Consequences of Vapor Enhancement on Selenium Speciation

These data suggested that NIES CRM 18 may contain part of its selenium as .... 11.97 ± 0.14 μg of Se L-1, measured, 11.7 ± 0.3 μg of Se L-1, mean ...
0 downloads 0 Views 118KB Size
Download PDF
Anal. Chem. 2006, 78, 8569-8574

Consequences of Vapor Enhancement on Selenium Speciation Analysis by HPLC/ICPMS Dijana Juresa, Doris Kuehnelt, and Kevin A. Francesconi*

Institute of ChemistrysAnalytical Chemistry, Karl-Franzens University Graz, Universitaetsplatz 1, 8010 Graz, Austria

Recent work has shown the presence of volatile selenium metabolites in human urine and suggested that these compounds could compromise quantitative selenium analyses by ICPMS. We show that with a commonly used sample introduction system (pneumatic nebulizer and spray chamber), two volatile selenium species recently identified in urine, namely, dimethyl selenide and dimethyl diselenide, gave greatly increased ICPMS responses (up to 58-fold) relative to selenite, an effect related to their volatilization in the spray chamber resulting in enhanced transport to the plasma. The quantitative consequences of this effect were demonstrated by measurement of total selenium and selenium species in certified reference material, NIES CRM 18 human urine. Direct flow injection analysis of the urine gave a total selenium concentration more than 2-fold higher than the certified value. These data suggested that NIES CRM 18 may contain part of its selenium as volatile species, and subsequent reversed-phase HPLC/ICPMS showed the presence of dimethyl selenide in addition to selenosugars and trimethylselenonium ion. Although the practice of quantifying unidentified chromatographic peaks against those of known compounds is common in speciation analysis, this approach when applied to NIES CRM 18 gave a value for the sum of selenium species which was twice the certified total selenium concentration. This work shows that the presence of volatile selenium species in urine precludes the use of flow injection analysis for total selenium measurements and imposes severe restrictions on the quantification of urinary selenium metabolites. In addition, it raises broader issues of the validity of the “dilute and shoot” approach to the determination of metals in clinical analysis of biological fluids. Understanding the biochemistry of the essential trace element selenium is important in human health because detrimental effects are commonly observed at exposure levels both just below and above a narrow optimum range. The selenium concentration in urine has long been recognized as a good marker of selenium body status,1 and determination of selenium species in urine, such as selenosugars and trimethylselenonium ion, may provide valuable additional information. * To whom correspondence should be addressed. Phone: +43 316 380 5301. Fax: +43 316 380 9845. E- mail: [email protected]. (1) Alaejos, M. S.; Romero, C. D. Clin. Chem. 1993, 39, 2040. 10.1021/ac061496r CCC: $33.50 Published on Web 11/15/2006

© 2006 American Chemical Society

Inductively coupled plasma mass spectrometry (ICPMS) is increasingly being used for determining both total selenium concentrations in urine and, when concatenated with an HPLC, urinary selenium species. The robustness and sensitivity of ICPMS as a selenium-selective detector offers the potential to analyze urine samples directly without sample pretreatment. The literature shows, however, an almost complete absence of quantitative data on selenium species in urine, and total selenium concentrations by direct ICPMS analysis of urine have been presented in only three studies,2-4 two of which reported high results in some samples.2,3 A prerequisite for quantitative analyses performed directly on urine samples is that the selenium species must produce the same signal response in the ICPMS. Factors influencing ICPMS response for the various selenium species are their transport to and decomposition in the plasma. Several studies have examined these factors with outcomes that suggest the effect depends on the type of nebulizer employed. For example, differences in responses for selenium species have been reported with an ultrasonic nebulizer5-7 or a nebulizer with a membrane desolvator,8 but no significant differences between species were observed with commonly used concentric nebulizers,5 a microconcentric or direct injection nebulizer,8 or a cross-flow nebulizer.6,9 There has been no study of the specific (separate) effect of transport efficiency on the responses for various selenium species. In our recent studies on selenium urinary metabolites, we had considerable difficulties with reliable quantification of selenium and selenium species when applying ICPMS and HPLC/ICPMS directly on the urine samples,10 and preliminary observations suggested that the problem stemmed from the presence of small and highly variable amounts of volatile selenium species. The present study investigates the ICPMS response of 12 selenium compounds with the focus on two volatile species, namely, dimethyl selenide and dimethyl diselenide, previously shown to be in some urine samples.10 Furthermore, we demonstrate the considerable practical consequences of the results from this study (2) Nakagawa, J.; Tsuchiya, Y.; Yashima, Y.; Tezuka, M.; Fujimoto, Y. J. Health Sci. 2004, 50, 164. (3) Wang, J.; Hansen, E. H.; Gammelgaard, B. Talanta 2001, 55, 117. (4) Heitland, P.; Ko ¨ster, H. D. J. Anal. At. Spectrom. 2004, 19, 1552. (5) Li, F.; Goessler, W.; Irgolic, K. J. J. Chromatogr., A 1999, 830, 337. (6) Gammelgaard, B.; Jøns, O. J. Anal. At. Spectrom. 2000, 15, 499. (7) Yang, K.-L.; Jiang, S.-J. Anal. Chim. Acta 1995, 307, 109. (8) Bendahl, L.; Gammelgaard, B. J. Anal. At. Spectrom. 2005, 20, 410. (9) Gammelgaard, B.; Jøns, O. J. Anal. At. Spectrom. 1999, 14, 867. (10) Juresa, D.; Darrouze`s, J.; Kienzl, N.; Bueno, M.; Pannier, F.; Potin-Gautier, M.; Francesconi, K. A.; Kuehnelt, D. J. Anal. At. Spectrom. 2006, 21, 684.

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006 8569

Figure 1. Structures of selenium compounds investigated in this study.

by analyzing selenium and selenium species directly in the certified reference material, NIES CRM 18 Human urine. EXPERIMENTAL SECTION Chemicals and Reagents. The 12 selenium compounds (Figure 1) investigated were as follows: sodium selenite (>99%) obtained from Merck (Darmstadt, Germany); sodium selenate (>99%), dimethyl selenide (DMSe, >99%), and D,L-selenomethionine (>99%) from Fluka (Buchs, Switzerland); dimethyl diselenide (DMDSe, >99%) from Acros Organics (Geel, Belgium); selenocystamine and D,L-selenocystine from Sigma-Aldrich (Vienna, Austria); trimethylselenonium ion (TMSe) as the iodide synthesized in-house using Hoffman’s11 procedure; and selenosugars 1 and 2,12 selenosugar 3,13 and methaneseleninate10 synthesized as previously reported. Other chemicals and reagents used were as reported in the previous study of Juresa et al.10 Reference Materials and Solutions. The urine certified reference material, NIES CRM 18 Human urine, was obtained (11) Hoffman, J. L. J. Chromatogr., A 1991, 588, 211. (12) Traar, P.; Belaj, F.; Francesconi, K. A. Aust. J. Chem. 2004, 57, 1051. (13) Kuehnelt, D.; Kienzl, N.; Traar, P.; Le, N. H.; Francesconi, K. A.; Ochi, T. Anal. Bioanal. Chem. 2005, 383, 235.

8570

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

from the National Institute for Environmental Studies (Ibaraki, Japan); standard reference material NIST SRM 1643e trace elements in water was obtained from the National Institute of Standards and Technology (Gaithersburg, MD); the stock standard solution of 1000 mg of Se L-1 as selenite used for calibration for total selenium by flow injection analysis ICPMS (FIA/ICPMS) was purchased from CPI (Santa Rosa, CA). Preparation of Standard Solutions of Selenium Compounds. From stock solutions (80-1000 mg of Se L-1), solutions with 10 mg of Se/L were prepared in water for 10 of the selenium species; DMDSe and DMSe, because of their nonpolar volatile nature, were prepared in ethanol at concentrations of 100 mg of Se L-1. Aliquots (0.5 or 0.05 mL, equivalent to 50 µg of Se for each compound) were mineralized as described below. The mineralized solutions were quantitatively transferred to polypropylene tubes and diluted with water to 50 mL or neutralized with NH3 and diluted with water to 50 mL to give a final nominal concentration of 100 µg of Se L-1. Similarly, solutions of 100 µg of Se L-1 of the intact (not mineralized) selenium compounds were prepared in 4% HNO3 or with neutralized (by addition of aqueous NH3) 4% HNO3. This neutralization procedure prior to addition of the compound was necessary for some of the compounds, namely, selenosugars 1 and 2, DMSe, and DMDSe, because in 4% HNO3 they were unstable and rapidly degraded to methaneseleninate (shown by HPLC/ICPMS). Although selenosugar 3 and methaneseleninate were stable in 4% HNO3, they were also prepared in neutralized HNO3 as a way of standardizing the procedure. All samples were prepared in triplicate. ICPMS responses were compared for the prepared solutions of diluted versus mineralized compounds by FIA/ICPMS under four different conditions (different nebulizers and flow rates, see below). Additionally, solutions (one from each of the triplicates) were analyzed by HPLC/ICPMS in order to confirm that (i) after dilution in the acid or neutralized acid matrix the species remained unchanged and (ii) after the mineralization step the species had been completely converted into inorganic selenium. The two volatile species, DMDSe and DMSe, were difficult to handle at the working concentrations (100 µg of Se L-1) employed in our experiments, and handling losses (ca. 20%) were observed during preparation of the working solutions from the stock solutions. This handling loss was quantified by transferring aliquots simultaneously from the working solution to two HPLC vials, which were immediately capped. One of the vials was used in the response comparison experiments and contained the appropriate matrix, while the second vial contained 50% HNO3 to quickly decompose the species to methaneseleninate. Quantification of the selenium in the vial containing HNO3, now present as methaneseleninate, provided an accurate concentration of the selenium in the experimental vial, present as the volatile DMDSe or DMSe. Microwave-Assisted Acid Mineralization. Aliquots (0.05 mL of 100 mg of Se L-1 of DMDSe and DMSe solutions; 0.5 mL of 10 mg of Se L-1 of other selenium compound solutions; and 0.5 mL of NIES CRM 18 Human urine) were placed in a 12 mL quartz vial of an autoclave mineralization system (ultraCLAVE 2, EMLS, Leutkirch, Germany) which contained 2 mL of HNO3, and water (1.95 mL for DMDSe and DMSe; 1.5 mL for other selenium compounds and for certified reference material) was added. The

autoclave was closed, loaded with argon to 4 × 106 Pa, and the mixture heated at 250 °C for 30 min, after which it was quantitatively transferred to polypropylene tubes and diluted with water to 50 or 10 mL (NIES CRM 18). Determination of Selenium by FIA/ICPMS. FIA/ICPMS was performed with an Agilent 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) consisting of a binary pump, a vacuum degasser, and a thermostated autosampler with a variable 100 µL injection loop. Samples were kept at 4 °C throughout the analysis. The mobile phase was 20 mM ammonium formate 3% MeOH at pH 3.0 (the same as that used for reversedphase HPLC in this study) at a flow rate of 1.0 or 0.5 mL min-1. The HPLC system was connected with PEEK (polyether ether ketone) capillary tubing (0.125 mm i.d.) to an Agilent 7500ce ICPMS equipped with an octopole reaction cell operated at a H2 flow rate of 3.5 mL min-1; solutions were introduced to the plasma via a Babington or microconcentric nebulizer and a Scott doublepass spray chamber cooled to 2 °C (capacity ca. 75 mL). When the microconcentric nebulizer was used, the flow of 1.0 mL min-1 was split by installing a T-piece after the HPLC system, and the flow to the nebulizer (0.27 or 0.53 mL min-1) was controlled with a peristaltic pump. Signals were monitored at m/z 77, 78, 80, and 82 and quantified using m/z 78. Quantification was performed by external calibration based on the peak area of standard solutions of selenium, prepared from the stock 1000 mg of Se L-1 solution purchased from CPI (Santa Rosa, CA) and in a matrix matching that of the samples (i.e., 4% HNO3 or neutralized 4% HNO3). The accuracy of the FIA/ICPMS method was checked by analysis of the standard reference material NIST SRM 1643e Trace elements in water (certified, 11.97 ( 0.14 µg of Se L-1, measured, 11.7 ( 0.3 µg of Se L-1, mean ( 1 SD, n ) 6). Determination of Selenium Species by HPLC/ICPMS. HPLC/ICPMS was performed with the Agilent 1100 HPLC system and an Agilent 7500ce ICPMS with a Babington nebulizer under conditions described above. Signals were monitored at m/z 77, 78, 80, and 82 using an integration time of 0.3 s. Quantification of identified compounds was performed against external calibration curves, based on m/z 78, of standard solutions of the respective compound, while unknown compounds and volatile compounds were quantified against selenosugar 1, a major urine selenium species. Selenium compounds in the solutions for ICPMS response analysis were investigated with three different chromatographic systems after dilution (1 + 9, v/v) with water. Anion-exchange chromatography was performed on a PRP-X100 column (4.1 mm × 100 mm) from the Hamilton Co. (Reno, NV) at 40 °C with an aqueous solution of citric acid (10 mmol L-1) at pH 4.8 (adjusted with aqueous ammonia) and a flow rate of 1.5 mL min-1 as the mobile phase. Some of the samples were analyzed by reversedphase chromatography, performed on a guard column for Waters Atlantis C18 column (4.6 mm × 20 mm; Waters Corp., Milford, MA) and a mobile phase of 20 mmol L-1 ammonium formate at pH 3.0, adjusted with formic acid, containing 3% MeOH at 30 °C at a flow rate of 1.0 mL min-1. For cation-exchange chromatography, a Hamilton PRP-X200 column (4.1 mm × 250 mm; Hamilton Co., Reno, NV) was used with a mobile phase of an aqueous solution of pyridine (10 mmol L-1) at pH 5.0, adjusted with formic acid, at 30 °C and a flow rate of 1.0 mL min-1.

Certified reference material NIES CRM 18 was reconstituted in a cool room at 4 °C with all laboratory ware and water cooled to 4 °C before use; it was analyzed after filtration through a 0.2 µm Nylon filter (ProFill, Markus Bruckner Analysentechnik, Linz, Austria) on a reversed-phase chromatography column Waters Atlantis C18 (4.6 mm × 150 mm) with a mobile phase of 20 mmol L-1 ammonium formate at pH 3.0, adjusted with formic acid, containing 3% MeOH at 30 °C and a flow rate of 1.0 mL min-1. To support the identification of selenium compounds present in the sample, the CRM urine was also analyzed using cation-exchange chromatography under the conditions described above. The presence of TMSe in the sample was additionally supported by HPLC/vapor generation/ICPMS under conditions reported elsewhere14 on a Shodex RSpak NN-614 column (6.0 mm × 150 mm; Showa Denko K. K., Kawasaki, Japan) at 30 °C with 20 mmol L-1 ammonium dihydrogen phosphate at pH 2.2 (adjusted with phosphoric acid) as mobile phase at a flow rate of 0.6 mL min-1. Injection volume was always 20 µL. Solvent Transport Efficiency. The solvent transport efficiency of the nebulizers was measured indirectly by nebulizing a known mass of mobile phase with the plasma on and weighing the mass of mobile phase collected at the drain of the spray chamber; the amount of sample transferred to the plasma was then estimated by difference. RESULTS AND DISCUSSION ICPMS Response of Selenium Species. To compare responses for the various selenium species we chose to prepare individual standard solutions and compare aliquots from these solutions before and after a microwave-assisted acid (HNO3) mineralization step. The assumptions were that mineralization to inorganic selenium was complete, no losses occurred during mineralization, and we would then be comparing the response of selenium in the intact selenium species with the equivalent quantity of selenium as inorganic selenium. These assumptions appeared valid because there was no indication of selenium lossessweighed (nominal) quantities of selenium species closely matched the measured values in most cases (Table 1)sand HPLC/ICPMS analysis of the mineralized samples showed that all the organic selenium species had been quantitatively converted into inorganic selenium. It was of interest, however, that the inorganic species after mineralization was selenite in all cases, even when selenate was the investigated species. One may have expected that use of HNO3 would have ensured that the selenium species would be mineralized to the most oxidized form, selenate. We have no ready explanation for this result; we note, however, that Wang et al.15 earlier reported that selenite is the selenium species formed after samples were treated with a microwave-assisted mineralization procedure using a H2O2/ H2SO4 mixture. The ICPMS responses for the intact selenium species and the equivalent amount of selenium as selenite were then compared by analyzing the solutions by flow injection analysis with a Babington nebulizer, which is the commonly used nebulizer for selenium speciation analysis with HPLC/ICPMS because it can handle high flow and high sample matrix. For the 10 polar species, (14) Kuehnelt, D.; Kienzl, N.; Juresa, D.; Francesconi, K. A. J. Anal. At. Spectrom. 2006, 21, 1264. (15) Wang, Z.; Gao, Y.-X.; Belzile, N. Anal. Chem. 2001, 73, 4711.

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

8571

Table 1. ICPMS Response of Different Selenium Species after Mineralization or Dilution Measured by Flow Injection/ICPMS Using a Babington Nebulizer and a Flow Rate of 1 mL min-1 a

Table 2. ICPMS Response of Volatile Selenium Species after Mineralization or Dilution Measured by Flow Injection/ICPMS with Different Nebulizer Transport Efficienciesa

measured [Se] in µg of Se L-1 compound Se(IV)b Se(VI)b selenomethionineb selenocystineb selenocystamineb TMSeb selenosugar 1c selenosugar 2c selenosugar 3c methaneseleninatec dimethyl diselenidec,d dimethyl selenidec,d

mineralized sample

diluted sample

99 ( 3 99 ( 1 96 ( 2 91 ( 1 97 ( 2 94 ( 3 102 ( 6 103 ( 3 102 ( 1 105 ( 4 83 ( 2 78 ( 1

100 ( 1 101 ( 1 96 ( 2 92 ( 1 95 ( 2 96 ( 1 99 ( 1 100 ( 3 97 ( 1 103 ( 2 3767 ( 253 4505 ( 79

ratio of measured [Se] in diluted samples to measured [Se] in mineralized sample nebulizer Babington microconcentric a

transport efficiency, %

dimethyl diselenide

dimethyl selenide

2.5 3.5 11 16

45 ( 4 31 ( 3 4.8 ( 0.5 1.9 ( 0.2

58 ( 2 39 ( 1 8.8 ( 0.3 4.4 ( 0.2

Values represent mean ( SD, n ) 3 individually prepared solutions.

a Values represent mean ( SD, n ) 3 individually prepared solutions. Prepared in 4% HNO3. c Prepared in neutralized 4% HNO3. d Values adjusted to accommodate ca. 20% handling losses experienced on preparation of working solutions from stock solutions (see Experimental Section).

b

the measured (mineralized) value, quantified against a standard selenium solution of selenite, always agreed within 10% of the nominal concentration (Table 1), and more importantly, comparison of diluted (intact selenium species) and mineralized (selenite) samples showed essentially identical responses in each case except for selenosugar 3 which showed a small (within 3% of each other) difference. On the basis of these data, use of selenite (or any of the other nine selenium species investigated) as a general quantification standard in speciation analysis by HPLC/ICPMS seems justified. Similar experiments performed with the two volatile species, however, produced markedly different results: whereas measured concentrations of the mineralized samples matched reasonably well the nominal value (they were each about 20% low owing to the difficulty in handling small quantities of volatile compounds), the response for the diluted samples was up to 58-fold higher than that for the mineralized samples (Table 1). We note that although the stock solutions of the volatile species were prepared in ethanol, subsequent dilutions were performed in water, resulting in a working solution containing 0.1% ethanol which contributed a negligible amount of signal enhancing carbon compared with the 3% methanol contained in the mobile phase used in the FIA measurements. Little work has been published on the consequences of volatility of species on their response when measured by ICPMS, but the studies of Botto and Zhu16 and by Langlois et al.17 are particularly informative in this regard. Botto and Zhu16 investigated the analysis of organic solvents for trace elements using an ultrasonic nebulizer with membrane desolvation and reported that only nonvolatile forms of the elements could be determined accurately. Langlois et al.17 compared the effect of two different nebulizers on the ICPMS response for iodine from volatile (CH3I) and nonvolatile (NaI) species and demonstrated that the 4-6(16) Botto, R. I.; Zhu, J. J. J. Anal. At. Spectrom. 1994, 9, 905. (17) Langlois, B.; Dautheribes, J.-L.; Mermet, J.-M. J. Anal. At. Spectrom. 2003, 18, 76.

8572 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

Figure 2. Reversed-phase HPLC/ICPMS chromatogram of NIES CRM 18. (Inset) NIES CRM 18 (solid line) and the same sample spiked with approximately 1 µg of Se L-1 as DMSe (dotted line). Conditions: Waters Atlantis C18 (4.6 mm × 150 mm; T ) 30 °C; 20 mmol L-1 ammonium formate, 3% MeOH, pH 3.0; flow rate, 1.0 mL min-1; injection volume, 20 µL).

fold greater response obtained for CH3I was related to evaporation in the spray chamber. We similarly tested the two volatile selenium species with both a Babington nebulizer and a microconcentric nebulizer at varying flow rates which equated to different transport efficiencies ranging from 2.5% to 16%. As the transport efficiency increased, the ratio of responses for the volatile intact selenium species and the equivalent quantity of selenium present as selenite decreased, so that at 16% transport efficiency the ratio was 1.9 for DMDSe and 4.4 for DMSe (Table 2). Moreover, the data show that this effect is greater for the more volatile DMSe (bp 57-58 °C) than for DMDSe (bp 155-157 °C) and support the conclusion that higher ICPMS response of volatile species is caused by evaporation within the spray chamber resulting in more efficient analyte transport to the plasma. For convenience, we will refer to this as a vapor enhancement effect. Consequences of the Vapor Enhancement Effect on the Quantification of Selenium and Selenium Species in Urine. We examined the practical consequences of these effects by analyzing the urine certified reference material NIES CRM 18 by FIA/ICPMS directly and after mineralization. Whereas the mineralized samples ([Se] ) 62 ( 4 µg L-1, n ) 3) matched well the

Figure 4. HPLC/vapor generation/ICPMS chromatogram of NIES CRM 18 sample (solid line) and the same sample spiked with 1.5 µg of Se L-1 as TMSe (dotted line) on the Shodex RSpak NN-614 (6.0 mm × 150 mm, T ) 30 °C; 20 mmol L-1 ammonium dihydrogen phosphate, pH 2.2; flow rate, 0.6 mL min-1; vapor generation conditions, 1% NaBH4 in 0.1 mol L-1 NaOH pumped with 0.5 mL min-1, 0.1 mol L-1 HCl pumped with 0.5 mL min-1; injection volume, 20 µL).

Figure 3. Chromatograms (HPLC/ICPMS) of NIES CRM 18 (solid line) and the same sample spiked with 3 µg of Se L-1 as TMSe, 5 µg of Se L-1 as selenosugar 3, and 20 µg of Se L-1 as selenosugar 1 (dotted line): (a) Reversed-phase chromatography (Waters Atlantis C18, 4.6 mm × 150 mm; T ) 30 °C; 20 mmol L-1 ammonium formate, 3% MeOH, pH 3.0; flow rate, 1.0 mL min-1; injection volume, 20 µL). (b) Cation-exchange chromatography (Hamilton PRP-X200, 4.1 mm × 250 mm; T ) 30 °C; 10 mmol L-1 pyridine, pH 5.0; flow rate, 1.0 mL min-1; injection volume, 20 µL).

certified selenium concentration (59 ( 5 µg L-1), the samples analyzed without mineralization gave results ([Se] ) 145 ( 7 µg L-1, n ) 4) more than 2-fold higher than the certified value. These data suggested that volatile selenium species might be present in the reconstituted urine CRM, and subsequent reversedphase HPLC/ICPMS (Figure 2) revealed two significant, longretained peaks (tR ) 20 and 74 min) in addition to several early peaks, three of which were identified as TMSe and selenosugars 1 and 3 (see below). The peak with tR ) 20 min was identified as DMSe by co-chromatography with standard material (Figure 2, inset). The peak with tR ) 74 min did not match standard DMDSe (tR ) 115 min) but was consistent with that expected for dimethyl selenenylsulfide (DMSeS), a species recently tentatively identified in urine by head space solid-phase microextraction and GC/MS.10 There is no commercially available standard of DMSeS; thus, identification of the peak at tR ) 74 min could not confirmed in this current ICPMS study.

TMSe, selenosugar 1, and selenosugar 3 were identified in the CRM urine by spiking experiments conducted under two sets of chromatographic conditions (Figure 3 a,b). Additionally, TMSe was determined by HPLC/vapor generation/ICPMS which, under certain conditions,14 is selective for TMSe (Figure 4). The selenosugars and TMSe were quantified against external calibration based on the individual compounds, and the values obtained were as follows: selenosugar 1, 18.6 ( 0.2 µg of Se L-1, n ) 4; selenosugar 3, 4.9 ( 0.1 µg of Se L-1, n ) 4; and TMSe, 1.7 ( 0.1 µg of Se L-1, n ) 4. In a previous study of selenium species in NIES CRM 18, Chatterjee et al.18 reported TMSe to be present at 3.04 ( 0.5 µg of Se L-1; selenosugars were not identified in that study, however, presumably owing to a lack of appropriate standards. Chatterjee et al.18 also reported selenite in the CRM urine, but we could find no evidence for the presence of selenite in our sample. Other selenium compounds were also present, but their chromatographic behavior did not match that of selenium standards available to us. We did not attempt to precisely quantify the volatile selenium species. Our experiments with flow injection analysis showed that this would require calibration against the same species, and it was not feasible to prepare standard solutions of the volatile compounds at the relevant low concentrations (sub-µg of Se L-1). However, to illustrate the significance of the vapor enhancement effect on quantitative selenium speciation analysis, we quantified all signals against selenosugar 1, a practice previously employed for unknowns in our laboratory. Under these conditions, DMSe was quantified at 64.3 ( 1.1 µg of Se L-1, n ) 4, the signal at tR ) 74 min, tentatively assigned to DMSeS, was quantified at 20.5 ( 3.6 µg of Se L-1, n ) 4, and the sum of all chromatographed species totaled 120 ( 2 µg of Se L-1, n ) 4, which is twice the certified total selenium concentration for this reference material. (18) Chatterjee, A.; Shibata, Y.; Tao, H.; Tanaka, A.; Morita, M. J. Chromatogr., A 2004, 1042, 99.

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

8573

More work is required to evaluate practical solutions to the problems of quantification for both total selenium and selenium species analyses. Because the volatile species appear to be present in only trace amounts but give disproportionately large signals, the problem might best be tackled by removing the volatiles by argon purging prior to analysis. The observations reported here have possible implications beyond selenium analyses in urine because sample preparation procedures based on simple dilution of biological fluids, so-called “dilute and shoot” methods, are increasingly being used in clinical analysis.19-22 Since dimethyl selenide is also excreted by exhalation,23 it might be detectable in blood samples, and thus, determination of selenium status by direct ICPMS measurement of (19) Barany, E.; Bergdahl, I. A.; Schu ¨ tz, A.; Skerfving, S.; Oskarsson, A. J. Anal. At. Spectrom. 1997, 12, 1005. (20) de Boer, J. L. M.; Ritsema, R.; Piso, S.; van Staden, H.; van den Beld, W. Anal. Bioanal. Chem. 2004, 379, 872. (21) Goulle´, J.-P.; Mahieu, L.; Castermant, J.; Neveu, N.; Bonneau, L.; Laine´, G.; Bouige, D.; Lacroix, C. Forensic Sci. Int. 2005, 153, 39. (22) Heitland, P.; Ko ¨ster, H. D. Clin. Chim. Acta 2006, 365, 310. (23) Kremer, D.; Ilgen, G.; Feldmann, J. Anal. Bioanal. Chem. 2005, 383, 509.

8574

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

blood plasma or serum using the dilute and shoot methodology is likely to meet the same quantification problems experienced with urine. Indeed, a study using ICPMS to directly measure selenium concentrations in diluted blood or serum reported high values (15-50% above recommended values) for all five reference materials investigated.19 Furthermore, volatile species of other elements may also be present in biological fluids, so dilute and shoot methodology for these elements may also need to be carefully checked before being implemented in routine clinical analyses. ACKNOWLEDGMENT We thank the Austrian Science Fund (FWF Project 16816-N11) for financial support and Pedro Traar for preparing the selenosugar and methaneseleninate standards.

Received for review August 11, 2006. Accepted October 4, 2006. AC061496R