Multivariate curve resolution techniques
can be used in order to
extract from spectroscopic data of chemical mixtures the contributions
from the pure components, namely, their concentration profiles and
their spectra. The curve resolution problem is by nature a matrix
factorization problem, which suffers from the difficulty that the
pure component factors are not unique. In chemometrics the so-called
rotational ambiguity paraphrases the existence of numerous, feasible
solutions. However, most of these solutions are not chemically meaningful.
The rotational ambiguity can be reduced by adding additional information
on the pure factors such as known pure component spectra or measured
concentration profiles of the components. The complementarity and
coupling theory (as developed in J. Chemometrics
2013
27, 106–116) provides a theoretical
basis for exploiting such adscititious information in order to reduce
the ambiguity. In this work the practical application of the complementarity
and coupling theory is explained, a user-friendly MATLAB implementation
is presented, and the techniques are applied to spectral data from
the rhodium-catalyzed hydroformylation process.