ABSTRACT
This chapter discusses the classic least-squares straight-line calibration, considering also common mistakes and misunderstandings. It provides some variations on the classical calibrations, such as one-point calibration, calibration models with heteroscedastic data, calibration using internal standards, calibration with standard addition, non-straight-line calibration, calibration after linearization transformation, inverse calibration, orthogonal calibration, and multivariate calibration. The relationship between the concentrations of the calibration standards and the instrumental response is expressed by a mathematical model, which allows predicting the concentrations of new sample. A distinction should be made between outliers in the calibration set when standards are injected once and outliers in replicate measurements at a given concentration, since the former will have more influence on the calibration model than the latter. Calibration lines are often used in chromatography, where the preference is given to a straight-line Ordinary least-square (OLS) model. The validation of the model is a critical step and a compulsory part in the construction of calibration curves.
