A dsorption on carbonaceous surfaces: cost-effective computational strategies for quantum chemistry studies of aromatic systems

We present a systematic analysis of the accuracy and efficiency of several computational quantum chemistry models for studying reactions involving aromatic systems. In particular, we have examined different multi-layer ONIOM models in which the whole system is divided into subsystems that can be treated at different levels of theory. The carbonaceous surface is modeled by a graphene layer that has unsaturated carbon atoms to represent active sites and has different shapes to simulate the local environment of the active sites of a carbonized material. We emphasized the model performance in predicting geometrical parameters, interaction energies and infrared spectra of carbon–oxygen complexes. We found that any attempt to partition the graphene layer into subsystems for employing different levels of theory yields considerable errors. However, it is possible to obtain reasonable accuracy by using the same level of theory for the whole system at different basis sets. This computational strategy can predict accurate geometrical parameters, interaction energies and infrared spectra of common oxygen complexes at lower computational cost.  2002 Elsevier Science Ltd. All rights reserved.