The starting-point of this article is the common problem encountered in both inductively coupled plasma atomic emission spectrometry (ICP—AES) and d.c. arc AES when used for trace analysis, namely the extraction of a true net analyte signal from the composite radiation observed in the spectral window of an analysis line. In this connection the author points, first, to the wealth of data available in the literature on d.c. arc spectrography, which again is of topical interest in the further development of ICP—AES, and second, to the lack of data on the relative intensities of spectral lines emitted for ICPs, which makes it difficult to choose reliable analysis lines when the matrix elements yield complex spectra. The main topic of the paper is the description of an approach currently under development at Philips Research Laboratories, the eventual goal of which is to arrive at a new tabulation of spectral interferences for ICP spectra. Essentially the previous experience in converting intensities from the NBS Tables of Spectral-Line Intensities into ICP sensitivities is extended to the prediction of the degree of spectral interference using (a) ICP sensitivity ratios of analysis and interfering lines, (b) concentration ratios of analytes and interferents, (c) wavelength distances between interfering and analysis lines, and (d) the spectral bandpass of the spectrometer. Computer programs have been written for a general-purpose approach and a specific approach. In the general-purpose approach, a tabulation is created which can be consulted for any set of conditions as dictated by the spectral apparatus, the sample composition and the analytes. However, the user of the table has to look up each item of interest and to make decisions. In the specific approach, all decisions are built into the computer program. The user only formulates the matrix and upper limits of concentrations and a set of analytes and their lower concentrations of interest, and the computer program provides a set of the most sensitive analysis lines that are interference free under the specified conditions and gives the reasons for disqualification of other sensitive lines. In the general purpose approach, the sensitivity ratios and concentration ratios are taken into account by the so-called “critical concentration ratio”, defined as the ratio of the concentrations of interferent to analyte at which the ratio of the intensities of interfering line to analysis line within the spectral window of the analysis line is equal to unity. Wavelength distance and spectral bandpass are involved through the calculation of the areas of the Voigt profiles of analysis and interfering lines that fall within a spectral window the width of which is equal to the full width at half maximum of the “average” spectral line in the ICP spectrum. A discussion on the assignment of an appropriate value to the a-parameter of the Voigt profile is included. In the specific approach, the majority of the procedures of the general-purpose approach are used. However, the program computes, instead of critical concentration ratios, the intensity ratios of interferent to analyte lines for the concentrations specified. Examples of both approaches are presented in tabular form. The present status of the work is reviewed, and envisaged projects including experimental verifications and extensions are discussed. Various gross errors inherent in the conversion of the intensities of the NBS Tables of Spectral-Line Intensities are pointed out. Results of the calculations are used to illustrate the effect of spectral resolution of the spectral interferences and conclusions with respect to the resolution of spectrometers and the number of spectral windows required in ICP—AES are drawn. Linking up with earlier work of other investigators, the author finally points to the necessity of using state-of-the-art spectrometers in order to circumvent the adverse effects of straylight on the detection limits. Experimental results obtained with a 10-years old monochromator illustrate the disastrous effect of scattered radiation which occurs not only with Ca and Mg as main components, but with all major constituents that yield line-rich spectra or spectra with very intense lines.
Read full abstract