ТОПОЛОГИЧЕСКОЕ (QSPR) МОДЕЛИРОВАНИЕ ФИЗИЧЕСКИХ СВОЙСТВ ПОЛИМЕРНЫХ СИСТЕМ ПОСРЕДСТВОМ ОПТИМАЛЬНЫХ ДЕСКРИПТОРОВ

Аннотация

Subjects of the inquiry: the research objects arc synthetic polymers and low-molecular compounds. The subject of the research was determination of the quantitative relationships between structure and property in the selected class of compounds and systems. Aim of the inquiry: development of topologic QSPR modeling for calculation of physical properties of polymers on the basis of information about their structure and with use of optimal descriptors based on statistical regularity as theoretical basis for determination of properties and purposeful synthesis of polymers. Method of inquiry: for determining of the “structure - property” dependencies optimization of correlation weights of local graph invariants has been applied. Per se this approach is a hybrid of the additive scheme and the regression analysis based on structural descriptors. The results achieved and their novelty: for the first time optimal descriptors have been used for the determining of the “structure - property” correlation. Application of local descriptors based on the optimization of correlative weights of local graph invariants for prediction of physical properties of single- and doublecomponent polymeric systems has been estimated. For the first time models of the glass transition temperature of polyarylenoxidcs, the melting temperature of oligophcnilens and octanol-water partition coefficients of organic substances have been estimated as well as the Flory-Huggins parameter, lower critical temperatures of blendings and the intrinsic viscosity of two-componcntal of “polymer-solvent” systems. Practical value: practical relevancy includes the fact that models of various characteristics of polymeric systems allows to predict by the calculation way above mentioned characteristics for wide range of single- and double-components “polymer-dissolvent” systems obtained which makes unnecessary complicated and expensive and sometimes even technically labor-intensive experiments. Sphere of usage: research method, obtained “structure - property” models and prediction algorithms can be used: • for analysis of set of polymeric structures in order to select prospective structures for technological implementation of practical tasks in physics and chemistry of high-molecular compounds; • for researches in the field material science during investigation of the “structure - property” dependencies for high-molecular single- and doublecomponents systems; • in educational process in material-sciences high education institutes.