The prediction of material properties is crucial to minimize the costs of experimental exploration of new materials. This will allow the design of the microstructure to create a material with the desired properties and the design of the optimal processing conditions. Therefore, we develop a computational tool for a priori prediction of the properties of nanostructured polymers dependent on processing conditions and surface energies of the components. Our hypothesis is that the relative nonpolar and polar surface energies of the components as well as the processing conditions, such as flow type and stress, control the nanoparticle dispersion of nanocomposites. Therefore, process-structure-process relationships have to be known. Up to now, nanocomposites have not been analyzed sufficiently regarding their process-structure-property relationships.
In a first step, melt mixing under varying process conditions, such as rotation speed, is performed to provide a systematic set of data on processing parameters and outcomes for systems in which interface parameters, such as surface energies, have been controlled. Generally, the polymer extrusion processing can be divided in the following functions: solid conveying, plastification, melt conveying, devolatilization, dispersion and distribution and pressurization. Analytical models are applied to correlate the specific energy input in the material and the surface energy with the nanocomposite microstructure. The processing conditions regarding laminar and elongational flow have to be well known. Polymers are used, that have various surface tensions, like PP and PMMA. The silica filler is modified in order to gain different surface energies of the filler. The shear stress and residence time are the key parameters in creating good dispersion and are utilized as processing descriptors parameters in the model. Therefore, the rotation speed of the extrusion processing is varied. Second, the microstructures are quantified via TEM and the image-based microstructure characterization is conducted to statistically quantify the microstructure using a two-point correlation function and microstructure descriptors. In a third step, the surface energies for each polymer and functionalized nanoparticle is calculated using heuristic Materials Quantitative Structure Property Relationships (MQSPR) models based on the chemical properties of the constituents.
The used descriptors are for example based on electron density distribution, Electrostatic Potential and Active Lone Pair potential. The physical properties are then related to the 3D microstructure and through statistical learning we have developed the analytical relationship between the selected structural parameters (descriptors) and the infinite dimensional 2-point correlation function.
With this method, we are able to predict material properties from the structure of nanocomposites and we can design the processing conditions to achieve the desired material structure.