Fully automated indirect hard modeling of mixture spectra

Indirect Hard Modeling (IHM) is a recently introduced phenomenologically motivated method for quantitative spectroscopic analysis. It is especially suited for highly interacting and even reactive mixtures with overlapping spectra. IHM is based on parametric pure component models that allow the thorough modeling of nonlinear effects. During calibration, the original procedure requires substantial user input for the selection of peaks affected by molecular interactions. The quality of the results therefore strongly depends on the expertise of the user. In order to overcome these limitations a new rigorous mathematical method is presented that adaptively identifies the affected peak parameters for each spectrum. The new approach provides an objective basis for IHM analysis. It eliminates the most time consuming step and thus facilitates the study of even more complex systems. Further it is fully automated, which makes IHM accessible even to inexperienced users. The core of the method consists of two algorithms for identification of molecular interaction which utilize results from regularization theory. Results for ATR - IR and Raman spectra are discussed. (c) 2007 Elsevier B.V. All rights reserved.