human health and disease has been reviewed in numerous scientific publications. Copper is an essential element whose role is complex in many body functions such as hemoglobin synthesis (in Fe utilization and Hb regeneration), connective tissue development, normal function of the central nervous system, and oxidative phosphorylation.1 More then 35 Cu-containing proteins and pigments have been isolated from tissues.1 The enzyme activity of these Cu (1) Vuori, E.; Huunan-Seppala, A.; Kilpio ˜, J. O. Scand. J. Work Environ. Health 1978, 4, 167. 10.1021/ac010319h CCC: $20.00 Published on Web 08/01/2001
© 2001 American Chemical Society
proteins may be disturbed in an early stage of Cu deficiency, which is associated with the following anomalies: loss of weight; microcytic hypochromic anemia; gastrointestinal disorders; impaired reproduction; and impeded parturition.2 Excessive Cu intake leads to accumulation of the metal in liver cells and hemolytic crisis, jaundice, and neurological disturbances.3 Chromium(III) is an essential (but also toxic) trace element for man, required for the maintenance of normal glucose, cholesterol, and fatty acid metabolism.4 Cr(III) plays a role in various enzyme reactions (e.g., thermoplastic and β-glucoronidase activity).4 Cr deficiency in animals is associated with impaired glucose tolerance, growth retardation, corneal opacity, necrotic liver degeneration, and neuropathy. Cr(VI) is highly toxic. Its acute toxic effects include an immediate cardiovascular shock and later effects on kidney, liver, NS, and blood-forming organs.4 Other clinical subacute and mainly chronic manifestations of exposure to Cr compounds are irritation of skin, typical bronchial asthma, and renal and liver dysfunction. Aluminum is the most abundant metal in the lithosphere. It is considered a nonessential trace element of low toxicity in the healthy human.5 Biological essentiality is found in brain development and activity as well as in nerve conductivity.6 Environmental Al is probably harmless for healthy humans, but in patients with renal failure, even a moderate intestinal absorption of Al may produce toxic effects.5 Toxic concentrations of Al are supposed to have ethological significance in primary degenerative dementia (Alzheimer’s disease or senile dementia), in dialysis dementia, and in dialysis-related osteomalacia. Increased Al hair levels have been found in 12-18-year-old boys with severe emotional problems.7 Manganese is also recognized as both an essential and a neurotoxic element. In normal conditions, it plays an important role in bone and tissue formation (normal growth), normal (2) Fischer, G. L. Sci. Total Environ. 1975, 4, 373. (3) Pethes, G. The need for trace element analyses in the animal sciences. In Element Analysis of Biological Materials; IAEA: Vienna, 1980; p 3. (4) Langård, S. Chromium. In Metals in the Environment; Waldro, H. A., Ed.; Academic Press: London, 1980; p 111. (5) Wawschinek, O.; Petek, W.; Lang J.; Pogglitsch, H.; Holzer, H. Mikrochim. Acta 1982, 1, 335. (6) Norseth, T. Aluminum. In Handbook on the Toxicology of Metals; Friberg, L., Nordberg, G. F., Vouk, V. B., Eds.; Elsevier/North-Holland: Amsterdam, 1979; p 275. (7) Rees, E. L. J. Orthomol. Psychiatr. 1979, 8, 37.
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001 4319
reproductive functions, and carbohydrate and lipid metabolism.8 The first Mn toxic symptoms may appear after durations of exposure ranging from 7 months to 20 years.9 Symptoms of central nervous system damage can be divided into three stages: the first includes asthenia, sleep disturbances, muscular excitability, and clumsiness of movement; the second, speech disorders, difficulty in walking, and exaggeration of reflexes. Fully developed clinical symptoms are maniacal or depressive psychosis and a class syndrome resembling Parkinsonism.10 A chronic nonspecific lung disease and Mn pneumonia may be observed as well.11 Hence it is clear that biomonitoring of body fluids (especially urine and blood serum) for Cu, Cr, Al, and Mn is essential for the control of nutritional deficiencies and perhaps for preventing their toxic effects in cases of environmental and occupational exposure. The levels of Cu, Cr, Al, and Mn reported in human urine are 6.1-30.3 µg of Cu/L, 0.20-10.72 µg of Cr/L, 3.5-31 µg of Al/L, and 0.08-2.67 µg of Mn/L, respectively.12 The concentration ranges found for each element vary due to sample contamination during sample collection and to the complex matrix present in the determination. Many researchers have determined these elements in urine by inductively coupled plasma atomic emission spectroscopy (ICP-AES),13 neutron activation analysis (NAA),14 inductively coupled plasma mass spectrometry (ICPMS),15,16 flame atomic absorption spectrometry, (FAAS) and graphite furnace atomic absorption spectrometry (GFAAS).17,18 Each method has its particular advantages and disadvantages. The most important commercially available multielement instruments for trace element determination are ICP-based. The applicability of ICP-based instruments has been somewhat limited by difficulty in dealing with high salt concentrations and the need for relativity large sample volumes. Consequently, separation or preconcentration of analyte from the matrix prior to measurement is necessary. Although ICPMS has emerged as a method with a power of detection at least comparable to that of GFAAS, the MS detector is rather complex and expensive, which has limited widespread use of ICPMS in routine work in laboratories and hospitals. Of these methods, GFAAS is probably the most popular technique today for the determination of these elements in clinical samples because of its speed, simplicity, good sensitivity, and low cost. There are numerous papers on the determination of trace elements in urine by GFAAS. Dube developed a method for determination of Cu in urine and blood samples based on Zeeman effect atomic absorption spectrometry. A detection limit of 0.75 µg/L for diluted urine samples and a precision of 2-8% (up to 65 (8) Leach, R. M.; Liburn, M. S. World Rev. Nutr. Diet 1978, 32, 123. (9) Teraoka, H.; Morii, F. J. Kobayashi, Eiyo Shokuryo 1981, 34, 221. (10) WHO Recommended Health-Based Limits in Occupational Exposure to Heavy Metals, Report of a WHO Study Group, Tech. Rep. Ser. No. 647, WHO, Geneva, 1980. (11) Piscator, M. Manganese. In Handbook on the Toxicology of Metals; Friberg, L., Nordberg, G. F., Vouk, V. B., Eds.; Elsevier/North-Holland: Amsterdam, 1979; p 185. (12) Tsalev, D. L. Atomic Absorption Spectrometry. In Occupational and Environmental Health Practice; CRC Press: Boca Raton: FL, 1984; Vol. 2. (13) Allain, P.; Mauras, Y. Anal. Chem. 1976, 51, 2089. (14) Bowen, H. J. M. CRC Crit. Rev. Anal. Chem. 1980, 10, 127. (15) Jouhanneau, P.; Raisbeck, G. M.; Yiou, F.; Lacour, B. Clin. Chem. 1997, 43 (6), 1023. (16) Headley, J. V.; Massiah, W.; Laberge, D.; Purdy, J. R. J. AOAC Int. 1996, 79 (5), 1184, (17) Clavel, J. P.; Jandon, M. C.; Galli, A. Ann. Biol. Clin. 1978, 36, 33. (18) Routh, M. W. Anal. Chem. 1980, 52, 182.
4320
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
µg/L Cu) was observed.19 Dube also determined Cr in urine by the same instrument. The reproducibility and limit of detection were in the order of (10% and 0.09 µg/L Cr, respectively.20 Burguera reported the effect of the chemical modifiers Eu, Mg(NO3)2, Pd, Eu-Pd, and Ni and the use of longitudinally (with deuterium lamp background correction) and transversally (with Zeeman effect background correction) heated atomizers on the determination of Cr in urine samples. A precision between 0.8 and 2.5% and a detection limit in the ppb range were found.21 D’Haese described different method of GFAAS with automatic sampling for determining Al in human serum, blood, urine, and tissue. The standard additions method and two ashing steps (700, 1400 °C) were used in all urine samples. The detection limit was 1.3 µg/L.22 Halls and Fell determined Mn in urine by GFAAS after dilution (1:1) with distilled water.23 However, a significant drawback of GFAAS is its single-element character, which can be timeconsuming if more than one element has to be measured. With the introduction of the multielement GFAAS, the analytical potential of GFAAS has been increased. However, the applications of the multielement GFAAS in the direct and simultaneous determination of trace elements in environmental and biological samples are not as much as one may expect. It has been used to determine several trace elements in samples with a simple matrix, such as portable water.24,25 The hydride-forming elements can be determined using the technique of “in-atomizer trapping” GFAAS,26-28 but its application is generally applied to portable water only. This technique has recently been used to determine elements in brackish water; the method of standard addition is essential for the analysis.28 Multielement GFAAS has been used to analyze samples with a complicate matrix, such as salts29 or silicate rock;30,31 however, separation or preconcentration of analyte from the matrix prior to measurement is necessary. The simultaneous determinations of Co and Mn32or Pb and Cd33in urine have been reported, the calibration slopes for aqueous standards and matrix-matched standards were not parallel, and therefore, matrix-matched standards were used for all determinations. Bencs et al.34 developed a method for the determination of chromium, molybdenum, and vanadium dopants in bismuth tellurite optical crystals by multielement GFAAS; by applying a multiple dosing and matrix prevaporization procedure (by the addition of triammonium citrate chemical modifier); the calibration could be performed using simple aqueous standards. We reported that (19) Dube, P. At. Spectrom. 1988, 9 (2), 55. (20) Dube, P. Analyst 1988, 113, 917. (21) Burguera, J. L. J. Anal. At. Spectrom. 1999, 14, 821. (22) D’Haese, P. C. Clin. Chem. 1985, 31 (1), 24. (23) Halls, D. J.; Fell, G. S. Anal. Chim. Acta 1981, 129, 205. (24) Latino, J. C.; Sears, D. C.; Portala, F.; Shuttler, I. L. At. Spectrom. 1995, (May/June), 121. (25) Sears, D.; Latino, J.; Portala, F.; Shuttler, I. L. PE Newsl. 1998, 11, 48. (26) Murphy, J.; Jones, P.; Schlemmer, G.; Shuttler, I. L.; Hill, S. J. Anal. Commun. 1997, 34, 359. (27) Murphy, J.; Schlemmer, G.; Shuttler, I. L.; Jones, P.; Hill, S. J. J. Anal. At. Spectrom. 1999, 14, 1593. (28) Elsayed, M.; Bjorn, E.; Frech, W. J. Anal. At. Spectrom. 2000, 15, 697. (29) Arpadjan, S.; Vassileva, E.; Momchilova, S. Analyst 1992, 117, 1933. (30) Gupta, J. G. S. J. Anal. At. Spectrom. 1993, 8, 93. (31) Gupta, J. G. S. Talanta 1993, 40 (6), 791. (32) Iversen, B. S.; Panayi, A.; Camblor, J. P.; Sabbioni, E. J. Anal. At. Spectrom. 1996, 11, 591. (33) White, M. A.; Panayi, A. At. Spectrom. 1998, 19(3), 89. (34) Bencs, L.; Szaka´cs, O.; Ka´ntor, T.; Varga, I.; Bozsai, G. Spectrochim. Acta 2000, 55B, 883,.
using Pd and Mg as the chemical modifier, one can determine Mo and V35 in seawater simultaneously by using the method of standard additions and determine Cu and Mn36simultaneously using calibration curves with simple aqueous standards. We reported recently that when a mixture of Pd and Mg with the purge gas (5% H2 in Ar) was used as the chemical modifier, Cu, Mn, and Mo in seawater samples could be determined simultaneously using calibration curves with simple aqueous standards.37 These results indicate that direct and simultaneously determination of trace elements from complicate matrix is possible if one can find the suitable chemical modifier. Many kinds of chemical modifiers were used in the analysis of urine by single-element GFAAS, such as Mg(NO3)2, Pd, Tritox100, NH4NO3, and a mixture of Pd and Mg(NO3)2. Schlemmer and Welz reported38 that the use of Pd in combination with magnesium nitrate was a fairly universally applicable modifier in GFAAS, and therefore it is recommended for multielement analysis. But the contamination of Cr present in the reagent grade Mg(NO3)2 solution caused a larger absorbance (0.387 A‚s/10 µL), so this material may not be suitable for this work. In contrast, Pd is a high-purity reagent and widely used in seawater, sediment, urine, and biological samples analysis. Palladium metal acts as the chemical modifier, although it normally is introduced into the furnace as a chloride or nitrate salt. Palladium metal is obtained through either thermal decomposition in the course of the pyrolysis stage or reduction of the palladium with reducing agents such as ascorbic acid, hydroxylamine hydrochloride, or hydrogen.39-41 Hydrogen is recommended for the reduction of Pd42 since it is free from the contamination that often arises from the addition of aqueous reducing agents, and also it promotes the formation of palladium metal earlier in the temperature program. Welz et al.39 eliminated chloride interference in the GFAAS determination of thallium in seawater by use of a mixture of Pd (NO3) 2 and Mg (NO3)2 as the modifier and 5% H2 in Ar as the purge gas. Creed et al.40 minimized chloride interference produced by combination acid digesting by using Pd/Mg (NO3)2 and 5% H2 in Ar as a modifier in GFAAS. Pd (or a mixture of Pd and Mg) with the purge gas (5% H2 in Ar) has been shown to be a very effective chemical modifier;37,39,40,42 however, it did not raise too much attention to the user of GFAAS. This is probably because this purge gas is not a general accessory of GFAAS; furthermore, it is not too difficult to find a chemical modifier without using this purge gas for singleelement determination. To develop a technique for direct and simultaneous determination of multielements in samples with a complicate matrix, one certainly needs a very powerful chemical modifier. The aim of this study is to develop the technique for the determination of Cu, Cr, Al, and Mn in urine directly and simultaneously with a multielement GFAAS (SIMAA 6000). Pd (35) Su, P.-G.; Huang, S.-D. J. Anal. At. Spectrom. 1998, 13, 641. (36) Su, P.-G.; Huang, S.-D. Spectrochim. Acta 1998, 55B, 708. (37) Chen, C.-L.; Danadurai, S. K.; Huang, S.-D. J. Anal. At. Spectrom. 2001, 16, 404. (38) Schlemmer, G.; Welz, B. Spectrochim. Acta, Part B 1986, 41, 1157. (39) Welz, B.; Schlemmer, G.; Mudakavi, J. R. J. Anal. At. Spectrom. 1988, 3, 695. (40) Creed, J.; Martin, T.; Lobring L.; O’Dell, J. Environ. Sci. Technol. 1992, 26, 102. (41) Voth-Beach, L. M.; Shrader, D. E. J. Anal. At. Spectrom. 1987, 2, 45. (42) Shrader, D. E.; Voth-Beach, L. M.; Rettberg, T. M. Res. Natl. Inst. Stand. Technol. 1988, 93, 450.
was used as the chemical modifier along with a special purge gas (5% H2 in Ar). A standard reference material (Seronorm Trace Elements Urine) was used to find the optimal temperature program and to confirm the accuracy of the technique. EXPERIMENTAL SECTION Instrumentation. An atomic absorption spectrometer, SIMMA 6000, from Perkin-Elmer, Norwalk, CT, was used for simultaneous multielementt analysis. A longitudinal Zeeman effect background correction system, a transverse heated graphite atomizer (THGA) and pyrolytically coated graphite tubes with platforms were used. The samples were injected from an AS-72 autosampler with an 80-position tray. The rate of flow of the normal gas (pure argon) or purge gas (5% H2 in Ar) was 250 mL/min. This flow was stopped during atomization. The procedure was controlled by the AA Winlab software version 2.3 (Perkin-Elmer). The lamps of Cu, Cr, Al, and Mn used were hollow cathode lamps (HCL) from PerkinElmer and the wavelengths used were Cu 324.8 nm, Cr 357.9 nm, Al 309.3 nm, and Mn 279.5 nm. Unless otherwise specified, each experimental result reported is the arithmetic average of three determinations. The peak area of the atomic absorption signal was used for the determination. A field emission gun scanning electron microprobe (ISM-6330F, JEOL) was used. Reagents and Standard Materials. All dilutions were made using Milli-Q ultrapure deionized water (Milli-Pore, Hamburg, Germany). Commercial Cu, Cr, Al, and Mn standards (1.000 g/L, Merck) were used. Throughout the procedures, Suprapure double sub-boiling distilled HNO3 (Merck) was used. A solution of Pd(NO3)2 was prepared by dissolving 100 mg of Pd metal powder (The particle size is 60 µm, 230 mesh, from Merck.) in 1 mL of concentrated nitric acid and diluting to 100 mL with deionized water. If solution was incomplete, 10 µL of Suprapure grade concentrated hydrochloric acid was added to the cold nitric solution, which was then heated gently until the solution cleared. The solution was then heated to gentle boiling in order to volatilize the excess chloride. The Chelex-100 ion exchange resin (sodium form, 100-200 mesh) used for preparing blank urine was obtained from Bio-Rad Laboratories. Standard reference material (SRM) Seronorm Trace Elements Urine from Sero (Billingstad, Norway) was used to test the accuracy of the results. Collection and Handling of Real Urine Samples. Urine samples obtained from healthy laboratory workers were voided directly into acid-washed polypropylene containers, and 0.35 mL of concentrated HNO3 was added to acidify the urine (100 mL). The samples were stored at 4 °C when the analysis of urine samples was not carried out immediately,. Blank Urine Preparation. The blank urine was prepared by adjusting the pH of the real urine sample to 5.0 and passing it through a conditioned Chelex-100 column at a flow rate of 0.20.3 mL/min. First, 4 g of resin was placed in a PTFE beaker with nitric acid (2 N, 30 mL) overnight; then it was transferred to a column. The resin was washed with nitric acid (2 N, 20 mL) and deionized water (50 mL). Aqueous ammonia (2 M, 20 mL) was added to convert the resin from the hydrogen form to the ammonium form. Finally, the resin was washed with excess deionized water until the pH was 7. Contamination Control. A class 100 laminar flow clean bench was used to prepare the solutions. All glassware, pipets, micropipet tips, autosampler cups, and polypropylene containers were acid Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4321
Figure 1. Influence of ashing temperature on the absorbance of dilute (1:1) Seronorm urine (20 µL) using a Pd modifier (5 µg). 5% H2 in 95% Ar was used as the purge gas before atomization. The atomization temperature is 2300 °C for this curve: (9) peak area signal and (0) background absorbance for Cu; (b) peak area signal and (O) background absorbance for Cr; (1) peak area signal and (3) background absorbance for Mn; (2) peak area signal and (4) background absorbance for Al.
Figure 2. Influence of atomization temperature on the absorbance of dilute (1:1) Seronorm urine (20 µL) using the modifier of Pd (5 µg). 5% H2 in 95% Ar was used as purge gas before atomization. The ashing temperature is 1250 °C for those curves: (9) peak area signal and (0) background absorbance for Cu; (b) peak area signal and (O) background absorbance for Cr; (1) peak area signal and (3) background absorbance for Mn; (2) peak area signal and (4) background absorbance for Al.
washed with 20% v/v HNO3 for 24 h and thoroughly rinsed three times with distilled water before use. Procedure. The lyophilized Seronorm Trace Elements Urine was added to exactly 5 mL with pure water, let stand for 30 min, and then transferred to a plastic tube. The urine standard and samples were diluted 1:1 with Milli-Q ultrapure water. A 20-µL aliquot of the diluted standard or samples and 5 µL of the Pd modifier solution were injected sequentially into the GFAAS. RESULTS AND DISCUSSION Effect of Ashing and Atomization Temperatures. The effect of ashing temperature on the atomic absorption and background absorption signals is shown in Figure 1. The 5% H2 in Ar was used in the drying and ashing steps and 100% Ar was used in the later steps. The importance of the addition of H2 was a detailed point of optimization discussed later on. For Cu and Mn, the atomic signal remained approximately constant as the ashing temperature varied from 1000 to 1350 and 1100 to 1350 °C, respectively. When the ashing temperature was increased to 1550 °C, the specimen evaporated considerably and the atomic signal decreased correspondingly. Relatively, Cr and Al were the refractory elements. Their atomic signals remained approximately constant as the ashing temperature varied from 1100 to 1500 (for Cr) and 1200 to 1550 °C (for Al). The background signals decreased with increasing ashing temperature and were reduced below 0.1 A‚s for ashing temperatures above 1250 °C. To measure four elements directly and simultaneously, the ashing temperatures from 1250 to 1350 °C were chosen for further testing. The effect of atomization temperature (with ashing temperature at 1250 °C) is shown in Figure 2. For Cu, Cr, and Mn, the atomic and background absorption signals decreased gradually with increasing atomization temperature. However, Al atoms were not easy to release from the complex matrix below an atomization temperature of 2100 °C. So the peaks do not converge to a limit and were broad. With increasing atomization temperature, the sensitivity for Al increased and the maximum signal appeared at 4322
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
Figure 3. Influence of palladium on the absorbance of dilute (1:1) Seronorm urine (20 µL). 5% H2 in 95% Ar was used as purge gas before atomization. The ashing temperature is 1250 °C, and the atomization temperature is 2300 °C; (9) peak area signal and (0) background absorbance for Cu; (b) peak area signal and (O) background absorbance for Cr; (1) peak area signal and (3) background absorbance for Mn; (2) peak area signal and (4) background absorbance for Al. Table 1. Different Ashing and Atomization Temperatures for Simultaneous Determination of Cu, Cr, Al, and Mn in Dilute (1:1) Sernorm Urine step
temp (°C)
ramp time (s)
hold time (s)
gas flow (mL min-1)
gas type
drying 1 drying 2 ashing atomization cleanout
110 130 variousa variousb 2500
1 15 30 0 1
20 20 20 5 5
250 250 250 0 250
S S S N N
a Ashing temperature: 1250, 1350, and 1350 °C. b Atomization temperature: 2200 and 2300 °C.
an atomization temperature of 2200 °C. For multielement analysis, atomization temperatures from 2200 to 2300 °C were chosen for further testing.. Effect of Pd(NO3)2. The absorption signals obtained using different masses of Pd as the chemical modifier are shown in
Table 2. Simultaneously Determined Values of Cu, Cr, Al, and Mn in Dilute (1:1) Sernorm Urine with Different Ashing and Atomization Temperatures found values (µg/L) °Ca
1250 Cu Cr Al Mn a
°Cb
30.7 ( 0.2 25.3 ( 0.0 212. ( 3 13.6 ( 0.06
2200 1300 °Cb
1350
28.8 ( 0.1 24.3 ( 0.1 192. ( 1 13.0 ( 0.1
Atomization temperature. b Ashing temperature.
°Cb
25.1 ( 0.2 24.5 ( 0.1 183 ( 2 13.4 ( 0.1 c
1250
2300 °Ca 1300 °Cb
°Cb
23.9 ( 0.2 19.7 ( 0.1 116 ( 3 10.6 ( 0.1
1350
23.8 ( 0.2 21.2(.0.0 103 ( 1 11.2 ( 0.0
reported values (µg/L)c
°Cb
19.8 ( 0.2 19.5 ( 0.3 98 ( 2 10.2 ( 0.1
28 20 132 13
Seronorm Trace Elements Urine from Sero (Billingstd, Norway).
Figure 3. The ashing temperature was 1250 °C, and atomization took place at 2300 °C. The purge gases used were the same as in the previous study. For Cu, the addition of Pd as the chemical modifier increased the atomic signal. It had been shown that Cu and Pd combine as an alloy, so Cu-Pd are able to stand higher ashing temperatures that would remove more of the matrix.43,44 The absorption signals remained approximately constant on the addition of Pd (3-9 µg) for Cu. For Mn, the optimum dosage of Pd was 5-9 µg due to the lower background signal. For Cr and Al, the background signals were lowest with the addition of 5 µg of Pd. For multielement analysis, 5 µg of Pd was used; this yielded results with greatest precision and lowest background. Analysis of Certified Reference Materials with Different Heating Modes. The temperature programs used are shown in Table 1. We tried to find the optimal temperature program by using the method of calibration curves to analyze the diluted standard urine with various temperature programs. The results are shown in Table 2. On comparing the recommended values and the found values of each element, we obtain suitable ashing and atomization temperatures for the determination of a single element. For Cu, the accurate values were an ashing temperature of 1250-1300 °C and an atomization temperature of 2200 °C. For Mn, best results were obtained at an ashing temperature of 1300 °C and an atomization temperature of 2200 °C. For Cr, an ashing temperature between 1250 and 1350 °C and an atomization temperature of 2300 °C provided good results. However, Al could not be determined accurately under any of these conditions. To measure all four elements simultaneously, we made further tests by replaced the purge gas (5% H2 in Ar) with pure Ar during the last ashing step (shown in Table 3). This ensures that the gas in the atomizer during the atomization step does not contain H2, which may affect the efficiency of atomization. The results are shown in Table 4. The values found for Cu and Mn were lower than the reported values, and the values found for Cr and Al were higher than the reported values. But all these found values were within a range of 90-110% of the reported values. The slope and linearity correlation coefficient (R) of calibration curve are shown in Table 5. The requirement to remove H2 gas prior to atomization step was also reported by Creed et al.,40 who found that if the 5% H2 gas mixture was not adequately purged from the atomizer prior to atomization, the Se signal was reduced by 18%. The incorporation of a purged step in Ar restores the Se signal to that found when no H2 was present. (43) Qiao, H.; Jackson, K. W. Spectrochim. Acta 1991, 46B, 1841. (44) Qiao, H.; Jackson, K. W. Spectrochim. Acta 1991, 47B, 1267.
Table 3. Optimum Temperature Program for Simultaneous Determination of Cu, Cr, Al, and Mn step drying ashing atomization cleanout a
temp (°C)
ramp time (s)
hold time (s)
gas flow (mL min-1)
gas typea
110 130 1250 1250 2300 2500
1 15 30 1 0 1
20 20 15 5 5 5
250 250 250 250 0 250
S S S N N N
read
read
S, 5% H2 in 95% Ar; N, 100% Ar.
Table 4. Simultaneously Determined Values of Cu, Cr, Al, and Mn in Dilute (1:1) Sernorm Urine with Optimal Temperature Program
Cu Cr Al Mn
found values (µg/L)
reported values (µg/L)
25.8 ( 0.4 21.6 ( 0.3 142 ( 2 11.8 ( 0.2
28 20 132 13
Table 5. Slope and Linearity Correlation Coefficient (R) of Calibration Curvea for Simultaneous Determination of Cu, Cr, Al, and Mn in Dilute (1:1) Sernorm Urine with Optimal Temperature Program
Cu Cr Al Mn
slope
R
0.003 45 0.009 45 0.001 44 0.007 87
0.999 0.999 0.997 0.999
a Calibration curve with the following concentrations: Cu, 6, 10, 14, 18 ppb; Cr, 3, 7, 10, 13 ppb; Al, 20, 44, 65, 85 ppb; Mn, 2, 4, 6.5, 9 ppb.
We also tested the analysis using pure Ar as the purge gas. The results are shown in Table 6. On comparing these with the results shown in Table 4, one sees that changing the purge gas does not affect the determination of Cr and Mn very much. For Cu, the value found was smaller than the reported value. For Al, the peaks became broad and the value found was larger than the reported value. The study demonstrated that Cu and Al in urine could not be determined accurately without H2. Scanning Electron Microscopic Examination of Palladium Distribution on Platforms. Voth-Beach and Shrader41 observed that the matrix affects the function of the Pd chemical modifier. Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4323
Figure 4. SEM graphs obtained by pipetting dilute (1:1) Seronorm urine (10 µL) and Pd (40 µg) onto the platform with different purge gas: (a) 5000× magnification; (b) 10000× magnification with 5% H2 in 95% Ar; (c) 5000× magnifcation; and (d) 10000× magnification with 100% Ar as the purge gas.
Table 6. Values, Slope, and Linearity Correlation Coefficient (R) of Calibrationa Curve for Simultaneous Determination of Cu, Cr, Al, and Mn in Dilute (1:1) Sernorm Urine with 100% Ar Purge Gas Program
Cu Cr Al Mn
found values (µg/L)
slope
R
22.4 ( 0.5 21.2 ( 0.4 217 ( 6 11.9 ( 0.3
0.003 77 0.009 94 0.000 99 0.009 82
0.995 0.999 0.989 0.999
a Calibration curve with the following concentrations: Cu, 6, 10, 14, 18 ppb; Cr, 3, 7, 10, 13 ppb; Al, 20, 44, 65, 85 ppb; Mn, 2, 4, 6.5, 9 ppb.
To develop the Pd capability of stabilizing metals, it was necessary to keep Pd in the reduced state. Also, the distribution and the particle size of the Pd in the furnace tube (studied using a scanning electron microscope) also affects the function of the modifier and the expression of the absorption peaks.37,43,44 In this study, we used a scanning electron microscope (SEM) to observe the platform and find the relationship between the purge gas used and the distribution of Pd particles on the platform (in the matrix of urine). Palladium (40 µg) and the diluted standard urine (10 µL) were pipetted into a platform in the graphite furnace, dried, and then heated to an ashing temperature of 1250 °C. After cooling, the platform was carefully transferred to the SEM. The results are shown in Figure 4a and b for experiments in which the furnace was heated to 1250 °C with 5% H2 in Ar as the purge gas; in Figure 4c and d, 100% Ar was used as the purge gas. The white points are Pd particles. The Pd particles are larger in the absence of H2, which makes diffusion difficult for the analyte as 4324 Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
evidenced by the fact that broadened peaks were observed. The particle diameters of are 0.4-1.2 µm. When H2 was used, the Pd particles were found to be smaller and distributed uniformly on the surface. The particle diameters of are 0.4-0.6 µm. This is due to the faster reduction of Pd by hydrogen and the maintenance of the Pd in its reduced form. In addition to this effect, H2 gas may also eliminate the chloride interference from NaCl in the urine samples. Real Urine Sample Analysis. The results of using the optimal temperature program to analyze real human urine samples are shown in Table 7. Recovery was tested by spiking urine with the trace elements (Cu, 4 µg L-1; Cr, 1µg L-1; Al, 5 µg L-1; Mn, 0.5 µg L-1). The results obtained for diluted urine were satisfactory. Dilution (1:1) decreased the background. The recoveries of the four elements were 100 ( 5%, except for Cu (80%). We have also compared the slopes of calibration curves with the slopes obtained with the method of standard additions. The slope ratios of each element were 0.89, 0.95, 0.88, and 1.04, respectively. The analyzed values were similar to each other. This means that matrix interference can be eliminated completely and that the developed methods have excellent reliability. Analytical Performance. Detection limits (DLs) and relative standard deviations (RSDs) are shown in Table 8. Detection limits were calculated as 3 times the standard deviation of seven replicate measurements of the urine blank. The detection limits were 0.08 µg L-1 for Cu, 0.05 µg L-1 for Cr, 0.06 µg L-1 for Al, and 0.06 µg L-1 for Mn. The RSDs for the direct and simultaneous determination of Cu, Cr, Al, and Mn in reference urine were below 2% with Pd as the chemical modifier.
Table 7. Results of Simultaneous Determination of Cu, Cr, Al, and Mn in Real Dilute (1:1) and Real Undiluted Urine Samples found values (µg/L)
Cu Cr Al Mn
background absorbance (A‚s)
dilute (1:1)
undiluted
dilute (1:1)
undiluted
dilute (1:1)
14.3 ( 0.1 0.78 ( 0.04 18.9 ( 0.2 0.06 ( 0.01
0.1463 0.0786 0.1250 0.1618
0.0510 0.0217 0.0312 0.0396
81.7 ( 0.3 102.6 ( 0.4 144.8 ( 0.2 75.3 ( 0.0
79.8 ( 0.2 95.0 ( 0.7 101.6 ( 0.7 98.2 ( 2.1
Table 8. Detection Limita (DL) and Relative Standard Deviations (RSDs) Obtained for Simultaneous Determination of Cu, Cr, Al, and Mn in Dilute (1:1) Real Urine Samples
Cu Cr Al Mn
recovery (%)
DL (µg/L)
RSD (%)
0.08 0.05 0.06 0.06
1.7 1.5 1.6 1.5
a Detection limit calculated as 3 times the standard deviation of seven replicate measurement of dilute (1:1) real urine blank.
Pd on the platform. The procedure is fast and accurate. The detection limits and precision are satisfactory. The use of 5% H2 in Ar eliminated chloride interference and showed a more uniform distribution of smaller particles of Pd on the platform. We have recently shown that a mixture of Pd (NO3)2 and Mg (NO3)2 with the purge gas (5% H2 in Ar) is an effective chemical modifier for the direct and simultaneous determination of three trace elements in seawater samples.37 We expect that Pd (or mixture of Pd and Mg) with the purge gas (5% H2 in Ar) may found to be the most useful chemical modifiers for the direct and simultaneously determination of multielements using GFAAS. ACKNOWLEDGMENT
CONCLUSION Cu, Cr, Al, and Mn in urine can be determined directly and simultaneously with high accuracy and precision using multielement GFAAS. A simple calibration curve can be used when Pd is used as the chemical modifier along with a special purge gas (5% H2 in Ar). The use of 5% H2 in Ar eliminated chloride interference and showed a more uniform distribution of smaller particles of
We thank the National Science Council of the Republic of China for support (grant NSC 88-2113-M007-001).
Received for review March 15, 2001. Accepted June 22, 2001. AC010319H
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4325
