SIMULATION OF C-13 NUCLEAR-MAGNETIC-RESONANCE SPECTRA OF ALKYL-SUBSTITUTED AROMATIC-COMPOUNDS

Carbon-13 nuclear magnetic resonance spectra for a collection of alkyl-substituted benzenes and polycyclic aromatic compounds have been simulated by using computer-based methodology. Several new types of structure-based parameters were developed and were shown to be useful In characterizing structural features related to the conjugated p systems present In these molecules. These new descriptors, along with others developed previously, were used to generate parametric equations for chemical shift prediction using a training set of 30 compounds. Two different approaches for the grouping of aromatic ring carbons were investigated and evaluated for use in generating accurate chemical shift models and complete simulated spectra. Partial simulated spectra were also constructed, such that the methyl carbon atom resonances were excluded, to assist in the evaluation of the ring carbon atom model equations. The external predictive ability of the derived models was evaluated through the simulation of 13C NMR spectra for a group of 44 compounds not Included in the model development process. External prediction accuracy improved when ring-bridging and non-ring-bridging carbons were treated separately.#. © 1990, American Chemical Society. All rights reserved.