Prediction of solubility and diffusion properties of pesticides in polymers

In the agrochemical industry, the use of controlled release technology for the delivery of pesticides to the environment has numerous advantages, from optimized delivery of the Active Ingredient (AI) to reduction of possible hazards to humans and the environment. The ability to model and study the delivery of AIs from controlled release devices is a very useful tool in product design where different pesticide-polymer combinations need to be tested to obtain the desired release behaviour. A model-based analysis of the release from different product alternatives provides, therefore, significant reduction of time and economical resources usually needed for the experimental measurements. The objective of this thesis is to develop controlled release mathematical models and integrate them with predictive models for the estimation of the properties required by the release models. With this, the delivery of an AI from a specific controlled release device can be studied without the need, in principle, of additional experimental measurements. According to the stated objectives, a model for the release of AIs from a microcapsule device has been developed, tested and further extended to include important special effects observed during the initial periods of delivery. The critical properties that control the delivery of an AI from a controlled release device have been identified as the solubility and the diffusivity (of the AIs) and therefore predictive models for their estimation are proposed. For the study of the solubility of the AIs in the polymers, an approach involving the estimation of activity coefficients of the compounds in polymers is suggested. A group contribution based model for the estimation of these activity coefficients is selected and further developed to allow for the calculations involving the complex pesticide molecules. In the case of the diffusion coefficient, the predictive model is based on the free volume theory of diffusion, with special attention to its application to large molecules representative of pesticides. The property models for solubility and diffusivity are integrated with the release model, and used to study the release behaviour of typical industrial pesticides encapsulated within microcapsules. With this, the overall performance of the predictive property models and the release models is assessed. Additional examples highlight that the developed models can also be applied to the release of pharmaceutical AIs. Finally, the basis for a predictive model-based analysis of the release behaviour of different pesticide-polymer alternatives has been established through a systematic study of the important issues, such as, need for predictive models and data needed for their development and verification.