Isotope mass bias

U) that can be used for isotope dilution mass spectrometry (IDMS), dual standards (233U/236U) for correction of instrumental isotopic mass bias, uranium-plutonium double spike mixtures and U/ Pu), and a series standards with certified content ranging from depleted uranium to LEU at 4.52%. In addition, there... 

Thermal ionization mass spectrometry (TIMS) suffers from time-dependent mass bias, referred to as mass fractionation, as a result of the finite amount of sample on the source filament and the more efficient thermal ionization of the lighter isotope. Mass bias correction is more crucial with multi-collector (MC)-ICP-MS as the latter suffers significantly larger bias and, as noted earher, it may not necessarily be constant over extended periods of time. Therefore, rigorous correction methods are required. Over the last few decades, several different mass bias correction methods have been successfully used for the determination of isotope amount ratios, as illustrated by Albarede et al. [16]. 

In principle, the three isotope method may be widely applied to new isotope systems such as Mg, Ca, Cr, Fe, Zn, Se, and Mo. Unlike isotopic analysis of purified oxygen, however, isotopic analysis of metals that have been separated from complex matrices commonly involves measurement of several isotopic ratios to monitor potential isobars, evaluate the internal consistency of the data through comparison with mass-dependent fractionation relations (e.g., Eqn. 8 above), or use in double-spike corrections for instrumental mass bias (Chapter 4 Albarede and Beard 2004). For experimental data that reflect partial isotopic exchange, their isotopic compositions will not lie along a mass-dependent fractionation line, but will instead lie along a line at high angle to a mass-dependent relation (Fig. 10), which will limit the use of multiple isotopic ratios for isobar corrections, data quality checks, and double-spike corrections. 

Figure 1. Transmission of Zn isotopes as a function of the collector position along the focal plane. This position is parameterized hy mass M. Transmission is the ratio between the number of ions of mass At, arriving at the collector at mass M divided by the number of atoms of mass At, introduced in the mass spectrometer. The parameter u is the linear mass bias coefficient.

Because there is so little mass bias in the mass analyzer, a discussion of ion transfer optics and collectors is not presented. The ion transfer optics of the magnetic sector mass analyzer, and the collectors used for isotope ratio measurements are critical design elements in all isotope ratio mass spectrometers and recent reviews of these items can be found in Habfast... 

Rigorous correction for instrumental mass bias is required if the precision of an isotope ratio measurement needs to be greater than l%o per mass unit. This concept is well illustrated by the definitive Ca isotope work of Russell et al. (1978), which used a double-spike approach. Prior to the Ca isotope investigation of Russell et al. (1978), natural mass-dependent Ca... 

To the first order, the mass bias on the isotopic ratio NJN can be evaluated from ... 

From a series of Zn isotope measurements, Marechal et al. (1999) deduced in this way that the patterns of the Zn mass bias on the Lyon Plasma 54 are best accounted for by the exponential law. 

A deviation from the ideal fractionation law will appear as a residual correlation between the ratios corrected for mass bias using the exponential law. It can be verified that even very small 8M/M in Equation (40) produces a potentially important isotopic effect on the order of 1 + P(5M,. - 5Mt)/Mj. The alignment of the correlation between the corrected ratios x = hir and y = In rj produced by sloping peaks in a log-log plot has a slope of 5M,/8M,. For a rormd peak, the second-order term should be included. 

So far, second-order corrections have only found their application for radiogenic isotopes (see a more extensive treatment in Albarede et al. 2004). The linear changes in the apparent mass bias of Nd with mass observed by Vance and Thirlwall (2002) is certainly an indication that high precision may benefit from such an elaborate scheme on at least some instruments. 

Matrix effects are typically divided into spectral (isobaric) and non-spectral types. The spectral or isobaric effects include 1) elemental isobaric interferences such as Cr at " Fe, 2) molecular interferences such as Ca O at Fe and Ar N at Fe, 3) double charge interferences such as Ca at Mg. Non-spectral matrix effects are largely associated with changes in the sensitivity of an analyte due to the presence of other elements (Olivares and Houk 1986). Changes in sensitivity correspond to a change in instrumental mass bias, and therefore non-spectral matrix effects can have a significant impact on the accuracy of isotope measurements. 

Figure 12. Plot of the Fe/ Fe ratio of an Fe standard, analyzed at 200 to 600 ppb concentrations, relative to the average Fe/ Fe of bracketing 400 ppb an Fe standard, versus the measured Fe volts (10 fl resistor). The measured Fe isotope composition varies relative to Fe concentration, which reflects differences in instrumental mass bias as a function of concentration. Data were taken over a 24 hour period using the University of Wisconsin-Madison Micromass IsoProbe.

Although mass bias effects relative to total ion intensity may be corrected using a working curve (Fig. 12), anomalous mass bias in natural samples due to the presence of other elements cannot be corrected. Figure 13 shows the effects of anomalous mass bias in a 400 ppb Fe ultra pure standard that has been doped with varying concentrations (up to 75 ppb) of Mg, Al, or La. Carlson et al. (2001) noted a similar effect of Al on the isotope composition of Mg standard solutions. These matrix elements were chosen because they would not produce isobars on the Fe mass spectrum (Fig. 13D) and for their spread in atomic mass Mg and Al both have masses less than Fe and La is greater. Additionally Mg and Al are major elements (7 and 3" most... 

Because of its isotopic variability, background must be reduced at any cost. The matrix of samples and standards must be reduced by appropriate chemistry to trace amounts, typically to a total concentration far smaller than the element to be analyzed. This requirement is most critical when the mass bias is inferred not internally from the sample itself, but externally from bracketing standards or from a different element used for isotopic normalization. Even the most dilute heavy species may drastically affect mass discrimination. 

Tf transmission for isotope i in cup y m mass bias factor for the linear law... 

Samples were prepared for Cu isotope analysis on the Multicollector Inductively-Coupled Plasma Mass Spectrometer (MC-ICPMS) at University of Arizona. The Cu-rich samples were loaded and dissolved in pure HNO3 and the Cu-poor samples were loaded and dissolved in a mixture of HCI and HNO3, Chromatographic separation of the Fe and Cu ions was deemed necessary for the Cu-rich samples. The diluted solutions were injected into the MC-ICPMS using a microconcentric nebulizer. Samples were run numerous times to increase precision. The Cu isotope ratios are reported in conventional per mil notation, relative to the NIST 976 standard. Mass bias was also accounted for by bracketing methods with the NIST 976 standard. 

Using the optimized data acquisition conditions, aspirate the mass bias solution, then the sample, then the mass bias solution again. Perform 10 replicate integrations on the i° Cd and Cd isotopes for each sample, and print out the results. 

Calculate the true io Cd i Ud isotope ratio, R, for the mass bias solution, using Eqn. B.2. 

---