ANALYTICAL SCIENCES DECEMBER 2006, VOL.

22 2006 The Japan Society for Analytical Chemistry

1605

Notes

Determination of Manganese in Urine and Whole Blood Samples by Electrothermal Atomic Absorption Spectrometry: Comparison of Chemical Modifiers
Frederico Garcia PINTO,* Ulisses Villela REY,* Eduardo Freitas FERNANDES,** Josianne Niccio SILVEIRA,** Leiliane AMORIM,** and Jos Bento Borba da SILVA*
*Departamento de Qumica, Universidade Federal de Minas Gerais, Avenida Antnio Carlos 6627, 31270-901 Belo Horizonte, MG, Brazil **Laboratrio de Toxicologia Ocupacional, Faculdade de Farmcia, Universidade Federal de Minas Gerais, Avenida Antnio Carlos 6627, 31270-901 Belo Horizonte, MG, Brazil

In this work, methodologies to determine manganese (Mn) in urine and whole blood by electrothermal atomic absorption spectrometry were developed. The use of Ru, Rh, and Zr as permanent modifiers, Pd as a modifier in solution, and the condition without modifier were investigated for the direct determination of Mn in urine and whole blood samples. The best results for Mn in urine and in whole blood were obtained without modifier use. The analytical characteristic, such as accuracy, precision and limit of detection of the proposed methodology were adequate. (Received January 6, 2006; Accepted July 20, 2006; Published December 10, 2006)

Because of their biochemical importance, the concentrations of trace and ultra-trace elements in body fluids such as serum, blood, and urine can often be used as biological indicators of human health, disease, and nutrition condition, and their determination has received great attention from biological and medical researchers.14 Manganese is largely distributed as particles on the earth crust, in water and in the atmosphere,5 being an essential element for human beings and animals. Normal human blood and urine levels of manganese are 2 8 g dL1 and 0.1 0.8 g dL1, respectively.6 Individuals exposed to high levels of manganese can suffer deterioration of neurological function, which begins with unspecific symptoms and neurological signals. These symptoms include anorexia, apathy, asthenia, headache, irritability, lethargy and weakness of the extremities with the final development of serious clinical conditions such as Parkinsons disease.7 One of the most usual analytical techniques to determine trace amounts of analytes in such samples is electrothermal atomic absorption spectrometry (ETAAS) due to its high sensitivity, selectivity, and simplicity.8 The use of chemical modifiers has become a routine in ETAAS to eliminate matrix interferences. After the introduction of the stabilized temperature platform furnace concept (STPF),9 new reagents have been researched as chemical modifiers. Chemical modification of the sample is recommended in the STPF concept to increase pyrolysis temperature, minimize matrix interferences in the vapor phase during atomization, stabilize the analyte on the platform while the vapor phase approaches thermal balance, and separate the analyte signal from the background. A large number of chemical modifiers have been proposed in the literature in the last years. A mixture of Pd and Mg nitrates was recommended as a universal chemical modifier for about 21 analytes.10 The use of graphite tubes treated with carbide To whom correspondence should be addressed. E-mail: bentojb@terra.com.br

forming elements (Zr, W, La and Nb)1113 and more recently with platinum group metals (Pd, Pt, Ir, Rh, and Ru)1419 as chemical modifiers in the direct determination of metals by graphite furnace has been studied. Treated surfaces have the advantage of acting as permanent chemical modifiers, allowing a great number of atomization cycles without the need of new modifier applications, besides making in situ cleaning of the modifier during treatment possible. Many authors have presented methods to determine Mn in whole blood and urine. Soko et al.20 proposed using a supported liquid membrane (SLM) technique for extraction and preconcentration of manganese in blood. Adsorptive stripping voltammetry (AdSV) has been applied to measure Mn concentration. Ohta et al.21 described a method for the direct determination of manganese in biological samples by using ETAAS with a molybdenum tube atomizer and a matrix modifier. The effects of large amounts of interferents were investigated and eliminated by the addition of thiourea. Lunas and Campos22 reported the effects of Mg(NO3)2, Pd(NO3)2 and the mixed nitrates of Pd + Mg modifiers on the determination of Mn in urine samples. The matrix effect was only removed when Pd(NO3)2 was used as a modifier. Recoveries were close to 100% with LOD of 0.6 g L1. In the present work, the uses of Ru, Rh, and Zr as permanent modifiers, Pd as a modifier in solution, and the condition without modifier were investigated for the direct determination of Mn in urine and whole blood samples by ETAAS.

Experimental
Apparatus All measurements were carried out with a Varian (Victoria, Australia) SpectrAA 220 atomic absorption spectrometer equipped with a graphite furnace and an autosampler EL

1606
98013384-2C (Varian) and the Polarized Zeeman Background correction procedure. The atomic signal was measured in peak height mode. A single element hollow cathode lamp for Mn from Varian was used as light source; it was operated at 5 mA with a spectral band-pass of 0.2 nm at 279.5 nm. The volume pipetted into the graphite furnace was 10 L of samples and calibration solutions and 5 L for modifier whenever used in solution. Argon, 99.999% (v/v) (White Martins, Contagem, Mg, Brazil), was used as carrier gas and gas flow was interrupted only during atomization. Pyrolytic-graphite-coated tubes without and with an integrated platform (Varian) were used throughout this work. The volume of sample and calibration solutions was 10 L. The temperatures used were 85 120C/40 s (ramp mode, urine) and 85 120C/55 s (ramp mode, whole blood) in the drying; 1100C/5 s (ramp mode, urine), 1000C/5 s (ramp mode, whole blood) and 1100C/20 s (step mode, urine), 1000C/20 s (step mode, whole blood) in ashing; 40C/18.2 s (ramp mode, urine) and 40C/18.0 s (ramp mode, whole blood) in the cool down; 2100C/1.2 s (ramp mode) and 2100C/2.0 s (step mode) in atomization and 2200C/2.0 s (ramp mode) for cleaning. The graphite tubes were treated with a permanent modifier by applying 50 L of a 1000-mg L1 solution of either Ru, or Rh, or Zr, independently for each modifier, and submitting the tube to a specific temperature program previously described.14 This procedure was repeated 10 times to obtain a 500-g deposit of permanent modifier. It is worthwhile to point out that the solvent drying temperature was also optimized in several stages to avoid solvent and analyte loss. Reagents and solutions All chemicals employed had analytical reagent grade unless otherwise specified. All solutions were prepared with high purity deionized water (resistivity 18.2 M cm) obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). The nitric acid 65% (w/w) and Triton X-100 used to dissolve and dilute the samples were from Merck (Darmstadt, Germany, No. 7587956). Before use, all glasses were washed abundantly with a detergent solution, rinsed with water, maintained in bath with nitric acid 2.8 mol L1 for a period not shorter than to 1 h, and later on rinsed several times with deionized water. The autosampler cups were submitted to the same treatment. The autosampler washing solution and diluent containing Triton X100 1.65 mmol L1 plus nitric acid 0.14 mol L1 were prepared to avoid analyte adsorption onto the surface of the container and to avoid clogging the autosampler pipette.23 The following 1000-g L1 stock solutions were used: ruthenium (Fluka, Buchs, Switzerland, No. 84033), rhodium (Fluka, No. 83722), zirconium (Aldrich, Milwaukee, Wis., USA, No. 27497-6), all in 1 mol L1 hydrochloric acid, and a palladium nitrate solution (Merck, No. B936989 710). A high purity standard reference solution of manganese 1 g L1 in 0.07 mol L1 HNO3 from Tritisol (Merck, Darmstadt, Germany) was used to obtain the reference analytical solutions. The limit of detection (LOD, g L1) was calculated using the equation LOD = 3 SBL/b, where SBL was the standard deviation of 10 blank measurements (nitric acid 0.14 mol L1) and b is the slope of the calibration curve. Samples and certified reference materials Urine and whole blood samples from healthy volunteers from Laboratrio de Toxicologia Ocupacional da Faculdade de Cincias Farmacuticas of Universidade Federal de Minas

ANALYTICAL SCIENCES DECEMBER 2006, VOL. 22

Fig. 1 Pyrolysis and atomization temperature curves obtained with different modifiers with 60 pg Mn in urine sample diluted in nitric acid 0.14 mol L1 and in 1.65 mmol L1 Triton X-100 at a ratio of 1:1 (filled symbol, pyrolysis; opened symbol, atomization).

Gerais (LATO-UFMG) were analyzed. Urine samples were prepared directly in the autosampler cups, diluted 1:1 with nitric acid 0.14 mol L1 and with 1.65 mmol L1 Triton X-100. For the dilution, a micro-pipette was used to homogenize the samples manually directly in the autosamplers cups. Whole blood samples were prepared in Eppendorf tubes, diluted 1:9 with nitric acid 0.14 mol L1 and with 1.65 mmol L1 Triton X100 and homogenized manually. Urine samples from Bio Rad (US Bio Rad Laboratories, Anaheim, USES), level 1 (Code 4001) and level 2 (Code 4051), were used to check the reliability of the entire proposed analytical method. Whole blood certified material was not available; thus the accuracy of the proposed method was evaluated through recovery study of spiked samples. The sample preparation for certified urine analysis was made using 1:10 dilution.

Results and Discussion

Pyrolysis and atomization temperature optimization All furnace program temperatures were optimized through pyrolysis and atomization temperature curves. The thermal stability of analytes in sample solutions with and without modifiers was studied. The selection for the modifier took into account the sensitivity, the absorption pulse form, and the low atomization temperature, which contributes to increase the graphite tube lifetime. The pyrolysis and atomization temperature curves were obtained for Mn for each matrix, by initially using the recommended atomization temperature and the best pyrolysis temperature selected in agreement with the criteria described, especially the sensitivity. The pyrolysis temperatures were determined for each matrix; 1100 and 1000C for urine and whole blood samples, respectively. The minimum atomization temperatures were determined by varying the temperature between 1700 and 2500C for urine and whole blood, and observing the sensitivity and the best pulse shape. Initially, analysis was carried out with and without integrated platform in graphite furnace for both matrices and absorbance reading in peak area or peak height modes without the use of the modifier. The best results were obtained for graphite tubes without integrated platform due to the high thermal stability of Mn and in peak height mode. The peaks were very symmetrical

ANALYTICAL SCIENCES DECEMBER 2006, VOL. 22

1607

Fig. 2 Pyrolysis and atomization temperature curves obtained with different modifiers for the background with 60 pg Mn in urine sample diluted in nitric acid 0.14 mol L1 and in 1.65 mmol L1 Triton X-100 at a ratio of 1:10 (filled symbol, pyrolysis; opened symbol, atomization).

Fig. 3 Pyrolysis and atomization temperature curves obtained with different modifiers with 46.8 ng Mn in whole blood sample diluted in nitric acid 0.14 mol L1 and in 1.65 mmol L1 Triton X-100 at a ratio of 1:10 (filled symbol, pyrolysis; opened symbol, atomization).

and returned to the baseline in 1 s. The effects of various chemical substances were investigated as chemical modifiers: palladium solution and ruthenium, rhodium, and zirconium as permanent modifiers (500 g of each independently) for urine and whole blood. The experiment was also carried out without modifier. Figure 1 shows the pyrolysis and atomization temperature curves of Mn in urine. The best results were obtained without modifier, with a symmetrical peak returning to the baseline in 1 s. The pyrolysis and atomization temperatures were 1100 and 2100C, respectively, and the characteristic mass, mo (1% absorbance) was 0.47 pg (recommended by the manufacturer without modifier in nitric acid 0.03 mol L1 of 1.3 pg). This is an important parameter for ETAAS sensitivity. With the permanent modifiers rhodium, zirconium, and ruthenium, it was observed that the Mn absorption pulses were wider and presented a tail with lower sensitivity. The tail was attributed to a possible memory effect caused by the interaction among Mn, the carbon on the graphite tube, and the modifiers, which are quite stable. With Zr as permanent modifier and Pd solution modifier, the sensitivity was about half as much as that without modifier. For ruthenium and rhodium, the thermal behavior of Mn was similar. The absorption pulses were larger and lower, and even when reading absorbance in peak area mode, the sensitivity decreased and the peaks were less reproducible, which affected measurement precision. To demonstrate how the modifiers affect the atomization background, we studied the atomization background profiles of the analytes. The analysis of the absorbances of the background, Fig. 2, revealed that with the use of permanent modifiers at low pyrolysis temperatures, the background values were much smaller compared to values without the use of modifier, ruthenium excepted. Additionally, background values decreased with the increase in pyrolysis temperature, which evidenced the destruction and the partial or the total concomitant elimination of the matrix from the sample. At the pyrolysis temperatures selected without modifier use, the background presented an absorbance value of 0.6476, which was perfectly corrected by the instrument (keeping in mind that the instrument correction system corrects background absorbance up to 2.0 absorbance units). For Mn in whole blood (Fig. 3), the best results were obtained without modifier, with

Fig. 4 Pyrolysis and atomization temperature curves for the background obtained with different modifiers with 46.8 pg Mn in whole blood sample diluted in nitric acid 0.14 mol L1 and in 1.65 mmol L1 Triton X-100 at a ratio of 1:10 (filled symbol, pyrolysis; opened symbol, atomization).

symmetrical peaks, that return to the baseline in 1 s. The pyrolysis and atomization temperatures were 1000 and 2100C, respectively, and the characteristic mass was 1.81 pg (recommended by the manufacturer without modifier use in aqueous solution of 1.3 pg nitric acid 0.03 mol L1). The absorption pulse obtained without the use of modifier was extremely sharp, which justified its larger sensibility when the absorbance reading was used in peak height. For readings with ruthenium and rhodium as permanent modifiers, the pulses were smaller, broad, and had a tail. Even with absorbance reading in the peak area, the sensitivity was lower and the peaks were less reproducible, which compromised measurement precision. With the use of either ruthenium or rhodium as permanent modifier, sensibility was threefold lower, while with permanent zirconium the sensibility was a little lower than without modifier use. The behavior of this analyte probably indicates an over-stabilization of Mn with these permanent modifiers and a memory effect due to the partial release of the analyte during the atomization step. On analyzing the background absorbances shown in Fig. 4, we observed that with the use of permanent modifiers, ruthenium excepted, the background values were much lower in comparison to values without modifier use, and

1608
Table 1 Matrix Urine Blood

ANALYTICAL SCIENCES DECEMBER 2006, VOL. 22

Characteristic masses (pg) of Mn obtained at optimal pyrolysis and atomization temperatures No modifier 0.47 1.81 Zirconium 1.02 2.15 Rhodium 8.59 6.00 Ruthenium 3.94 4.63 Palladium 0.61 Recommended 1.30 1.30

Table 2 Recovery of Mn from spiked urine and whole blood samples without modifier use Expected concentration/g L1 Urine 3 6 8 12 6 9 12 Recovery, % 99 109 107 100 107 102 99 RSD (n = 3) 8.2 9.6 7.0 6.0 5.4 2.4 2.7

Whole blood

similarly to the pattern observed for urine, absorbance decreased with the increase in pyrolysis temperature. However, for the selected pyrolysis temperature without modifier use, the background had an absorbance value of 0.0407 and the background was perfectly corrected. The best results obtained for manganese in urine and in whole blood were without any chemical modifier. This can be explained by the fact that the Mn that comes inside the graphite tube has a stable thermal behavior possibly due to the formation of refractory compositions of this element with the carbon on the graphite tube (Mn3C),24 the so-called refractory carbides. To overcome this problem, it was necessary to use a high atomization temperature for the complete atomization of the analyte, which was demonstrated by the pyrolysis and atomization curves obtained for urine and whole blood with the use of permanent modifiers. We observed that the increase in the atomization temperature of the analyte provoked an increase in the absorbance signal of Mn. However, there is an inconvenience; the increase in atomization temperature affects the lifetime of the graphite tube, which is costly and should be preserved for as long as possible. Although the use of modifier in ETAAS for manganese is controversial, Pd and Mg(NO3)2 have been successfully adopted to determine manganese in whole blood25 and urine.26 Table 1 shows the characteristic masses of the two matrices investigated in this work with each modifier. In relation to the background absorbance, it was observed that it was perfectly corrected for the two samples studied as the pyrolysis temperature was increased (approximately above 1000C). Analytical performance Calibration curves using matrix-matched calibration were constructed for the two matrices studied. Each point of the calibration curve of urine was obtained with solutions containing 500 L urine, Mn analytic solutions, and the diluent in a final volume of 1000 L. Each point for whole blood was obtained with solutions containing 100 L blood, Mn analytic solutions, and the diluent in a final volume of 1000 L. For both samples, the calibration range was from 2.5 to 12.5 g L1. The maximum value according to the legislation6 was of 8 g L1 for urine and between 7.5 10.5 g L1 for whole blood. The r values were higher than 0.998 for both samples. The characteristic masses were of 0.47 and 1.81 pg and the

limits of detections were of 0.2 and 0.3 g L1 for urine and whole blood, respectively. The relative standard deviation (n = 3) were lower than 3% for both samples. Recovery experiments were carried out for each matrix at the optimized pyrolysis and atomization temperatures and without the use of modifier. The recoveries were determined from triplicates and are shown in Table 2. To determine the recovery of Mn in urine and whole blood, the analyte concentration in each matrix was determined and subtracted from values of the same samples spiked with different manganese concentrations. It was verified that the recoveries of Mn at all levels for the two matrices investigated in this work were always close to 100%, indicating an acceptable accuracy. The standard deviations were usual for graphite furnace analysis. The results obtained for Mn in urine certified materials from Bio-Rad were 57.5 and 72.2 g L1, while the certified values were 59.3 and 70.0, respectively. These results demonstrated that for the two levels we obtained a good agreement with certified values. Manganese can be determined with sensitivity, accuracy, and precision in urine and whole blood samples without modifier use. Recoveries of spiked urine and whole blood samples and the Mn values determined for certified urine gave values close to 100%, indicating the acceptable accuracy of the methodologies proposed. The lifetime of the graphite tubes was about 400 atomization cycles.

Acknowledgements
The authors are grateful to Conselho Nacional de Pequisa e Desenvolvimento Tecnolgico (CNPq) for financial support. J. B. B. Silva, F. G. Pinto, and U. V. Rey thank CNPq for the scholarships.

References
1. R. M. Roalt-Malone, Bioinorganic Chemistry, 2002, Wiley, New York. 2. H. Vanhoe, R. Dams, and J. Versieck, J. Anal. At. Spectrom., 1994, 9, 23. 3. E. Fujimori, T. Hayashi, K. Inagaki, and H. Haraguchi, Fresenius J. Anal. Chem., 1999, 363, 277. 4. C. Prohaska, K. Pomazal, and I. Steffan, Fresenius J. Anal. Chem., 2000, 67, 479. 5. H. A. Waldron, Metals in the Environment, 1980, Academic Press, 199. 6. WHO, Manganese and Its Compounds, 1999, Concise International Chemical Assessment Document, Geneva, 12. 7. D. G. Barceloux, Manganese Clin. Toxicol., 1999, 37, 293. 8. B. Welz and M. Sperling, Atomic Absorption Spectrometry, 3rd ed., 1999, Wiley-VHC, Germany. 9. W. Slavin, D. C. Manning, and G. R. Carnrick, At. Spectrosc., 1981, 2, 137. 10. B. Welz, G. Schlemmer, and J. R. Mudakavi, J. Anal. At. Spectrom., 1992, 7, 1257.

ANALYTICAL SCIENCES DECEMBER 2006, VOL. 22

11. E. C. Lima, F. J. Krug, and K. W. Jackson, Spectrochim. Acta, Part B, 1998, 53, 1791. 12. D. L. Tsalev, V. I. Svaleikova, and P. B. Mandjukov, Spectrochim. Acta Rev., 1990, 13, 225. 13. E. C. Lima, F. J. Krug, A. T. Ferreira, and F. Barbosa Jr., J. Anal. At. Spectrom., 1999, 14, 269. 14. J. B. B. Silva, M. B. O. Giacomelli, I. G. Souza, and A. J. Curtius, Microchem. J., 1998, 60, 249. 15. J. B. B. Silva, M. A. M. Silva, A. J. Curtius, and B. Welz, J. Anal. At. Spectrom., 1999, 14, 1737. 16. C. Rademeyer, B. Radziuk, N. Romanova, N. P. Skaugset, A. Skogstad, and Y. Thomassen, J. Anal. At. Spectrom., 1995, 10, 739. 17. D. L. Tsalev, A. DUlivo, A. Lampugnani, M. Di Marco, and R. J. Zamboni, J. Anal. At. Spectrom., 1995, 10, 1003. 18. E. Bulska, W. Kandler, and A. Hulanicki, Spectrochim. Acta, Part B, 1996, 51, 1263.

1609
19. A. B. Volynski and S. V. Tikhomorov, J. Anal. Chem., 1998, 53, 718. 20. L. Soko, L. Chimuka, E. Cukrowska, and S. Pole, Anal. Chim. Acta, 2003, 485, 25. 21. K. Ohta, S. Itoh, S. Kaneco, and T. Mizuno, Anal. Sci., 1992, 8, 423. 22. A. S. Luna and R. C. Campos, At. Spectrosc., 1999, 20, 108. 23. B. R. Nunes, C. G. Magalhes, and J. B. B. Silva, J. Anal. At. Spectrom., 2002, 17, 1335. 24. H. I. Tekgul and S. Akman, Spectrochim. Acta, Part B, 1997, 52, 621. 25. F. R. Moreira and F. Pivetta, At. Spectrosc., 1998, 19, 137. 26. B. S. Iversen, A. Panayi, J. P. Camblor, and E. Sabbioni, J. Anal. At. Spectrom., 1996, 11, 591. 27. S. Oga, Fundamentos de Toxicologia, 1996, Atheneu Editora, So Paulo.