Toward a generalized algorithm for the automated analysis of complex anisotropic NMR spectra

An existing algorithm, founded on the works of Stephenson and Binsch, for the automatic analysis of isotropic or simple anisotropic NMR spectra has been improved to treat very complex NMR spectra of molecules dissolved in nematic solvents. The main options added to the original algorithm are a wider choice of smoothing functions; the use of the principal component regression method; and the possibility of selecting molecular coordinates, order parameters, and spectral parameters as variables of the problem. By means of these new options, it has been possible to analyze automatically NMR spectra (even depending on 27 spectral parameters) of 16 molecules in an anisotropic environment. Details of each case are discussed. (C) 1998 Academic Press.