Drift characteristics of the PE6000 ICPMS for solution analysis of geological samples

M. Norman, GEMOC, Macquarie

One of the principal limitations on the accuracy of an ICPMS determination is the drift of the instrument response relative to the calibration standard during the course of a run. This drift is often a complex function of mass and can be the single most important factor in determining the accuracy of an analysis (Eggins et al., 1997, Chemical Geology 134, 311-326). This is especially a problem for geological samples with their inherently complex matrix, and usually requires addition of multiple internal standards to the analysed solution to provide an adequate correction.

This abstract describes the drift characteristics observed for two sets of whole rock analyses conducted recently (April 22-24, 1997) on the Perkin Elmer 6000 ICPMS at Macquarie University. Samples consisted of 100 mg of basalt, pelagic sediments, and ultramafic xenoliths dissolved in concentrated HF+HNO3. A mixed internal standard consisting of 6Li, As, Rh, In, Re, and Bi was added, and the solution brought to a final volume of 100 ml with 2% HNO3. The nominal dilution factor of the final solutions were 1:1000, and nominal concentrations of the internal standards were 30 ppb Li and 10 ppb for the other elements. Both runs were calibrated against the well characterized basalt standard BHVO-1, and another basalt from Kilauea volcano (KIL1489) was run peridocially as a drift monitor.

Figures 1 and 2 show the intensities of each analysis of KIL1489 normalised to the first analysis of this sample in each run. These data show that a significant loss of intensity for the light masses occurred early in each run, which appears to be a fairly typical pattern for solution ICPMS analysis regardless of the type of instrument. In these experiment, the heavier masses show less overall change in signal intensity, although the drift is clearly a non-linear function of both mass and time

Drift corrections that simply normalize a group of elements to a single internal standard are clearly inappropriate if better than 10-20% accuracy is required, as it is in most modern geochemical research problems. Fortunately, these data suggest that the response characteristics for the PE 6000 are relatively smooth over limited regions of the mass range, so that corrections based on interpolations between the internal standard masses should improve both the accuracy and the precision based on replicate determinations of a single sample, although additional corrections based on a drift monitor analysed every few samples may be necessary to achieve the highest levels of accuracy and precision.

Understanding, controlling, and correcting for instrument drift such as this should be a high priority of all manufacturers of ICPMS instrumentation.

Back to the GEMOC Abstract Titles Page