Quantifying relationships among the molecular weight distribution, non-Newtonian shear viscosity, and melt index for linear polymers

This paper presents methodologies to quantify the relationships among the molecular weight distribution (MWD), steady-shear non-Newtonian viscosity (i.e., flow curve), and melt index (MI) of three linear low-density polyethylenes manufactured using the same technology. With the aid of computer-aided process simulation tools (such as POLYMERS PLUS), polymer producers can predict accurately the MWD from manufacturing conditions. Our methodologies help the polymer producers to extend their simulation model to predict flow curves and MI from the MWD. To do this, this paper employs (1) a modified Carreau-Yasuda (CY) model or Bersted's partition model to relate the MWDs and flow curves, (2) Bremner and Rudin's model to relate the weight-average molecular weight (MWW) and MI, and (3) Rohlfing and Janzen's model to relate the flow curve and MI. We show that the Carreau-Yasuda, Bersted, and Bremner and Rudin models work very well for correlating our data, predicting flow curves that average 3-7% error and MI values that average 2% error. In addition, for the case in which we lack MI data, we use the CY or Bersted model predictions of the flow curve to generate MI values through Rohlfing and Janzen's model. These predictions are very good, averaging only 0.5-3% error. We also show how to use the CY or Bersted model and Rohlfing and Janzen's melt-indexer model to estimate closely the low shear rate region of the flow curve using MWD/MI data alone. This case corresponds to one in which we lack flow curve data. Last, we provide practical guidelines for polymer manufacturers who want to predict the flow curve and MI using the MWD.