Real-time internal standardization with an axially-viewed inductively coupled plasma for optical emission spectrometry

An evaluation of precision improvements using real-time internal standardization with an axially-viewed inductively coupled plasma (ICP) is presented. New findings are presented with respect to the nature of the noise in the analytical signals from the axial ICP. It is observed that a high degree of correlation exists in the line signals from the axial ICP. Using the yttrium ion line at 371.030 nm as the internal standard, the analytical precision after the application of real-time internal standardization is maintained between 0.1 and 0.2% relative standard deviation (RSD) for ion lines. Precision improvement factors of 3 to 4 are obtained by comparison with the uncorrected results. With atomic lines, real-time internal standardization using the yttrium ion line is less effective, yielding precision values between 0.2 and 0.7% RSD. The precision improvement factors for atomic lines are between 1.5 and 3. Thus, real-time internal standardization provides significant improvements in the RSDs of the line signals. The limits of these improvements are explored and an equation is presented which yields the fundamental shot noise limit for precision. Shot noise limited precision is demonstrated. However, this is not possible for all elements using a single internal standard signal. The effectiveness of real-time internal standardization is shown to be dependent on the nature of the specific spectral line. With the axially-viewed ICP, the dominant phenomenon preventing the full benefit of internal standardization from being obtained is the amplitude of the noise in the line signals and not the degree of correlation between analyte and internal standard signals. A trend is observed for atomic transitions in which lower excitation energy is correlated with higher relative noise amplitudes. This finding is in contrast with previously published work on the radially-viewed ICP. An explanation of this result is proposed which takes into account the influence of vaporizing sample droplets in the observation volume.