Abstrait

Evaluation of the linear calibration in terms of uncertainties

Jens E.T.Andersen

Amethod is presented for the assessment of the calibration line in analytical chemistry. It is based on a combination of a chemometric data interpretation and a calculation of the minimum standard deviation(STDEV) of experiments. The more favourable concentration range of linear calibration could be obtained by iteration. This condition links the lower limit of analysis(LLA) to an upper limit of analysis(ULA), and thus completes the statistically appropriate extension of the calibration line. In addition, a minimumSTDEVofmeasurementwas expressed in terms of STDEV’s on slope(S) and on intercept(S) and the calculation was performed by using amodified version of the lawof propagation of errors(LPE). Themethod was validated on experiments of gas chromatography(GC), liquid chromatography(LC), electrochemistry, flow-injection analysis(FIA), atomic emission spectrometry(AES), flame atomic absorption spectrometry (FAAS) and inductively-coupled-plasmamass-spectrometry(ICP-MS). The comparison of theory with experiments suggests that an appropriate concentration range of calibration was obtained by iteration when the intercept decreased to a value below a factor of two of the corresponding standard deviation. The proposed method is uncomplicated and facilitates the determination of the concentration range of calibration and the LLA in one working operation.