Deductive Prediction of Measurement Precision from Signal and Noise in Photomultiplier and Photodiode Detector of Liquid Chromatography

Abstract

Recently, a probability theory to predict the precision or relative standard deviation (RSD) of measurements in analytical instruments has been proposed. The aim of this paper is to examine the precision of data obtained by a high-performance liquid chromatograph (HPLC) equipped with photodiode detector and photomultiplier on the basis of the above-mentioned theory. The baseline drift, which is often formulated as 1/f noise, is approximated by the mixed random process of white noise and Markov process. Of six parameters necessary for the uncertainty prediction, three parameters are determined from the power spectral density of the baseline drift: the standard deviations (SD), w, of the white noise and the SD, m, and retention parameter, ρ, of the Markov process. The others are signal domain, k_f, signal area, A, over domain, k_f, and independent error, I (mainly from the injection error). No arbitrary constants are involved. The prediction is shown to be superb for peaks with various areas, heights and widths over a wide concentration range in the HPLC analysis for some aromatic compounds. The applicability of the uncertainty prediction in analytical chemistry is discussed.