Anal. Chem. 2006, 78, 6126-6132
δ34S Measurements of Sulfur by Multicollector Inductively Coupled Plasma Mass Spectrometry Robert Clough,† Peter Evans,‡ Tim Catterick,‡ and E. Hywel Evans*,†
Speciation and Environmental Analysis Research Group, School of Earth Ocean and Environmental Sciences, University of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK, and LGC, Queens Rd, Teddington, Middlesex, TW11 0LY, UK
An accurate and precise method for the determination of δ34S measurements by multicollector inductively coupled plasma mass spectrometry has been developed. Full uncertainty budgets, taking into consideration all the uncertainties of the measurement process, have been calculated. The technique was evaluated by comparing measured values with a range of isotopically enriched sulfur solutions prepared by gravimetric addition of a 34S spike. The gravimetric and measured results exhibited a correlation of R2 >0.999. Repeat measurements were also made after adding Na (up to 420 µg g-1) and Ca (up to 400 µg g-1) salts to the sulfur standard. No significant deviations in the δ34S values were observed. The Russell correction expression (Ingle, C.; Sharp, B.; Horstwood, M.; Parrish, R.; Lewis, D. J. J. Anal. At. Spectrom. 2003, 18, 219) was used to correct for mass bias on the 34S/32S isotope amount ratio from the mass bias observed for the 30Si/28Si isotope amount ratio. Consistent compensation for instrumental mass bias was achieved. Resolution of the measured δ34S values was better than 1‰ after consideration of all uncertainty components. The technique was evaluated for practical applications by measurement of δ34S for a range of mineral waters by pneumatic nebulization sample introduction and the analysis of genuine and counterfeit pharmaceuticals using both laser ablation sample introduction and liquid chromatography. For the former two cases polyatomic interferences were resolved by operating the MC-ICPMS in medium resolution, while for the chromatographic analyses polyatomic interferences were minimized by the use of a membrane desolvator, allowing the instrument to be operated at a resolution of 400. Relative differences in sulfur isotopic abundances, observed in sulfur-containing compounds from different sources,2 mainly result from the reduction of sulfate ions by anaerobic bacteria, which excrete H2S enriched in the lighter S isotopes. This variability has been exploited in a number of applications including the identification of water sources,3,4 monitoring of the effective* Corresponding author. E-mail:
[email protected]. Fax: +44 (0)1752 233040. † University of Plymouth. ‡ LGC. (1) Ingle, C.; Sharp, B.; Horstwood, M.; Parrish, R.; Lewis, D. J. J. Anal. At. Spectrom. 2003, 18, 219. (2) Rosman, K.; Taylor, P. D. P. Pure Appl. Chem 1998, 70 (1), 217-235.
6126 Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
ness of constructed wetlands for controlling acid mine drainage,5 determination of sulfate retention patterns and mechanisms in aerated forest soils,6 modeling of volcanic gas fluxes,7 identification of organically fed beef cattle,8 studies of the earth’s early atmosphere and environments,9 and inferring photochemical reactions in nebular gas.10 Gas source isotope ratio mass spectrometry (GS-IRMS), in which S is measured as SO2 or SF6, is the customary technique used for the measurement of relative sulfur isotopic abundance variations. However, the complex, multistep, sample preparation methods employed to convert the various S-containing species from geological, biological, and environmental matrixes into SO2 or SF6 may lead to additional S isotopic fractionation or contamination from the reagents employed. Furthermore, different δ34S values for the same material are obtained depending on which species, either SO2 or SF6, is introduced into the GS-IRMS.11 Oxygen isotopic variation12 must also be accounted for when SO2 is the analyte. In addition, the interaction of SO2 with the mass spectrometer has historically limited the measurement of δ34S to dedicated instruments, because residual SO2 can affect the measurement of other key isotopic systems. These difficulties have limited the use of δ34S values as a comprehensive survey tool, though continuous-flow methods are now changing this situation. In recent years, sector field and collision/reaction cell inductively coupled plasma mass spectrometry (ICPMS) has been used for the measurement of sulfur isotope amount ratios.13-16 However, due to the sequential data acquisition of these instruments, i.e., (3) Allen, D. Groundwater 2004, 17-31. (4) McArdle, N. C.; Liss, P. S. Atmos. Environ. 1995, 29, 2553-2556. (5) Hsu, S. C.; Maynard, J. B. Environ. Eng. Policy 1999, 1, 223-233. (6) Mayer, B.; Prietzel, C.; Krouse, H. R. Appl. Geochem. 2001, 16, 10031019. (7) Sakai, H.; Cassadevall, T.; Moore, J. Geochim. Cosmochim. Acta 1982, 46, 729-738. (8) Boner, M.; Forstel, H. Anal. Bioanal. Chem. 2004, 378, 301-310. (9) Farquhar, J.; Wing, B. Trans. Inst.f Min. Metall., Sect. B 112 B156-B157. (10) Rai, K.; Jackson, T.; Thiemens, M. Science 2005, 309, 1062-1065. (11) Qi, H. P.; Coplen, T. B. Chem. Geol. 2003, 199, 183-187. (12) Coplen, T.; Bohlke, J.; De Bievre, P.; Ding, T.; Holden, N.; Hopple, J.; Krouse, H.; Lamberty, A.; Peiser, H.; Revesz, K.; Rieder, S.; Rosman, K.; Roth, E.; Taylor, P.; Vocke, R.; Xiao, Y. Pure Appl. Chem. 2002, 74, 1987-2017. (13) Mason, P. R. D.; de Hoog, C. J.; Meffan-Main, S. Ninth Annual V. M. Goldschmidt Conference, 1999, LPI Contrib. 971. (14) Evans, P.; Wolff-Briche, C.; Fairman, B. J. Anal. At. Spectrom. 2001, 16, 964-969. (15) Mason, P.; Kaspers, K.; Bergen, M. J. Anal. At. Spectrom. 1999, 14, 10671074. (16) Prohaska, T.; Latkoczy, C.; Stingeder, G. J. Anal. At. Spectrom. 1999, 14, 1501-1504. 10.1021/ac060875h CCC: $33.50
© 2006 American Chemical Society Published on Web 07/13/2006
each isotope of interest is monitored in turn so ion beam fluctuations caused by plasma flicker are not accounted for, the precision obtained in these studies has limited the efficacy of the technique to identifying only gross variations in sulfur isotopic composition. The arrival of multicollector (MC) ICPMS instruments, which can be used to make simultaneous ion signal measurements, has introduced new opportunities for the precise measurement of isotope amount ratios with instrumental precision between 0.1 and 99% for uwithin, which was calculated by combining the standard uncertainty of each replicate according to error propagation laws. This highlights the importance of calculating a full uncertainty budget because, if the measurement precision is reported as simply the standard deviation of replicate measurements, the actual measurement uncertainty will be underestimated. Analysis of Pharmaceuticals. In a further development of the analytical technique, δ34S values were obtained for solid samples by direct introduction to the plasma by laser ablation. It has been reported that laser ablation can cause isotopic fractionation, in addition to that caused by the ICPMS instrument, evidenced by a rise in the measured 65Cu/63Cu isotope amount ratio with time.31 Figure 6 shows that this effect was not observed in this study as the measured 32S/34S isotope amount ratio varies randomly with time so we believe that the data are internally consistent. δ34S measurements were obtained for genuine Viagra tablets and counterfeit samples from two different sources. Table 3 shows the results obtained for these analyses. Significant differences were (31) Jackson, S.; Guˆnther, D. J. Anal. At. Spectrom. 2003, 18, 205.
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
6131
Figure 6. Variability of the 32S/34S isotope amount ratio with time during LA-MC-ICPMS of Pfizer Viagra. Table 3. δ34S Values for Genuine and Counterfeit Viagra δ34S relative to Pfizer Viagra
counterfeit Viagra 1 counterfeit Viagra 2
bulk tablet by laser ablation (‰)
sildenafil citrate by HPLC (‰)
-8.0 ( 0.36 +10.5 ( 0.39
-3.0 ( 0.9 +5.0 ( 1.1
measurable for counterfeit Viagra 1 and counterfeit Viagra 2 relative to genuine Pfizer Viagra, δ34S of -10.5 and +8.0‰, respectively. Subsequently, the active ingredient of the tablets, sildenafil citrate, was extracted and δ34S measurements made by HPLC-MC-ICPMS. In this case, mass discrimination was performed by the bracketing technique using the previously characterized SRM 3154 as the standard. The Axiom instrument employed for this work was operated in low resolution. Polyatomic interferences, which had been minimized by the use of a membrane desolvator, were accounted for by baseline subtraction of the raw ion counts for the chromatographic peak obtained. Isotope amount ratios were calculated using the pseudo-steady-state approach, which allows an estimate of internal precision to be made from a single sample injection.32 The δ34S values obtained for sildenafil citrate are presented in Table 3. Again significant differences were detectable, -3‰ for counterfeit 1 and +5‰ for counterfeit 2 relative to (32) Clough, R.; Truscott, J.; Belt, S.; Evans, E. H.; Fairman, B.; Catterick, T. Appl. Spectrosc. Rev. 2003, 38, 101.
6132 Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
genuine Pfizer Viagra. The different values obtained for the whole tablet and active ingredients can be accounted for by sulfur impurities in the cellulose of the tablet formulation. δ13C and δ15N values were also obtained for the genuine and counterfeit samples by GS-IRMS. No difference was observed for the δ13C measurements. However, the δ15N values also showed significant differences between the samples, -4‰ for genuine Viagra, -10‰ for counterfeit 1 and - 8‰ for counterfeit 2. The counterfeit drug market is estimated to be worth in excess of $30 billion per annum globally, with an estimated 10% of drugs sold in the UK contributing to this figure. This has important implications to both pharmaceutical companies and human health, as counterfeit drugs often contain sub- or superpotent ingredients. Current anticounterfeit measures include impurity profiling and packaging inspection. This work has shown that δ34S measurements of sulfur-containing drugs, separately or in conjunction with δ15N measurements, can act as valuable tools in detecting counterfeit pharmaceutical products. CONCLUSIONS A new and robust technique for measuring accurate and precise δ34S values has been developed, for both liquid and solid samples and individual chemical species, requiring minimal sample preparation. Silicon has been shown to be an effective method for mass discrimination correction when δ34S values are required. The use of silicon also reduces the analytical time, when compared to the sample/standard bracketing correction technique, and can compensate for matrix-induced changes to the plasma brought about by the presence of other ions in the sample. The application of a technique that does not require sample bracketing may have particular relevance where the use of bracketing standards certified for δ34S is not feasible due to difficulties in matrix matching. ACKNOWLEDGMENT The work described in this paper was supported by the Department of Trade and Industry (UK) as part of the National Measurement System Valid Analytical Measurement Program. Thanks are due to Ken Carter at LGC for the provision of δ13C and δ15N values for sildenafil citrate. Received for review May 12, 2006. Accepted June 5, 2006. AC060875H