Additive Background Correction in Multivariate Instrument Standardization

Nonideal instruments give responses that do not follow Beer's law. Most spectrometers have some component of stray light that adds a structured, rather than constant or linear, background to absorbance spectra. When instrument standardization is performed using the multiplicative model introduced by Wang et al. (Anal. Chem. 1991, 63, 2750), the model does not explicitly include the additive term. This term, if uncorrected, can cause multivariate standardization to provide an incomplete transfer of calibration models between instruments. Using both simulation and real data, the current paper demonstrates the effect of the additive background term in multivariate standardization and provides an effective means by which such additive differences can be corrected.