Original Articles
Correction of instrumentally produced mass fractionation during isotopic analysis of fe by thermal ionization mass spectrometry

https://doi.org/10.1016/S1387-3806(99)00158-XGet rights and content

Abstract

High-precision (∼0.015%/mass) isotope ratio measurements of Fe may be obtained by using magnetic-sector thermal ionization mass spectrometry (TIMS), where rigorous correction of instrumentally produced mass fractionation can be made. Such corrections are best done by using a double-spike approach, which was first introduced several decades ago. However, previous derivations do not lend themselves to the high-precision isotope analysis that modern TIMS instruments are capable of because of various assumptions of mass fractionation laws or constant atomic weights. Moreover, some of these previous approaches took iterative approaches to the calculation, and none presented detailed error propagations. Here we present a completely general derivation to the double-spike approach that may be used for any appropriate isotope system and is applicable to the mass fractionation laws that are known to occur in TIMS. In addition, we present an assessment of error propagation as a function of algorithm and spike isotope composition. This approach has produced the highest precision Fe isotope ratio measurements yet reported, on the order of ±0.2 to 0.3 per mil for the 54Fe/56Fe ratio, that correct for instrumentally produced mass fractionation and yet retain natural, mass-dependent isotopic variations in samples.

Introduction

There have been relatively few isotopic studies of intermediate-mass elements, in comparison to studies of mass-dependent fractionation of the light stable isotopes (e.g. [1]), or studies of heavier elements that are part of radioactive decay systems (e.g. [2]). However, intermediate-mass elements may undergo natural, mass-dependent isotopic fractionation, as has been discussed for Ca [3], [4], [5]. In the case of Fe, its importance in metabolism, as well as nucleosynthetic processes, has inspired a number of previous Fe isotope investigations (Table 1; [6], [7], [8], [9], [10], [11], [12]). Most of these studies used magnetic-sector TIMS in an effort to maximize the precision. Alternative methods, such as inductively coupled plasma mass spectrometry (ICP-MS) have advantages in higher ionization efficiencies as compared to TIMS, but suffer from ArN+ and ArO+ interferences on 54Fe and 56Fe, respectively, which presents significant difficulty in obtaining precise isotope ratios for all of the Fe isotopes.

A significant limitation of isotopic analysis using TIMS is the potentially large and variable mass fractionation that occurs in a TIMS source. A common approach to removing instrumentally produced mass fractionation is to adjust the measured ratios of interest to a single reference ratio. This approach, often referred to as “internal normalization,” is useful where one isotope may independently vary due to radiogenic in-growth from a radioactive parent isotope (e.g. [2]), or from nucleosynthetic processes (e.g. [2], [8]). Although internal normalization is capable of producing high-precision Fe isotope ratios, on the order of ±0.01% (or 0.1 per mil) by using TIMS [8], this approach completely removes any natural, mass-dependent isotopic fractionation that may exist in the sample.

Two approaches may be taken that correct for instrumentally produced mass fractionation and yet retain natural, mass-dependent isotopic variations in samples. One is an empirical approach, where the average instrumental mass fractionation is estimated by isotopic analysis of gravimetrically prepared isotopic standards. This approach was applied to Fe by Taylor et al. [13] and Dixon et al. [9], [10], who report the reproducibility of Fe standards as ±0.3 per mil (‰) per mass. However, the data reported for samples [9] are considerably more variable and contain internal inconsistencies among the Fe isotope ratios, suggesting that the true precision on samples of unknown isotopic composition and matrices is on the order of 1–3 per mil per mass using the empirical approach.

It has long been known that the most rigorous approach for correcting instrumentally produced mass fractionation and, at the same time, retaining natural isotope variations in samples, is the double-spike method (e.g. [13], [14], [15]). The advantage of the double-spike approach is that it is not affected by variations in sample matrix, sample loading, or source conditions, factors that can produce significant uncertainties in the empirical approach to correcting instrumental mass fractionation.

In this contribution, we present a completely general formulation of the double-spike approach that avoids previous approximations and assumptions that either prevented precise isotopic measurements, involved iterative calculations, or were derived for a specific isotope system.

Section snippets

Double-spike method

The double-spike technique is superior to the approaches that have been previously taken for Fe isotope analysis, because it allows rigorous instrumental mass fractionation corrections to be made, and simultaneously preserving naturally occurring, mass-dependent isotope variations (e.g. [13], [14], [15]). The double-spike approach can only be used for elements with four or more isotopes and in its most general application, requires two analyses; an analysis of an unspiked aliquot of the unknown

General derivation of the double-spike solution

The fractionation curve for the unspiked sample (unknown) (curve U in Fig. 1) is described by XUT=XUM(1+aXf1),YUT=YUM(1+aYf1),ZUT=ZUM(1+aZf1) where f1 is the mass fractionation factor for the sample (unknown), and aX, aY, and aZ are the mass-difference coefficients, as modified by the exponential approximation [Eq. (5)], and isotope ratios X, Y, and Z are defined previously.

The fractionation curve for the mixture (spike + sample) (curve M in Fig. 1) is described by XMT=XMM(1+aXf2),YMT=YMM(1+aYf2

Mass analysis of Fe

All isotope analyses were conducted at the University of Wisconsin Radiogenic Isotope Laboratory using a Mircomass Sector 54 TIMS mass spectrometer. This instrument is fitted with seven Faraday collectors and an analog Daly detector. Filament ribbon used for Fe isotope analysis is 99.999% pure (“zone-refined”) Re and is 0.0010 in. thick by 0.030 in. wide. After adding an Al2O3 slurry to a single blank filament, the Fe sample (∼4 μg of Fe) is loaded, followed by 1 M H3PO4, followed by silica

Precision of Fe isotope measurements

Our lab standard is a J-M Fe standard, and this can be measured to a precision of ±0.26 per mil (1 SD) for the 54Fe/56Fe ratio, using 21 analyses of 12 mixtures (Xspike = 0.13–0.38; Table 4). Using the average ratios (when available) of mixtures that have been run in duplicate produces a precision of ±0.12 per mil (1 SD) for the 54Fe/56Fe ratio. There is no correlation between the calculated true isotopic composition of the standard with that of the fraction of spike in the mixture (Fig. 5),

Conclusions

Instrumentally produced mass fractionation in a thermal ionization mass spectrometer may be rigorously corrected using the double-spike approach. The closed-form derivation we present is completely general, being applicable to any element with four or more isotopes, and provides an excellent approximation to exponential mass fractionation in the TIMS source, over the range of measured instrumentally produced mass fractionation. In addition, it is shown that the spike isotope composition has a

Acknowledgements

This research was supported by NSF grant no. OPP-9713968 and NASA grant no. NAG5-6342, and the NASA Astrobiology Institute. We thank Herb Wang and Zulcas Baumgartner for helpful discussions.

References (30)

  • W.A. Russell et al.

    Geochim. Cosmochim. Acta

    (1978)
  • J. Skulan et al.

    Geochim. Cosmochim. Acta

    (1997)
  • A. Götz et al.

    Int. J. Mass Spectrom. Ion Processes

    (1988)
  • I.D. Hutcheon et al.

    Geochim. Cosmochim. Acta

    (1987)
  • P.D.P. Taylor et al.

    Int. J. Mass Spectrom. Ion Processes

    (1992)
  • T. Walczyk

    Int. J. Mass Spectrom. Ion Processes

    (1997)
  • P.D.P. Taylor et al.

    Int. J. Mass Spectrom. Ion Processes

    (1993)
  • N.H. Gale

    Chem. Geol.

    (1970)
  • G.L. Cumming

    Chem. Geol.

    (1973)
  • M.J. Dallwitz

    Chem. Geol.

    (1970)
  • M.H. Dodson

    Geochim. Cosmochim. Acta

    (1970)
  • A. Hofmann

    Earth Planet. Sci. Lett.

    (1971)
  • B. Hamelin et al.

    Geochim. Cosmochim. Acta

    (1985)
  • G. Manhes et al.

    Earth Planet. Sci. Lett.

    (1978)
  • B.L. Beard et al.

    Geochim. Cosmochim. Acta

    (1999)
  • Cited by (151)

    • Radioisotope mass spectrometry

      2020, Handbook of Radioactivity Analysis: Volume 1: Radiation Physics and Detectors
    View all citing articles on Scopus
    View full text