(559g) Systematic Kinetic Modeling and Optimization of Polycondensation Reactions

Polycondensation is one of the dominant mechanisms for the synthesis of a variety of polymers. Polyesters, formed by condensation reactions between acids and alcohols, and polyols, which are condensation polymers of alcohols represent a wide class of commodity and specialized polymers being researched and applied worldwide today. These polycondensation systems are usually characterized by reversible reactions that require the removal of water or alcohol, by some means, to drive the growth in molecular weights and polymer properties. Furthermore, by-product formation reactions need to be minimized to the extent possible. The challenge here at the research stage for a new polymer is to identify the optimum conditions that deliver the desired polymer properties with minimized by-product formation.

We have developed a systematic methodology that allows us to address the above challenge for polymer research initiatives involving multiphase polycondensation reactions. The first step in addressing this challenge is the development of the reaction mechanism and a first principles kinetic model, which is based on tracking functional groups in the reaction system. The kinetic model is then used to guide the experiment design to further clarify the reaction mechanism and kinetics, through a step we call virtual experimentation. Following the development of an adequate kinetic model, dynamic optimization is applied to develop the optimum conditions to deliver desired polymer properties. This allows us to quantitatively determine the ability of this reaction system to achieve our objectives in terms of polymer properties and economic metrics. While the aspects of the above approach are well known, we have been able to execute effectively on these aspects by applying a systematic software framework to aid in the process of mechanism discovery, kinetic estimation and optimization, thus enhancing the productivity in the research workflow. The software (REX) allows us to perform parameter estimation and dynamic simulation and optimization, thus allowing the researcher to focus on the chemistry and the physical aspects of the systems.

This approach has allowed us to systematically address the common questions encountered in several projects, such as:

What is the optimum monomer ratio (Alcohol/Acid Ratio) for the polymer ?

Issues: If the monomer alcohol is volatile, which is common, the ratio should be higher than 1, but a ratio that is too high could result in lower average molecular weights due to loss of balance in terminal groups. Furthermore, the presence and speed of the trans-esterification reactions affects this ratio significantly.

What is the optimum pressure profile ?

Issues: The removal of water is a key factor in increasing the molecular weights, which is usually accomplished by lowering the reactor pressure as the reaction proceeds. In batch polymerization, early pressure reduction can speed up the polymerization but also results in potential loss of volatile monomer, whose concentration is high at the early stage.

What is the optimum temperature profile ?

Issues: This is determined by activation energies of the desired reactions versus the thermal scission and byproduct formation reactions.

How does monomer solubility affect the polymerization ?

Issues: The monomer solubility (especially for the acids in the case of polyesters) affects the polymerization by altering the apparent acid alcohol ratio. This affects the speed of consumption of alcohol monomer, thus affecting the optimum pressure and temperature profiles.

We present case studies highlighting how answers were developed to the questions posed above, resulting eventually in yield and property improvements in the desired polymer. The development of the functional group based kinetic models will also be described. This approach has been applied to several projects at Mitsubishi Chemical. The key point is to start the analysis at an early stage, thus allowing systematic reasoning to guide the research process.