Forward Modeling of Infrared Spectra of Heterogeneous Ziegler–Natta Catalysts from Structural Distributions

Abstract

Computational chemistry has occupied an important place in the study of heterogeneous catalysts by supporting experimental findings. In a backward modeling approach, a few models are proposed based on intuition and their existence is justified by fitness to experimental results, limiting their capability to represent distribution. In heterogeneous catalysts, however, distribution is a given, due to their complex surface, and fundamental to understand their performance. In this study, we propose a forward modeling approach, where an ensemble of models is created nonempirically, and the comparison with experiments is done only after their creation. We calculated the infrared intensities of an ensemble of 132 Ziegler–Natta catalyst (ZNC) structures to recreate the ν(C═O) band of the experimental infrared spectrum in an ab initio way. The obtained peak position was 1 to 6 cm–1 close to experiments and distribution was essential to obtain a single broad band, as individual models had multiple well-separated peaks. By calculating the contribution of each subpopulation to the band, we were able to determine that the diethyl phthalate (DEP) modulator adsorbing on the catalyst had separate peaks depending on the adsorption mode, 17 cm–1 apart. In addition, we discuss the sensitivity of the ν(C═O) to the local environment of DEP and the origin of the broadness of the band in ZNCs.