Abstract A new method for the determination of mercury and selenium in tobacco by inductively coupled plasma mass spectrometry (ICP-MS) using isotope dilution calibration was studied. The samples were digested with aqua regia in a microwave oven. The isotope ratios used for quantification were ²⁰¹Hg/²⁰²Hg and ⁷⁷Se/⁸²Se. A NaBH₄ solution was used as reducing agent. Five certified tobacco samples were analyzed and the concentrations obtained were in good agreement with the certified values. The detection limits in the samples were 0.5 ng g^-1 for mercury and 2 ng g^-1 for selenium, respectively. The method is precise, accurate and rapid.
Keywords  mercury; selenium; ICP-MS; isotopic dilution calibration; microwave digestion; Tobacco

1. INTRODUCTION
Mercury and selenium are important metal ions existing in tobacco. The Quality Standards says that the concentration of mercury should not be exceeding 0.2 m g g^-1 in tobacco because of its high toxicity. However, the selenium is helpful to our healthy when its concentration in tobacco within an appropriate range. ^[1] Therefore, the determination of trace mercury in tobacco and tobacco additive is important. Presently, there are several analytical methods for the determination of mercury and selenium including spectrophotometry, atomic absorption spectrometry, fluorescence and mass spectrometry (MS). ^[2-18] Among the analytical techniques employed, inductively coupled plasma mass spectrometry (ICP-MS) is especially important due to its several advantages, such as its multi-element capacity, high sensitivity and ability to measure isotopic ratios, which allows isotope dilution calibration (ID).^[11-18] In this work, a method for the determination of mercury and selenium in tobacco, after microwave-assisted acid digestion with aqua regia of the samples followed by chemical vapor generation with retention of the analyte vapor in an Ir-treated graphite tube of an electrothermal vaporizer using isotope dilution calibration (ID-CVG-ETVICP-MS) is optimized and described. The results show that this method is precise, accurate and rapid.

2 EXPERIMENTAL
2.1 Instrumental
All the measurements were carried out on an inductively coupled plasma mass spectrometer ELAN 6000 (Perkin Elmer SCIEX, Thornhill, ON, Canada). For the ETV system, a hydride generator MHS-15 (Perkin Elmer, CO, USA) was coupled to a Perkin-Elmer HGA 600 MS electrothermal vaporizer and a Perkin Elmer AS-60 autosampler and manually operated, as described previously.¹⁸ A 3% (m/v) sodium borohydride solution stabilized with 1% (m/v) sodium hydroxide was used as reducing agent in the MHS-15. The reducing agent was injected during 5 s using an argon pressure of 250 kPa. The generated vapors were transported to the ETV by an Ar flow through a 10 cm glass tube (0.1 cm i.d.) connected to a polytetrafluoroethylene (PTFE) tube (60 cm long, 0.5 cm i.d.). Pyrolytic coated graphite tubes (Perkin-Elmer, Part No. B050 8371) were used. Argon of 99.996% purity (Kunming Yangqi, Kunming, P.R.China) was used. An internal Ar flow rate of 0.1 L min^-1 in the vaporizer, optimized previously,^[18] summed to the "nebulizer" gas flow rate of 1.06 L min^-1 during vaporization, resulting in a total flow rate of 1.16 L min^-1, transported the aerosol to the plasma. The instrumental conditions are shown in Table 1 and the ETV temperature program, which also was optimized previously^[18] is shown in Table 2. The samples for the microwave digestion were weighed using a M2P microbalance (Sartorius, Göttingen, Germany). A Model WL 5001 microwave system (1000 W, Fei Yue Analytical Instrument Factory, Shanghai, China) was employed for the digestions of the samples.
2.2 Reagents and materials  
All the reagents were of analytical grade. The water used was de-ionized in a Milli-Q system (Millipore, Bedford, MA, USA). Nitric acid (High purity grade, Tianjing, P.R.China) and hydrochloric acid (High purity grade, Tianjing, P.R.China) were used. The reducing agent was prepared by dissolving NaBH₄ (Fluka, Buchs, Switzerland) in sodium hydroxide (High purity grade, Tianjing, P.R.China) stored in a polyethylene flask and kept under refrigeration for no longer than two days. A stock solution of IrCl₃ (Fluka, Buchs, Switzerland), 1000 mg L^-1 Ir, was used for the treatment of the tube with the permanent modifier as described previously.²⁰ The enriched isotope materials were from the Cambridge Isotope Laboratories Inc. (Andover, MA, USA). The abundances of the enriched isotopes were: 96.35% of ²⁰¹Hg and 93.48% of ⁷⁷Se. The isotope compositions of the enriched materials were measured and the found values agreed with the informed values. Stock solutions of 30 mg L^-1 for mercury and 200 m g L^-1 for selenium were prepared by dissolution of an accurately weighed amount of the solid material (HgO and elemental Se) in nitric acid and diluted in 5% v/v HNO₃. The following isotope ratios were used in the calculation of the concentrations: ²⁰¹Hg/²⁰²Hg and ⁷⁷Se/⁸²Se. The ratio of the signal intensities of the isotopes in the samples without the addition of the enriched isotope was compared to the natural ratio in order to check for spectral interference and mass discrimination. The correction factor for the mass discrimination was calculated by comparing the natural isotope ratio with the measured ratio for the sample without the spike. The software automatically uses the correction factor to correct the measured altered ratio.
2.3 Sample preparation
The samples analyzed are Certified Chinese Standard Material. The standard values were determined by Atomic Absorption Spectrum (AAS), Atomic Fluorescence Spectrum (AFS), Neutron Activation Analysis (NAA), Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), and Ion Chromatography (IC) method.
    An aliquot of approximately 250 mg of the sample was weighed directly in a PTFE flask and put in the microwave oven. Then, the solutions of the materials enriched with the isotopes ²⁰¹Hg and ⁷⁷Se were added to the flask in an adequate amount in order to obtain an altered ratio close to 1, to minimize the measurement errors. Due to the low concentrations of selenium and mercury in the samples, masses of the enriched materials at ng level were added. The masses were calculated using the ID equation for an altered ratio of 1 and converting the masses to volumes of the enriched materials solutions. The sample aliquot without spike was digested with 3.5 mL aqua regia plus 1.0 mL of deionized water in the microwave bombs. The sample aliquot with the spikes was also digested with 3.5 mL aqua regia plus a certain volume of deionized water, which was the complement to 1 mL of the volumes of the spiking solutions. In this way, all sample aliquots, without and with spikes, were digested in the same volume of liquid. The bombs were sealed tightly and then positioned in the carousel of the microwave oven. The system was operated at full power for 6.0 min. After the digestion was finished, all of the solutions were colorless, indicating an effective digestion. Following that, 1 mL of the digested sample was mixed with 1 mL of concentrated hydrochloric acid and heated to 90 °C for 30 min, in order to guarantee the lower oxidation state for selenium, favoring its hydride generation.^[7] After cooling the solution, the final volume of 10 mL was made up with deionized water, which was not strictly necessary, as isotopic dilution calibration was used, assuming that equilibration of the added isotope with the isotope in the sample have occurred during digestion. However, this final dilution to the same final volume for each sample aliquot, allowed a better control of the counting signals.
2.4 Analytical procedure
For the determination of mercury and selenium in samples, a 1 mL aliquot of the final sample solution (without or with spike), was transferred to the reaction flask of the hydride generator and the temperature program of the ETV was started. The glass tube of the hydride generator was manually introduced into the graphite tube. During the pre-heating step, the reducing agent was added to the reaction flask for 5 s and the generated vapors were transferred to the graphite tube by using argon for 30 s in the MHS-15. Before the vaporization step, the glass tube was removed and the graphite tip of the ETV arm closed the dosing hole of the graphite tube. Reading was carried out during the vaporization stage, in which the vapor is transported from the ETV to the plasma by a total argon flow rate of 1.16 L min^-1.

3.3 Figures of merit and analytical application
The limit of detection (LOD) is defined as the minimum concentration or weight of analyte that can be detected at a known confidence level. The limits of detection, in the sample, calculated as a function of the enrichment of the isotopic spike and the linear calibration detection limits for each isotope ²¹ are 0.5 and 2 ng g^-1 for mercury and selenium, respectively, indicating that the proposed procedure is able to detect very low concentrations of the analytes in tobacco samples. In a very recent paper, Santos et al.⁵ have investigated the chemical vapor generation for mercury and selenium in biological samples as slurries in different media, prior to detection by in ICP-OES, using external calibration. For 20 mg of solid sample in 15 mL of slurry, the obtained detection limits (3 s, n = 10) for the proposed procedure were 80 and 100 ng g^-1 for mercury and selenium, respectively, much higher (about 110 times for mercury and 33 times for selenium) than those obtained in this work, as expected due to the higher sensitivity of the ETV-ICP-MS technique. Vieira et al.¹⁸ have proposed a method for the determination of arsenic, mercury, selenium and tin in sediment by slurry sampling CVG-ETV-ICP-MS, also with trapping of the vapor on a treated graphite tube, but using slurry sampling and external calibration against aqueous standards, obtaining about the same LOD (3 s, n = 10) for mercury, 0.80 ng g^-1, as in this work, and 300 times better LOD for selenium, 0.01 ng g^-1.

Figure 1. Transient signals for Hg and Se in reference material GB-2102 (a) without and (b) with the addition of the enriched isotope.

    In Figure 1 some typical transient signals for mercury and selenium in reference material GB-2102, before and after addition the enriched isotopes, are shown. As shown in the Figure the condition of an altered isotopic ratio close to 1 was obtained for mercury and selenium.
    The accuracies of the procedures were estimated by the analysis of fix certified reference materials. As shown in Table 3, the found concentrations are in agreement with the certified values, according to the t-test for a confidence level of 95%. The relative standard deviations (RSD) were the range from 2.8 to 3.4% for selenium and from 2.8 to 4.1%, for mercury, indicating an adequate precision. Probably, an efficient digestion of the sample, with complete destruction of the analytes organic compounds by microwave digestion with aqua regia was attained, providing the equilibration of the added isotopes with the isotopes in the sample, which is a basic requirement for the ID calibration. After equilibration, there is no need to use an exact volume of the sample solution.

Table 3 Determined concentrations in m g g^-1 for mercury and selenium in certified tobacco samples by CVG-ETV-ICP-MS using isotope dilution calibration after microwave digestion with aqua regia and treatment with hydrochloric acid to reduce the oxidation number of selenium (n = 5)

4 CONCLUSIONS
It was demonstrated that mercury and selenium can be determined in tobacco samples after microwave-assisted acid digestion with aqua regia followed by their determination by CVG-ETVICP-MS using vapor trapping in the graphite tube and isotopic dilution calibration. The treated tube with Ir was efficient for the trapping of the mercury cold vapor and the selenium hydride. The proposed procedure using ID compensate eventual analyte loss during the steps of sample preparation, including digestion and reduction of the oxidation state for selenium. The reduction of the oxidation state for selenium with hydrochloric acid prior to its hydride generation is necessary, even using isotope dilution calibration. For mercury, the reduction step with hydrochloric acid can be omitted.

Acknowledgement  This work was supported by the National Natural Science Foundation of China (20471051) and the Key Natural Science Foundation of Yunnan Province (2005K021)

REFERENCES
[1] Chen Z Y, Kong WS, Liu W et al. Chinese. J. Tob. Res. 2003, 2, 115
[2] Matusiewicz, H.; Sturgeon, R. E.; Berman, S. S.; J. Anal. At. Spectrom. 1989, 4, 323.
[3] Vieira, M. A.; Ribeiro, A. S.; Curtius, A. J.; Anal. Bioanal. Chem. 2004, 380, 570.
[4] Cal-Prieto M J, Felipe-Sotelo M, Carlosena A et al. Talanta 2002, 56, 1.
[5] Dos Santos, E. J.; Herrmann, A. B.; Frescura, V. L. A.; Curtius, A. J.; Anal. Chim. Acta 2005, 548, 166.
[6] Welz, B.; Sperling, M.; Atomic Absorption Spectrometry, 3^rd ed., Wiley-VCH: Weinheim, 1999.
[7] Petersson, J.; Olin, A.; Talanta 1991, 38, 413.
[8] Jin, Q.; Liang, F.; Zhang, H.; Zhao, L.; Huan, Y.; Song, D.; Trends Anal. Chem. 1999, 16, 479.
[9] Brisbin, J. A.; Caruso, J. A.; Analyst 2002, 127, 921.
[10] Houk, R. S.; Anal. Chem. 1986, 58, 97A.
[11] Dias, L. F.; Saint’Pierre, T. D.; Maia, S. M.et al. Spectrochim. Acta, Part B 2002, 57, 2003.
[12] Pozebon, D.; Dressler, V. L.; Curtius, A. J.; J. Anal. At. Spectrom. 1998, 13, 1101.
[13] Saint' Pierre, T. D.; Maranh?o, T. A.; Frescura, V. L. A.; Curtius, A. J.; Spectrochim. Acta, Part B 2005, 60, 605.
[14] Vanhaecke, F.; Boonen, S.; Moens, L.; Dams, R.; J. Anal. At. Spectrom. 1997, 12, 125.
[15] Tibi, M.; Heumann, K. G.; J. Anal. At. Spectrom. 2003, 18, 1076.
[16] Chang, C.; Jiang, S.; J. Anal. At. Spectrom. 1997, 12, 75.
[17] Maia, S. M.; Pozebon, D.; Curtius, A. J.; J. Anal. At. Spectrom. 2003, 18, 330.
[18] Vieira, M. A.; Saint' Pierre, T. D.; Welz, B.; Curtius, A. J.; J. Anal. At. Spectrom. 2004, 19, 297.
[19] Vieira, M. A.; Ribeiro, A. S.; Dias, L. F.; Curtius, A. J.; Spectrochim. Acta, Part B 2005, 60, 643.
[20] Vieira, M. A.; Curtius, A. J.; Welz, B.; Spectrochim. Acta, Part B 2002, 57, 2057.
[21] Yu, L. L.; Fassett, J. D.; Guthrie, W. F.; Anal. Chem. 2002, 74, 3887.