Classification of model topologies using the δ versus G^* plot

Abstract

Most studies dealing with the quantitative characterization of long-chain branching (LCB) concentrate only on the degree of branching as the characteristic measure. A more complete description of branching structure requires knowledge of the topology, meaning position and length of every single branch. Topology is a vital factor in rheological behavior. However, it is not easy to isolate the influences of the molecular weight distribution (MWD) from LCB effects. To overcome this limitation, we combine MWDs with linear viscoelastic measurements to give a more comprehensive fingerprint of branched samples. The δ versus G^* plot proves highly beneficial as part of the method, representing all samples on one comparable scale, eliminating fundamental differences between samples that are not due to LCB. We use the obtained datasets for evaluation of a classification algorithm aimed at assigning unknown samples to one of several known topological classes. Classification is based on a phenomenological observation of the samples’ characteristics, emphasizing the departure of rheological behavior from the expected data for completely linear topologies. We attempt to evaluate the applicability of linear viscoelastic data for the identification of topology. Our work starts out from a novel idea to treat these data and presents first results. The method is promising but requires more working data to reach its full potential. Without this, classification success remains limited.