In English

Abstract

A set of characteristics calculated within the scope of quantum chemistry methods may be assigned to local ones changing from atom to atom in complex systems. Simple averaging of the related values gives rather poor characteristics of the systems because various fractions of certain atoms can have different surrounding and, therefore, different characteristics, which may not correspond to the average one. The aim of this study is searching a more appropriate pathway to transform local characteristics, e.g., atomic charges, into nonlocal ones based on the distribution functions. The distribution functions of atomic charges (CDF) could be considered as a simple tool to analyze nonuniform complex systems since specificity of different fractions of atoms reflects in the CDF shape. As a whole, the approach accuracy and efficiency depend on the quality and appropriateness of molecular and cluster models used, as well as on the quantum chemical methods (ab initio, DFT, and semiempirical) and the basis sets used. Nanosystems with dozens of molecules (clusters, domains, nanodroplets), modelling a liquid phase or interfacial layers, and solid nanoparticles of almost real sizes (> 40 units, > 2 nm) may be considered as more appropriate models of real systems than the models with several molecules and small clusters (< 20 units, < 1 nm). This approach has been applied to a set of representatives of such various materials as activated carbon, porous and nanoparticulate silicas unmodified and modified interacting with nitrogen, methane, water, human serum albumin (HSA) binding doxorubicin molecules. This approach may give information useful upon the analysis of any complex system.