Anal. Chem. 1994,66, 1316-1322
Determination of Tellurium in Urine by Isotope Dilution Gas Chromatography/Mass Spectrometry Using (4-Fluorophenyl)magnesium Bromide as a Derivatizing Agent and a Comparison with Electrothermal Atomic Absorption Spectrometry Suresh K. Aggarwai,tl* Michael Kinter,s James Nichoison,g and David A. Herold'~t~l1 VA Medical Center-1 13, San Diego, California 92 16 1, Fuel Chemistry Division, Bhabha Atomic Research Center, Trombay, Bombay 400 085, India, Department of Pathology, University of Virginia, Charlottesville, Virginia 22908, and Department of Pathology, University of California-San Diego, La Jolla, California 92037
The antitumor drug AS- 101 [ammoniumtrichloro(dioxoethylene-O,O')tellurate(IV)] is the first tellurium-containing compound that has been identified as possessing immunomodulating properties and minimal toxicity. We have developed a stable isotope dilution gas chromatography/mass spectrometry method using 120Te as an internal standard and (4fluoropheny1)magnesium bromide as a derivatizing agent for Te determination in urine. The urine samples were digested using HNO3 H202 prior to derivatization with lithium bis(trifluoroethy1)dithiocarbamate at a pH of 3. The trifluorodiethyldithiocarbamate of tellurium was reacted with the Grignard reagent in anhydrous diethyl ether to obtain Te(FC&)2 for GC/MS analysis. All isotope ratio measurements were made by selected ion monitoring with a Finnigan MAT 8230 organic mass spectrometer using a 10-m fused silica capillary column. Overall percision values for the five major Te isotopes relative to ljOTe were 0.6-3.1% when 10-ng samples of chelated Te were analyzed. No appreciable memory or carry-over effect was observed when two synthetic mixtures differing in W e : l q e ratios by a factor of 50 were sequentially analyzed. The isotope dilution GC/MS method was validated by determining Te in urine samples and comparing the values with electrothermal atomic absorption spectrometry. Te concentrations were determined in the 100-500 pg/L range with CVs of 1-4%.
+
Tellurium is a rare trace element with previous interest mainly limited to the semiconductor fabrication industry. Interest in the metabolism of this element has assumed importance due to a new drug. The antitumor drug AS-101 [ammoniumtrichloro(dioxoethylene-0,0~tellurate(IV)]is the first Te-containing compound that has been identified to possess immunomodulating properties and minimal toxicity. 1,2 In view of the extremely low doses proposed for use in t VA Medical Center. 8 Bhabha Atomic Research 1 University of Virginia
Center.
of California. (1) Sredni, B.; Caspi, R. R.; Klein, A.; Kalcchman, Y.; Danziger, Y.; BenYa'akov, M.; Tamari, T.; Shalit, F.; Albeck, M. Nature 1987, 330, 173-176. (2) Nyska, A.; Waner,T.;Pirak, M.; Albock, M.; Sredni, B. Arch. Toxicol. 1988, 63, 386-393. I University
1316 Analytical Chemistry, Vol. 66,No. 8, April 15, 1994
laboratory animals and humans, it is necessary to develop sensitive, precise, and accurate analytical methods for Te determination in biological specimens. Very few studies are reported in the literature for Te determination in biological samples. Siddik and Newman3 published an electrothermal atomic absorption spectrometry (EAAS) method using platinum as a modifier for Te determination in biological samples to facilitate pharmacokinetic studies with AS-101. Detection limits of 50, 5 , and 170 ng of Te/mL (or g) were observed for urine, plasma, and tissues, respectively. These authors stressed the importance of the volatile nature of Te, which requires the use of either a low ashing temperature or a suitable chemical modifier for thermal stabilization of Te at the higher ashing temperature. An electrothermal vaporization inductively coupled plasma source mass spectrometry (ETV-ICPMS) technique using hyamine hydroxide as a matrix modifier was also explored by Newman and co-workers for Te determination in biological fluid^.^ Detection limits of Te, using the isotope 13@Te,in the urine and plasma were reported to be 2.7 and 5.7 ng/mL, respectively. Wittmaack and co-workers5 compared the use of stable enriched (124Te, 126Te)and radioactive (121mTe, 123mTe) isotopes for determining the biokinetic parameters of Te in rabbits. These authors used EAAS for total Te measurement and secondary ion mass spectrometry (SIMS) for determining isotope ratios of stable Te isotopes. However, due to restrictions limiting the use of radioactive isotopes for studies on human subjects, it is becoming increasingly important to use stable enriched isotopes for studying the intestinal absorption and the metabolic behavior of trace elements and therefore mass spectrometry techniques are becoming important. A detailed report on the Te content of vegetation has also been published recentlye6 Other methods that have been used for determining Te are flame atomic absorption spectrometry in samples with high Te content or conversion of the metal to gaseous hydride for EAAS. (3) Siddik, Z. H.; Newman, R. A. A d . Biochem. 1988, 172, 190-196. (4) Newman, R. A,; Osborn, S.;Siddik, 2.H. Clin. Cfiim.Acta 1989,179, 191196. ( 5 ) Kron, T.; Wittmaack, K.; Hansen, C.; Werner, E. Anal. Chem. 1991,63,26032607. (6) Cowgill, U. M. B i d . Trace Elem. Res. 1988, 17, 43-67.
0003-2700/94/0366-13~6~04.50/0
0 1994 American Chemlcal Society
In our laboratory, we have been developing isotope dilution gas chromatography/mass spectrometry (GC/MS) methods for trace element determination in biological ~amples.~-20 We have been motivated to develop GC/MS methods for trace elements due to the increased availability of benchtop instruments equipped with automated sample injector systems in biomedical, clinical, and environmental laboratories engaged in determining organic compounds. This GC/MS methodology would provide these laboratories with an independent definitive analytical technique in addition to the commonly used EAAS methodology, thus making the GC/MS approach attractive and easy to use compared to other specialized mass spectrometric techniques such as thermal ionization mass spectrometry (TIMS), ICPMS, and SIMS. In addition, the use of the isotope dilution technique provides two important advantages: (1) there is freedom from matrix effects and (2) precision and accuracy of analyses are not affected by incomplete recovery. This is particularly important for a volatile element like Te. Since the mass spectrometry approach is based on themeasurement of isotope ratios after the addition of internal standard, the results are not degraded due to losses resulting from high volatility or other chemical processing. Nonetheless, it is preferred to have the maximum recovery and minimize the losses when dealing with samples containing extremely small amounts of the trace element. As the method employs an ideal internal standard i.e., an enriched isotope of the same element, a high degree of accuracy can be readily achieved in concentration determinations. One of the problems preventing the widespread use of GC/ MS for trace element determination has been the lack of recognition of suitable chelating agents. Other limitations have been the relatively poor precision of the isotope ratio data compared to TIMS and thechemistry involved in digestion and metal chelate preparation. The precision values obtained for isotope ratio measurements by TIMS cannot be surpassed by any other mass spectrometric technique, but GC/MS provides a workable compromise between the precision obtained and the chemistry involved in sample preparation. In order to exploit the high precision of TIMS, extensive sample preparation and purification is mandatory. However, the (7) Aggarwal, S.K.; Kinter, M.; Wills, M. R.; Savory, J.; Herold, D. A. Anal. Chim. Acta 1989, 224, 83-95. (8) Aggarwal, S.K.; Kinter, M.; Wills, M. R.; Savory, J.; Herold, D. A. Anal. Chem. 1989, 61, 1099-1103. (9) Aggarwal, S. K.; Kinter, M.; Wills, M. R.; Savory, J.; Herold, D. A. A w l . Chem. 1990,62, 111-115. (10) Aggarwal, S.K.; Kinter, M.; Herold, D. A. J . Am. SOC.Mass Spectrom. 1991, 2, 85-90. (1 1) Aggarwal, S. K.; Kinter, M.; Herold, D. A. A w l . Biochem. 1991,194, 140145. (12) Aggarwal, S. K.; Kinter, M.; Herold, D. A. J . Chromatogr. Biomed. Appl. 1992, 576, 297-304. (13) Aggarwal, S.K.; Kinter, M.; Herold, D. A. Anal. Biochem. 1992,202,367374. (14) Herold, D. A.;Aggarwal, S.K.; Kinter, M. Clin. Chem. 1992,38,1647-1649. (15) Aggarwal, S. K.; Orth, R. G.; Wendling, J.; Kinter, M.;Herold, D.A. J . A w l . Toxicol. 1993, 17, 5-10. (16) Aggarwal, S. K.; Gemma, N . W.; Kinter, M.; Nicholson, J.; Shipe, J. R., Jr.; Herold, D. A. Anal. Biochem. 1993, 210, 113-1 18. (17) Aggarwal, S.K.; Kinter, M.; Herold, D. A. Proceedings of the 39th ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, TN, May 19-24, 1991; p 226. (18) Aggarwal, S. K.; Kinter, M.; Herold, D. A. Proceedings of the 40th ASMS Conferenceon Mass Spectrometry and Allied Topics, Washington, D.C., May 31-June 5, 1992; p 684. (19) Baird, G. S.;Herold, D. A.; Aggarwal, S.K.; Fitzgerald, R. L. Proceedings of the 41st ASMS Conference on Mass Spectrometry and Allied Topics, San Francisco, CA, May 30-June 4, 1993; p 736. (20) Aggarwal, S.K.; Kinter, M.; Herold, D. A. B i d . Truce Elem. Res., in press.
uncertainties due to sample variability, particularly in the case of biological samples, can override the advantage of high precision achievable by TIMS. Another popular mass spectrometry technique for trace element determination, ICPMS, has the advantage of multielement capability and rapid analysis. A number of laboratories are trying to develop this promising technique for biological specimens. However, in general, there are a number of isobaric interferences from oxides, hydroxides, argides (argon adducts), etc. Further, the presence of large amounts of salts suppress (sometimes enhance) the analyte signal, and this may necessitate the digestion of biological samples and a purification step. In one of the studies reported for the determination of Te compounds21 in waste water streams by liquid chromatography/ICPMS using 100 ppb lo3Rh as an internal standard, the peaks of 125Teand lz8Te isotopes were observed to increase when the same sample was analyzed repeatedly. It took 1 h to bring the peaks at these masses down to the base line level. Selective volatilization of the Te-containing compounds during nebulization in ICPMS was thought to be responsible for this. GC/MS is free from these problems since the metal chelate formation provides for purification as well as preconcentration of trace element. In addition, the measurements are done in the high-mass range, corresponding to metal chelate ions, and the region of high mass is free from any isobaric or other interferences. Furthermore, the GC column provides additional selectivity owing to differences in the retention times of different metal chelates. GC/MS using a suitable chelating agent also22 offers the potential to determine more than one trace element simultaneously. For successful use of any metal chelate for isotope ratio or concentration determination by GC/MS, there should be no carry-over or memory effect. Freedom from the memory effect is imperative for the accurate determination of a wide range of isotope ratios in unknown samples. We have been addressing this vital point in our investigations and have demonstrated the applicability of lithium bis (trifluoroet hy1)dithiocarbamate [Li(FDEDTC)] chelating agent for Ni,8 Cr9, Co,12 Pt,1°J6 and Hg.20 However, strong memory effects were observed for Cu,ll Pb,17-19 and Zn7 and a small memory effect was observed for Cd,l5 and this led to the exploration and development of alternative chelating agents. It is interesting to note that, in addition to the work being pursued in our laboratory, an increasing number of researchers are using GC/MS for trace elements such as Pb,23Hg,24Se,25-27and Cr.28 In this paper, we continue the development of isotope dilution GC/MS methods and report the results of studies performed to establish an assay for Te in urine using (4fluoropheny1)magnesium bromide (Grignard reagent) as the derivatizing agent. The precision and accuracy of the Te
-
(21) Kinkenbcrg,H.;Vandenual,S.;Frusch,J.;Terwint, L.; Beeren,T. At. Spectrosc.
1990, 11, 198-201. (22) Rigin, V. I. Fresenius 2.Anal. Chem. 1989, 335, 15-18. (23) Fcldman, B. J.; Mogadedi, H.; Osterloh, J. D. J . Chromatogr. 1992, 594,
275-282.
(24) Brunmark, P.; Skarping, G.; Schutz, A. J . Chromatogr. 1992, 573, 3541. (25) Ducros, V.; Favier, A.; Guigues, M. J . Trace Elem. Electrolytes Health Dis. 1991, 5, 145-154. (26) Maser-Veillon, P. B.; Mangles, A. R.; Patterson, K. Y.; Veillon, C. A ~ l y s t 1992, 117, 559-562. (27) Reamer, D. C.; Veillon, C. A. J. Nurr. 1983, 113, 786-792. (28) Veillon, C.; Wolf, W. R.; Guthrie, B. E. Anal. Chem. 1979, 51, 1022-1024.
Analytical Chemism, Vol. 80, No. 8, April 15, 1994
1317
Table 1. Furnace Condttlona for Daonnlnatlon of To by EAAS
time (8)
4
process
temp ("C)
ramp
hold
dry dry dry dry char cool atomize clean
80 120 200 600 900 (1150)" 20 2000 2500
1 1 10 10
5 5 15 5 20 8 5 5
5 1 0 1
For Pt modifier.
isotope ratio measurements were evaluated by replicate analyses of natural Te at nanogram levels on different days. The memory effect in determining the altered isotope ratios was studied by sequential analyses of samples with different isotope ratios. Finally, the isotope dilution GC/MS method for determining Te in urine samples was validated by the analysis of urine samples using a standard addition approach and comparing the results with those from EAAS.
EXPERIMENTAL SECTION Instrumentation. The mass spectrometer used was a doublefocusing,reverse-geometry instrument (Model 8230, Finnigan MAT, San Jose, CA) with a SpectroSystem 300 data system. The instrument was operated in the electron ionization (EI) mode using 70-eV electrons with a source temperature of 200 OC, the conversion dynode at -5000 V, and the secondary electron multiplier at 2400 V. The fragmentation pattern of the metal chelate was recorded by employing the exponential scanning mode (1 s/decade) and adjusting the source and the collector slits to obtain a triangular peak shape. Data for isotope ratio measurements were acquired in the selected ion monitoring mode (SIM) using voltage peak switching. The source and the collector slits were adjusted to obtain trapezoidal peaks with flat tops (collector slit width > source slit width) and the data were acquired at 2 Hz, yielding -20 data points across the 10-s-wideGC peak. A mass resolution of 1000 was used throughout. The isotope ratios were calculated by integrating the ion current for various chromatographic peaks. The overall linearity of the mass spectrometer was confirmed by injecting methyl stearate into the GC column and measuring the isotope ratios with m / z values 299:298,300:298,301:298, and 302:298, differing by 4 orders of magnitude (data not shown). Other details of the measurement methodology have been p ~ b l i s h e d . ~ A Varian 3700 gas chromatograph equipped with an Altech SE-30 (Altech Associates, Inc., Deerfield, IL) bonded-phase fused silica capillary column, 10 m X 0.32 mm, with a 0.25pm film thickness was used. Te chelate samples were injected by using an on-column injector (OCI-3, Scientific Glass Engineering, Austin, TX) at an oven temperature of 100 OC followed by a 15 OC/min temperature ramp to 300 "C. Highpurity helium was used as a carrier gas. A Perkin-Elmer 5 100atomic absorption spectrometer was used for analyzing urine samples. A pyrolytically coated graphite tube with stabilized temperature platform furnace (STPF) was used. The furnace was temperature programmed as shown in Table 1. Absorbance was measured at 214.3 nm 1310 Analytical Chemistry, Voi. 66, No. 8, April 15, 1994
using a Zeeman background correction procedure with an integration time of 5.0 s. There was a 300 mL/m argon gas flow throughout the analysis except that there was no internal gas flow during atomization. Reagents. 12@Te-enrichedmetal (56 atom 7%) used as an internal standard for isotope dilution was obtained from Oak Ridge National Laboratory (Oak Ridge, TN). Certified atomic absorption standard (Te in 2 wt % KOH), purchased from Aldrich Chemical Co. (Milwaukee, WI) was used as the primary standard. Double subboiling quartz distilled HNO3 in a Teflon bottle was obtained from the National Institute of Standards and Technology (NIST, Gaithersburg MD). Ultrex grade NH40H solution (30%) was obtained from J. T. Baker Chemical Co. (Phillipsburg, NJ), andstabilized H202 (50%) was obtained from Fisher Scientific (Fairlawn, NJ). Lithium bis(trifluoroethy1)dithiocarbamate [Li(FDEDTC)] was synthesized by reacting bis(trifluoroethy1)amine (PCR Inc., Gainesville, FL) and n-butyllithium (Aldrich) in an inert atmosphere at -70 OC followed by the addition of carbon disulfide (Aldrich) .29 (4-Fluoropheny1)magnesiumbromide (CFPMB; 2.0 M solution in diethyl ether) and anhydrous diethyl ether were also obtained from Aldrich. Though Te is a rare element and contamination is not a major problem, all the usual steps were taken to minimize the potential for adventitious Te contamination from the apparatus, reagents, personnel, and laboratory environment, as previously reported for other trace elements.819 Acid-leached polypropylene/Teflon labware, deionized water (DW), powder-free gloves, and a clean hood with laminar flow were used. Preparation and Standardization of Internal Standard Solution. A 120Teinternal standard solution was prepared by dissolving I 2 T e metal in 7 M HNOs and making up to volume with 1 M HNOs. More dilute solutions were prepared from this stock solution, on a weight basis, for isotope dilution experiments. The isotopic composition of Te in the internal standard was determined experimentally by preparing the Te(FCaH4)2 chelate. The internal standard solution was calibrated as previously reported by reverse-isotopedilution GC/ M S using the natural Te primary standard.15 The concentration of Te in the internal standard solution was calculated using the isotope ratios of m/z 310:312, 310:313, 310:314, 310:315, 310:316, 310:318, and 310:320. Digestion of Urine Samples and Chelate Formation. A known volume (1 mL) of the urine sample spiked with an accurately weighed amount of Te standard solution was mixed with a weighed amount of the 120Teinternal standard solution in a Teflon beaker. The Te internal standard added to the urine sample was selected to obtain an isotope ratio in the mixture corresponding to the geometric mean of isotope ratio m/z3 10:320of theexpected high and low concentration limits of the samples and the internal standard. The samples were treated with 1 mL of concentrated HNOs and digested with HN03 -+ H202 as described p r e v i o ~ s l y . ~The . ~ dried residue was redissolved in 1 mL of deionized water and the solution was transferred to a polypropylene (PP) centrifuge tube with a conical bottom. The solution was allowed to cool to room temperature and adjusted to pH 3 using a 4% solution of NH40H. Subsequently, 1 mL of pH 3 acetate buffer' was ( 2 9 ) Sucre, L.;Jennings, W. Anal. Lori. 1980, 13,497-501.
lool 80
,
84
1 '
190
I
1w
100,000 801
50.000 20,000 20,000 *
m/z-313
10,000.
m/z-312
2:w
230
I
,
330
-
,
.
I
4:OO
Retention Time (min)
0
50
,
3:00
100
150
200
250
300
350
m/z Fburo 1. E1 mass spectrum of Te(FCeH4)2. The molecular ion Is Indicated by M'+, and the Inset shows a blowup of the molecular ion
Figuro 2. Selected ion tracing showing the peaks of the molecular ion that correspond to the various Te isotopes. '0°1
to illustrate the information In the Isotopic cluster. 80
added followed by 200 pL of a 20 mM solution of Li(FDEDTC) chelating agent in deionized water. The samples were vortexed for 5 min, and the Te(FDEDTC)2 chelate formed was extracted with 1 mL of toluene. The organic extract containing the Te chelate was evaporated to dryness at 60 OC under a stream of argon gas in the laminar flow hood. The dried residue was redissolved in 200 pL of diethyl ether, and -200 pL of Grignard agent, 4-FPMB, was added. The solution in the PP tube was shaken to allow complete mixing. The excess of Grignard agent was destroyed by adding 100 pL of 10% isopropyl alcohol in toluene followed by 1 mL of 1 M HNOs. The Te(FC&)2 chelate was extracted with 1 mL of toluene, evaporated to dryness, and reconstituted in 10-50 pL of methylene chloride for analysis. Cas Chromatopaphy/Mass Spectrometry. Te isotope ratios were measured, in duplicate, by injecting 1 pL of the chelate solution and monitoring the isotopic cluster corresponding to the molecular ion. Prior to Te isotope ratio measurements, the focusing conditions of the mass spectrometer were optimized and mass calibration established by using perfluorokerosene (PFK). As the measurements were performed without using a lock mass, it was found necessary to experimentally determine the m/zvalue of the peak maximum by obtaining a histogram of the ion current adjacent to and including the calculated m/z value of a given ion. Data were acquired in a selected ion monitoring experiment, and quantitation was achieved by use of the chromatographic peak areas. The concentration of Te was calculated from different isotope ratios by the usual method of isotope di1uti0n.l~ Preparation of Samples for EAAS Analysis. A linear calibration curve in the concentration range of 0-100 pg of Te/L was established using diluted samples from the atomic absorption standard. The urine samples were diluted 5-fold and used for Te determination employing Mg and Pt as matrix modifiers in separate experiments. The concentration of Mg modifier was 1.4 pg/mL and that of Pt modifier was 10 pg/ mL. Sample (1 5 pL) and matrix modifier solution (1 5 pL) were pipeted together. We used aqueous Te standards and normal urine as the matrix spiked with Te standards.
RESULTS AND DISCUSSION The E1 mass spectrum of Te(FC&)2 shown in Figure 1
IU
1 1
284
1
282
I
iMi*' 284
m/z Figure 9. E1 mass spectrum of Te(CsHI) with Te isotopic information Illustrated In the blowup of the molecular ion.
exhibits an intense molecular ion peak, designated by Me+, containing Te isotopic information and was used in these experiments. Figure 2 is a selected ion monitoring trace showing the peaks at m/z values 312,313,314,3 15,3 16,3 18, and 320 correspondingrespectively to 122Te,luTe, 124Te,1z5Te, 126Te, 12*Te, and 130Te isotopes in the molecular ion Te(FCaH4)2*+in the mass spectrum. The intense peak at m / z 190 corresponds to the fragment ion (FCsH4)2+. An attempt was also made in the initial stages of the work to use phenylmagnesium bromide for derivatization. Figure 3 gives the complete mass spectrum of Te(C&)z, which shows the molecular ion Te(CaHs)z*+a t m / z 284 and the fragment ion at m / z 154 from (CaH5)2+. This was not pursued further due to the lower volatility of the unfluorinated derivative as well as the low intensity of the molecular ion in the mass spectrum. It may be noted that Te(Fc6H4)~can also be prepared via Te(DEDTC)Z using the commercially available sodium diethyldithiocarbamate or ammonium pyrrolidine carbodithioate. We selected Li(FDEDTC) as the first step to derivatize Te with an objective of carrying out the GC/MS analysis using Te(FDEDTC)2. This proved unsuccessful and therefore we reacted these preformed derivatives with the Grignard agent. Precision and Accuracy in Isotope Ratio Measurements. Table 2 shows the precision data of isotoperatio measurements of natural Te using Te(FCaH4)z. Chelated Te solution (1 pL) containing -10 ng of Te was injected. Five replicate determinations were made on each of three days. The mean values for each day were used to calculate the mean of means Anasl)rticel Chemistry, Vol. 86, No. 8, Aprll 15, 1994
1319
Table 2. Precision in Isotope Ratio Determination of Natural Te as Te(FCeH4)z
isotope ratio' (mean f rsd %) 310320
day 1 (n = 5) day 2 (n = 5) day 3 (n = 5) mean of means within-run precisionb(%) between-run precision (%) overall precision' (%)
d 0.0028 f 30 0.0033 f 11 0.0030 15.7 11.6 19.5
312:320 0.0774 f 3.4 0.0793 f 2.7 0.0781 f 1.6 0.0783 1.4 1.2 1.8
313:320 0.0390 f 3.4 0.0408 f 1.8 0.0406 f 3.6 0.0401 1.8 2.5 3.1
314320 0.1547 f 1.0 0.1568 f 2.0 0.1559 f 1.0 0.1558 0.8 0.7 1.1
315:320
316320
318:320
0.2315 f 2.1 0.2338 f 2.7 0.2346 f 1.3 0.2333 1.2 0.7 1.4
0.5970 f 2.1 0.6097 f 0.8 0.6045 f 1.4 0.6037 0.9 1.1 1.4
0.9534 f 1.3 0.9556 f 0.8 0.954 f 0.7 0.9543 0.6 0.1 0.6
a Not corrected for any mass discrimination factor. Calculated using the formula S i = ( E ~ ~ l s ~ ) 1 /where 2 / n , si representa the standard deviation obtained on each individual day and n is the number of days. Overall precision St was calculated by combining within-run precision (Si) and between-run precision (S.) according to the formula S t = (Si+ Se)1/2.The ratio 310320 was not measured for day 1.
Table 3. Calculated and Measured Abundance8 of Different Ions in Natural Te(FC8H4)z
isotope
atom % abund 0.096 2.60 0.91 4.82 7.14 18.95 31.69 33.80
molecular ion ( M + ) calcd measd (m/z) abunda (%) abundb (%) ion
310 312 313 314 315 316 318 320
0.094 2.53 1.23 4.83 7.60 19.45 31.04 33.20
0.098 2.55 1.31 5.08 7.60 19.67 31.10 32.59
Table 4. Calculated and Meamred Abundance8 of Different Ions In Enrlched 120Te(FC8H,)2
atom % abund' 55.98 f 0.30 8.25 f 0.20 1.46 f 0.10 3.80 f 0.10 3.99 f 0.10 7.94 f 0.20 9.43 f 0.20 9.14 f 0.20
molecular ion ( M + ) calcd measd (m/z) abundb( % 1 abundc (%) ion
310 312 313 314 315 316 318 320
54.29 8.46 2.51 3.95 4.38 8.26 9.21 8.94
53.80 8.64 2.61 4.16 4.45 8.41 9.21 8.73
'Including the contributions of the isotopes of Te and carbon in the ion Te(FCsH4)2'+. Not corrected for any mass discrimination factor.
Values given by Oak Ridge National Laboratory. Including the contributions of the isotopesof Te and carbonin the ionTe(FC&)2*+. Not corrected for any mass discrimination factor.
and its standard deviation, referred to as between-run precision in this table. The within-run precision was obtained by considering the standard deviation values obtained on the individual days. Overall precision was calculated by combining the within-run and between-run precision values. This was done to evaluate the effects of any variations in the mass spectrometer operating parameters that may affect the isotope ratio data from one day to another. No significant differences were observed among the isotope ratios measured on different days. Overall precision values of 0.6-3.1% are obtained for the different isotope ratios, except for the m / z 310:320 ratio which gave a precision of -20% due to the low natural abundance (0.096%) of lZ0Te. Obviously, the best precision of 0.6% was obtained for m / z 3 18:320 ratio with value close to unity. No correction has been applied to the data shown in this table for any mass discrimination among the different isotopes. Any mass discrimination factor would be canceled in an isotope dilution experiment in which the internal standard solution is calibrated in the same experiment. Tables 3 and 4 give the data for the isotopic composition of Te in the natural Te and lZ0Te-enrichedinternal standard, respectively. The atom percent abundances of different Te isotopes in natural Te are the recommended values based on the values measured experimentally by various l a b o r a t ~ r i e s . ~ ~ The atom percent abundances of different Te isotopes in 120Te internal standard are the values provided by Oak Ridge National Laboratory. Tables 3 and 4 also include the calculated and measured values for the abundances of various
Te-containing peaks in the molecular ion, Te(FC6H&*+, for the natural Te and 12Teinternal standard, respectively. The calculated abundances have been obtained by including the contributions of the isotopes of Te and C in the molecular ion.31,32As can be seen, there is good agreement between the calculated and measured abundances in the natural Te and lzoTe tracer. This shows that within the experimental uncertainties, there was no appreciable mass discrimination due to the use of different accelerating voltages since isotope ratios were determined by voltage peak-to-peak switching. This lack of bias is likely due to the small mass difference at the relatively high m / z values of the ions in the chelate. Evaluation of Memory Effect. Memory effect refers to carry-over in the GC/MS system during sequential analyses of samples with different isotope ratios. This carryover can adversely affect the accuracy of the measurement of altered isotope ratios. In these studies, the memory effect was evaluated by determining isotope ratios different from those of natural Te. For this purpose, two synthetic mixtures differing in the m / z 310:320 ratio by a factor of 50 were prepared by mixing weighed aliquots of the primary standard and the enriched lzoTe solution in differing proportions by containing almost equal amounts of total Te. Replicate determinations, each with 10 ng of Te, were made for each mixture. The analyses were carried out in the following sequence: five injections of mixture 1, five injections of mixture 2, followed by five injections of mixture 1. The mean value
(30) DeBievre, P.; Barnes, I. L. I n f . J . Mass Specfrom. Ion Processes 1985, 65, 211-230.
1320
Analytical Chemistty, Vol. 66,No. 8, April 15, 1994
(31) Beynon, J. H. MassSpectromefryandIfs Applications fo Organic Chemistry; Elsevier: Amsterdam, 1960. (32) Beynon, J. H.; William, A. E.Mass and Abundance Tablesfor Use in Mass Specfromerry;Elsevier: Amsterdam, 1963.
3'3
1
Mixture 1
Mixture 1
concentration of Tea (MR of Te/a of solution) (mean f SD)a rsd (%) range of values
ratio 310312 310313 310314 310315 310316 310318 310320
$ 2
E N -
0.08
Splke Solutlon by
Table 5. Determlnatlon of Te In Reverae-Iwtope Dllutlon QC/MS
0.06
10.54 f 0.27 10.67f 0.50 10.47f 0.13 10.26 f 0.17 10.11f 0.17 10.10f 0.11 10.05 f 0.07
2.6 4.7 1.2 1.7 1.7 1.0 0.7
10.15-10.80 9.98-11.14 10.37-10.65 10.01-10.39 9.89-10.28 9.96-10.20 9.99-10.15
Number of samples n = 4.
a
0.04
0.02
Table 6. Determlnatlon of Te In Urlne by Irotooe DIMlon QC
0.00 1
2
3
4
5
6
7
8
concn of Te (pg/L) using various ion ratios
9 1 0 1 1 1 2 1 3 1 4 1 5
Analysis Number
Flgure 4. Sequential study of two synthetic mlxtures differing In the mlr 310:320 ratio by a factor -50. The absence of observable memory effect is most clearly demonstrated by analyses 11-15 belng equal to analyses 1-5.
determined for the isotope ratio m / z 310:320 from analyses 1-5 from mixture 1 is 2.934 f 3.8%. Thecorresponding value for analyses 11-15 for the same mixture 1 is 2.925 f 3.4% after analyses 6-10 from the mixture 2 with a mean value of 0.0706 f 3.3%. In the presenceof a memory effect, the isotope ratios for analyses 11-1 5 for mixture 1 would have shown an increasing trend and their mean value would have been lower than the mean value for analyses 1-5. Figure 4 shows the m / z 3 10:320 isotope ratios determined in this experiment. As can be seen from the constancy in the isotope ratios determined by replicate injections, there is no significant memory effect observed in analyzing the two mixtures with widely different isotope ratios of 120Te:130Te.This facilitates the analysis of unknown specimens with varying amounts of Te, minimizes the need for optimum addition of internal standard to restrict the range of isotope ratios, and enhances the confidence in the isotope ratio data. Calibration of I*@Te Internal Standard Solution. The lzoTe internal standard solution was calibrated by reverse-isotope dilution using a primary standard of natural Te. For this standardization, four samples were prepared by mixing weighed aliquots of primary standard and lZoTeinternal standard solutions to achieve an optimum isotope ratio m / z 310:320 in the mixtures. Chelates were prepared and the different isotope ratios were measured. The Te concentration in the internal standard solution was calculated using seven different isotope ratios. The values shown in Table 5 were in good agreement with one another. The standard deviation observed for the concentration values from isotope ratios m / z 310:312 and 310:313 are higher due to the low natural abundances of lzzTe(2.60%) and lz3Te(0.91%), respectively, in the Te standard. Moreover, the concentration values determined from the ratios m / z 310:312, 310:313, 310:314, and 3 10:315 are higher compared to other values because the internal standard addition was not optimized for these ratios in the spiked mixtures. Also the areas for the ion peaks at m l z 3 12,3 13,3 14, and 3 15 were small due to lower abundances of the corresponding Te isotopes in the natural Te and the internal standard, and this gives rise to a larger error in the concentration data.
expected concn (pg/L)
310316
310318
310320
concn (pg/L)a
110.2 224.8 298.3 397.3 520.2
105.4 209.4 295.8 401.2 529.0
109.7 220.1 309.2 401.0 512.6
100.4 215.9 306.8 400.4 527.2
105.2 f 4.7 215.1 f 4.4 303.9 f 7.1 400.9f 0.4 522.9 f 9.0
Mean f SD.
a
Table 7. Detennlnatlon of Te In Urlne Samples by E M S measd concqa (pg/L f rad %)
no. 1
2 3 4 5
expected concn(&L) 110.2 224.8 298.3 , 397.3 520.2
Mg
Pt
a
U
a
u
33f7 75f2 97f15 150f 1 208f 1
121f2 258*4 421*4 536f5 717f5
34f43 56f26 75f3 95f 16 102f7
119f49 196f27 326f 11 352f9 527f 1
0 a and u denote the calibration curve obtained by using aqueous standards and urine-based Standards, respectively.
Results on Urine Samples. The calibrated I2OTe internal standard solution was used to quantify Te in urine to which known amounts of Te were previously added. The results obtained are shown in Table 6. The concentration values were calculated using the isotope ratios m / z 310:316, 310: 318, and 310:320 as discussed above. As can be seen, the Te concentration values calculated by use of the three different isotope ratios are in good agreement with one another as well as with the added values. The results obtained for the Te determination in urineusing EAAS are given in the Table 7. It is clearly seen that the use of a calibration curve based on aqueous standards does not lead to correct results. The values obtained for Te concentration in urine samples are lower by a factor of 2-3 compared to the added values. This is not surprising and is attributed to the commonly known matrix effects in EAAS. It was observed that the slopeof thecalibration curve with magnesium nitrate as the matrix modifier and using urine-based standards was much lower (0.0005) compared to the aqueous standards (0.0019), and this accounts for the lower Te concentration values observed in urine samples. A calibration curve based on standards containing urine yields much better results with either of the two matrix modifiers (Pt or Mg), and the results obtained are in good agreement with the expected values and those determined by isotope dulition GC/MS. Mg appears Analytical Chemistry, Vol. 66,No. 8, April 15, 1994
1321
to be a better matrix modifier compared to Pt reported previ~usly.~ This signifies the need to critically evaluate the different matrix modifiers for Te determination in biological samples using EAAS. The isotope dilution mass spectrometry method does not have any such matrix effects and therefore offers superior results in terms of precision and accuracy. The EAAS data obtained by using a recently published method3 are included to point out the drawbacks of the EAAS method compared to the GC/MS technique.
CONCLUSIONS Stable isotope dilution gas chromatography/mass spectrometry with 4-FPMB as a chelating agent can be used for Te determination in biological samples at micrograms per liter with a high degreeof precision and accuracy. The absence of memory effect enhances the confidence in measuring altered isotope ratios. The present method would supplement EAAS methods, providing an alternative physicochemical principle for the quantitation of Te. ACKNOWLEDGMENT Funding for the purchase of the high-resolution mass
1322 Analytical Chemistv, Vol. 88, No. 8, April 15, 1994
spectrometer was obtained from the National Institutes of Health, Division of Research Resources Shared Instrumentation Grant Program, Grant 1-S10-RRO-2418-01. Additional funding from the John Lee Pratt Fund of the University of Virginia is gratefully acknowledged. S.K.A. thanks the Division of Experimental Pathology, Department of Pathology, University of Virginia Health Sciences Center, for a postdoctoral fellowship and the authorities at Bhabha Atomic Research Center, Trombay, Bombay400 085, India, for granting leave. The authors thank Patrick K. Anonick for assistance in the synthesis of the chelating agent and J. Savory and M.R. Wills for their interest in the present work and for allowing use of the facilities available in the trace metals laboratory.
Recelved for review August 13, 1993. Accepted January 11, 1994.'
Abstract published in Aduancc ACS Abstracts, March 1 , 1994.
