Der einfluss des untergrunds auf die gestalt spektrochemischer eichkurven

Abstract

Spectrochemical calibration curves in the Δ Y — log k - diagram turn out to be strongly curved in the region of faint line intensities, if the intensity of the spectrum background is not taken into consideration. If the background intensity is deducted, then in general a straight line calibration curve results. The equations for the (true) straight line and for the uncorrected calibration curve are set up. Each may be calculated from the other. If the curvature of the calibration curve is not too pronounced, it may be replaced in limited concentration ranges by a straight line which approaches it as closely as possible. The equations for the calculation of these substituted straight lines (“Ersatzeichgerade”) are developed. It is shown that the true straight line and the calibration curve can with certain assumptions be calculated from the substituted straight lines, and conversely. Thus it is possible from the data for a substituted straight line to obtain the calibration values for lower concentration ranges for which a simple extrapolation would lead to large systematic errors. For the systematic errors of analysis which arise through the use of a substituted straight line, a simple formula is given. From this is derived a graphic representation from which one can see whether or not a background correction is to be made, if the systematic errors in a definite concentration range are to be kept under a pre-established limit. The theory is illustrated by an experimental example (determination of Mn in Al). In the appendix it is also shown what effect may be expected on the background correction in the supposition that all intensity ratios are calculated with a false filter value. Even then the calibration curve is straightened by means of the background correction, although the corrected curve is not precisely a straight line.