External Linear Calibration With a Nonzero Intercept

1c External Linear Calibration With a Nonzero Intercept 

Such non-zero intercepts in calibration curves are examples of bias errors. The detection and characterization of bias errors, often appearing together with simultaneous proportional systematic errors (e.g., those associated with values of F 1), is a major thrust of the approach (Section 8.5.Id) promoted by Youden and Cardone (Youden 1947,1960 Cardone 1983,1986 Ferrus 1987). 

Again a worst scenario corresponds to a linear external calibration with a non-zero intercept, no internal standard and no matrix blank available. Such circumstances lead to a modified form of Equation [8.64a] by following the same reasoning  

However, Equation [8.64a] that relates qa to the concentration of analyte in the unknown sample stiU applies, so elimination of q a from Equations [8.64a] and [8.67b] as before, gives  

This discussion emphasizes that non-zero values for A usually indicate potential uncertainties which are best avoided, if possible, by diagnosing the cause and taking remedial action, e.g., if A 0 a more selective analytical method such as different cleanup and chromatography and/or increased mass spectrometer resolution or, if A 0, changing columns and/or silylation to deactivate appropriate portions of the chromatographic train etc. However, if none of these remedies improve the situation, assumption of the appUcabdity of the calibration parameters B and A to analysis of sample extracts gives the working relation Equation [8.67]. 

---