Anal. Chem. 2004, 76, 322-327
Fe Isotope Variations in Natural Materials Measured Using High Mass Resolution Multiple Collector ICPMS G. L. Arnold,*,† S. Weyer,‡,§ and A. D. Anbar†,|
Department of Earth and Environmental Sciences, University of Rochester, Rochester, New York 14627, ThermoFinnigan GmbH, Bremen, Germany, and Department of Chemistry, University of Rochester, Rochester, New York 14627
We present the first measurements of Fe isotope variations in chemically purified natural samples using high mass resolution multiple-collector inductively coupled plasma source mass spectrometry (MC-ICPMS). High mass resolution allows polyatomic interferences at Fe masses to be resolved (especially, 40Ar14N+, 40Ar16O+, and 40Ar16OH+). Simultaneous detection of Fe isotope ion beams using multiple Faraday collectors facilitates highprecision isotope ratio measurements. Fe in basalt and paleosol samples was extracted and purified using a simple, single-stage anion chemistry procedure. A Cu “element spike” was used as an internal standard to correct for variations in mass bias. Using this procedure, we obtained data with an external precision of 0.030.11‰ and 0.04-0.15‰ for δ56/54Fe and δ57/54Fe, respectively (2σ). Use of Cu was necessary for such reproducibility, presumably because of subtle effects of residual sample matrix on mass bias. These findings demonstrate the utility of high-resolution MC-ICPMS for high-precision Fe isotope analysis in geologic and other natural materials. They also highlight the importance of internal monitoring of mass bias, particularly when using routine methods for Fe extraction and purification. The development of multiple-collector inductively coupled plasma mass spectrometry (MC-ICPMS) has stimulated interest in the stable isotope systematics of metals, particularly transition metals.1-12 Interest in Fe isotopes is especially intense because †
Department of Earth and Environmental Sciences, University of Rochester. ThermoFinnigan GmbH. | Department of Chemistry, University of Rochester. § Present address: Institut fu ¨ r Mineralogie, Universita¨t Frankfurt, 60054 Frankfurt, Germany. (1) Mare´chal, C. N.; Te´louk, P.; Albare`de, F. Chem. Geol. 1999, 156, 251273. (2) Anbar, A. D.; Roe, J. E.; Barling, J.; Nealson, K. H. Science 2000, 288, 126128. (3) Barling, J.; Arnold, G. L.; Anbar, A. D. Earth Planet. Sci. Lett. 2001, 193, 447-457. (4) Siebert, C.; Nagler, T. F.; Kramers, J. D. Geochem. Geophys. Geosyst. 2001, 2, article no. 2000GC000124. (5) Rouxel, O.; Dobbek, N.; Ludden, J.; Fouquet, Y. Chem. Geol., in press. (6) McManus, J.; Nagler, T. F.; Siebert, C.; Wheat, C. G.; Hammond D. E. Geochem. Geophys. Geosyst. 2002, 3, article no. 1078. (7) Rehka¨mper, M.; Mezger, K. J. Anal. At. Spectrom. 2000, 15, 1451-1460. ‡
322 Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
of the ubiquity of this element, its importance in biology and the environment, and the possible use of Fe-bearing biominerals as “biosignatures”.11 However, there are two critical issues introduced by MC-ICPMS that must be addressed for accurate and precise Fe isotope analysis: (1) polyatomic interferences; (2) mass bias control. First, Fe isotope analysis by MC-ICPMS is a special challenge because 40Ar14N+, 40Ar16O+, and 40Ar16OH+ are produced in plasma sources. These polyatomic ions have the same nominal mass (hereafter referred to as isobaric) as 54Fe, 56Fe, and 57Fe, respectively. In a standard ICP source (wet plasma), the magnitudes of these interferences can be 5-50% of the Fe ion beams. Hence, they can be very problematic when measuring natural variations in the isotopic composition of Fe that are commonly 0.1-1‰. Second, the magnitude of mass discrimination during ICPMS analysis (hereafter referred to as mass bias) for Fe is typically of order 3-4%/amu. This is much larger than the magnitude of natural mass-dependent variation observed for Fe. Therefore, it is critical to correct or compensate for this effect to very high precision. Three approaches have previously been used to address the interference problem in Fe isotope MC-ICPMS analysis: desolvating nebulizers;2,13 collision cells;10 “cold plasma” operating conditions.14 When coupled to multiple collector instruments, each of these methods has proven capable of reducing interferences sufficiently to permit determination of mass-dependent variations in the isotopic composition of Fe extracted from natural materials. However, desolvation nebulizers reduce polyatomic interferences but do not fully eliminate them. This necessitates large sample sizes (e.g., 20 ppm Fe standard solutions).13 In cold plasma there (8) Walczyk T.; Von Blanckenburg, F. Science 2002, 295, 2065-2066. (9) Zhu, X. K.; Guo, Y.; Williams, R. J. P.; O’Nions, R. K.; Matthews, A.; Belshaw, N. S.; Canters, G. W.; de Waal, E. C.; Weser, U.; Burgess, B. K.; Salvato B. Earth Planet. Sci. Lett. 2002, 200, 47-62. (10) Beard, B. L.; Johnson, C. M.; Skulan, J. L.; Nealson, K. H.; Cox, L.; Sun, H. Chem. Geol. 2003, 195, 87-117. (11) Beard, B. L.; Johnson, C. M.; Cox, L.; Sun, H.; Nealson, K. H.; Aguilar, C. Science 1999, 285, 1889-1892. (12) Rouxel, O.; Ludden, J.; Fouquet, Y. Chem. Geol. 2003, 200, 25-40. (13) Belshaw, N. S.; Zhu, X. K.; Guo, Y.; O’Nions, R. K. Int. J. Mass Spectrom. 2000, 197, 191-195. (14) Kehm, K.; Hauri, E. H.; Alexander, C. M. O’D.; Carlson R. W. Geochim. Cosmochim. Acta 2003, 67, 2879-2891. 10.1021/ac034601v CCC: $27.50
© 2004 American Chemical Society Published on Web 12/09/2003
is little to no attenuation of hydroxides, so ArOH+ persists as an interference problem at mass 57.14 In addition, cold plasma running conditions result in a loss of Fe ionization efficiency, also requiring large sample sizes.14 In collision cells ArOH+ is reduced, but not completely eliminated.10 Careful measurement of mass 57 is important because it allows comparison of variations in two ratios, 56Fe/54Fe and 57Fe/54Fe. If variations are the result of massdependent chemical processes, variations in 56Fe/54Fe and 57Fe/ 54Fe should be correlated with a slope of ≈2/3. Therefore, precise measurement of 57Fe/54Fe is critical for quality control in Fe stable isotope research.2,10 As a result, Fe isotope analysis is among the most demanding and least routine of MC-ICPMS applications, despite widespread interest. Recently, Weyer and Schwieters15 reported Fe isotope results from a new MC-ICPMS instrument (ThermoFinnigan Neptune) capable of multiple collector analysis at high mass resolution with a resolving power of Rpower ) (M/∆M g 10 000).15 High-resolution multicollection promises to be a comparatively straightforward solution to the interference problem for Fe isotopic analysis. The initial study of Weyer and Schwieters15 was confined to solutions of high-purity reference materials and experimental products in which Fe was essentially the only dissolved cation other than H+. In practice, high-precision isotopic analyses of Fe in natural samples, in which Fe is rarely the predominant component, are more challenging. Matrix components may affect mass bias in ways that are difficult to recognize and poorly understood.16 Therefore, in stable isotope studies, matrix-induced variations in mass bias may degrade precision and give rise to systematic inaccuracies in the apparent deviation of isotopic composition between matrix-contaminated samples and clean standards. Such problems are dramatically ameliorated by purification of Fe, typically using chromatographic methods. However, the residual matrix even in nominally purified samples may be problematic when examining isotopic variations 95% yield for this procedure on aliquots of each sample drawn before and after Fe separation. Fe concentrations were determined using either UVvis spectrophotometry20 or high-precision isotope dilution ICPMS. Data are reported using the δ notation common in geochemical research, relative to the Fe standard IRMM-014:21
δ56/54Fesample )
(
(56Fe/54Fe)sample
(56Fe/54Fe)IRMM-014
)
- 1 × 1000
(1)
Measurement of the IRMM-014 standard and other standards of known isotopic composition were interleaved with measurements of the natural samples. Other standards included a “gravimetric” Fe standard. The gravimetric Fe standard was prepared by the addition of a known amount of 54Fe tracer to a JMC-Fe standard. The 54Fe tracer was made from an 54Fe-enriched metal (97.7% purity) purchased from Oak Ridge National Laboratory. The metal was dissolved in concentrated HNO3, and its concentration determined by reverse isotope dilution. All quantities were determined by careful gravimetry. The true δ56/54Fe of the gravimetric Fe standard is -10.4 ( 0.2 (2σ) relative to the IRMM014 standard. Also included was an Fe solution made from a synthetically produced akagenite (sample AKA-4). The akagenite was produced by the slow precipitation of FeOOH from Merck FeIIICl3‚6H2O solution and has a reference δ56/54Fe value of -0.61 ( 0.08‰ (2σ).15,22 Samples and standards were diluted to 3 ppm total Fe. A Cu standard (JMC-ICP solution Specpure Lot No. 200536E) was added to each sample before analysis to a concentration of 3 ppm. High-Resolution Multicollection Analysis. Samples were analyzed on a ThermoFinnigan Neptune (factory demonstration instrument in Bremen, Germany). The instrumental procedures are presented in detail in Weyer & Schwieters15 and summarized only briefly here. The Neptune is a double-focusing multiple collector ICPMS which has the capability of high mass resolution measurements in multiple collector mode. This is achieved by three switchable entrance slits located before the ESA (electrostatic analyzer) and fixed resolution detector slits, resulting in one low and two high mass resolution modes (which are called “medium” and “high” mass resolution). (19) van der Walt, T. N.; Strelow, F. W. E.; Verheij, R. Solvent Extr. Ion Exch. 1985, 3, 723-740. (20) To, T. B.; Nordstrom, D. K.; Cunningham, K. M.; Ball, J. W.; McCleskey, R. B. Environ. Sci. Technol. 1999, 33, 807-813. (21) Taylor, P. D. P.; Maeck, R.; De Bie`vre, P. Int. J. Mass Spectom. Ion Process. 1992, 121, 111-125. (22) von Blanckenburg, F.; Bo ¨ttcher, M. E. Eur. J. Mineral. 2001, 13, 193.
Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
323
Table 1. Likely Polyatomic Interferences on Fe Isotopes and the Required Mass Resolutiona
mass 54Fe 56Fe 57Fe a
44Ca12C
40Ar14N
40Ca14N
40Ar16O
40Ca16O
40Ar16OH
40Ca16OH
28Si28Si
55.955480
53.965458 2087
53.965665 2070
55.957299
55.957506
56.965124
56.965331
55.953854
2502
2479 1915
1902
2723
Fe atomic masses:
54Fe
) 53.939612;
56Fe
) 55.934993;
57Fe
)
RESULTS AND DISCUSSION As discussed above, there are two main issues that must be addressed to obtain both precise and accurate δ56/54Fe analysis by MC-ICPMS: correction for isobaric interferences; correction for mass bias. Our evaluation of both issues begins with a consideration of the mass bias behavior of the instrument. Instrumental Mass Bias. Instrumental mass bias has been described by a number of different empirical functions or “laws”. The two most common of these laws applied to high-precision Analytical Chemistry, Vol. 76, No. 2, January 15, 2004
i
56FeH+
56.953422
56.942764
3159
7671
2957
56.935396.35,36
All measurements in this study were performed with the Neptune medium mass resolution entrance slit. Each detector was positioned individually to discriminate the polyatomic interferences (mainly 40Ar16O+, 40Ar16OH+, and 40Ar14N+) at the edge of the detector slits so that only the undisturbed Fe beams were detected. With the Neptune medium mass resolution entrance slit a resolving power of 8000-9000 is achieved. This is sufficient to fully resolve polyatomic interferences, which differ from the respective Fe isotopes by less than 4000 M/∆M (Table 1) and to produce flat top peak sections enabling high-precision measurements. This includes all likely interferences except 56FeH+, which is not fully resolved from 57Fe+ in medium mass resolution mode on the Neptune. The measurements were performed using 1011 Ω resistors for all Fe masses. Unlike other methods,13 a 1010 Ω resistor is not needed because the amplifiers of the Neptune have an extended dynamic range of 50 V. The concentration matched samples (3 ppm Fe + 3 ppm Cu) were introduced in wet plasma conditions together with a low-flow (50 µL/min) PFA nebulizer and standard cones. A typical signal achieved for 3 ppm Fe solution in medium mass resolution was 10-15 V of 56Fe. The abundances of the polyatomic interferences were on the order of 10 mV for 40Ar14N+, 500 mV for 40Ar16O+, and 5 mV for 40Ar16OH+, resulting in relative abundances (compared to 54Fe+, 56Fe+, and 57Fe+, respectively) of 1-5%. The tail contribution of the polyatomic interferences on the respective Fe isotopes is negligible (
