Modeling polyurethanes reactions using differential scanning calorimetry and micro-calorimetry

In resin transfer moulding (RTM) a composite material is made by injecting a reactive resin through a fiber bed. To make a polyurethane based composite, a reactive resin starts as a mixture of two liquids, a polyisocyanate and a polyol, and due to polymerization ends up as a polyurethane network. As a result, the rheological behavior changes from viscous to visco-elastic to elastic. To control and predict this rheological behavior, the kinetics of this reaction must be studied and modelled.
Using differential scanning calorimetry (DSC) and micro-calorimetry, the heat flow generated during the polymerization reaction can be measured as a function of time and temperature. As the heat flow is directly correlated with the reaction rate, the kinetics of the polymerization can be measured. Here a combination of isothermal and non-isothermal measurements is used in combination with different mixing rations of the initial compounds, to have a broad range of conditions in which the reaction takes place. Using a mechanistic model, where the global reaction is subdivided in a set of (elementary) reactions, the heat flow profiles can be simulated. Using a least squares curve fitting algorithm on the experimental and the simulated heat flows, the reaction rate parameters of the kinetic model can be determined. An advantage of this mechanistic modeling approach is that the concentrations of the different functional groups, like primary and secondary alcohols, presence of ether groups, catalysts, … are explicitly calculated and can be used to model the kinetics for different compositions. A robust model is achieved if it can predict the polymerization progress for all different conditions at reasonable accuracy. In a next step, the mean molecular weight can be estimated using a statistical method, the so-called “in-out recursive analysis” of Macosko and Miller. The growth of the molecules is calculated based on the conversion of the functional groups, leading to molecular weight, gelation, crosslink density and elasticity modulus calculations. In the next step, the changing molecular weight distribution will be linked to the evolution of the viscosity.