Eigenvector mapping: a method for discerning solvent effects on vibrational spectra.

Abstract

This paper reports a density functional theory (DFT) analysis of the adenine spectra in a hydrogen-bonding environment. We compare the theoretical vibrational spectra of 26 model systems in which water has been hydrogen bonded to adenine with the experimental frequencies of the solid state infrared spectra (150-1700 cm(-1)) of polycrystalline adenine and the experimental frequencies observed in matrix isolation spectra of adenine [J. Phys. Chem. 100 (1996) 3527]. The vibrational eigenvectors of adenine are compared by taking the dot product to determine how the normal modes of the 15-adenine atoms are affected by different hydrogen bonding geometries. Using the isolated adenine molecule as a reference permits a comparison of different calculated spectra in terms of the projections of various normal modes and the determination of the potential energy redistribution among normal modes. This method creates a map of the normal modes using the isolated adenine molecule as a reference. Improvement in agreement between the polycrystalline data and a model of adenine with four waters is most striking. The improvement in the fit between matrix isolation data and a model of adenine with a single water was not as dramatic as the fit seen for the polycrystalline data, but the fact that a single hydrogen-bonded water shifted the spectra of the model to a closer fit than that of isolated adenine is important. We call this method eigenvector mapping. The eigenvector mapping method can be used to extract the normal modes of a parent molecule from a solvent model system. The application of this method is important because it aids in the interpretation of complex molecular interactions in terms of the spectrum of an isolated molecule. The eigenvector mapping procedure will be shown to greatly improve the correspondence between the model and the experimental data.