A real can of worms

released 12.02.08

By J. William Bell

University of Michigan engineers model the entangled polymer strands that make up plastic products and textiles on your store shelves.

Studying molten plastics, according to chemical engineering professor Ron Larson, is like opening a can of worms. "It's even worse than that, because some of the worms are branched, Hydra-headed mutants. Long strands of polymers branch off from one another at multiple points, a tangle of jiggling knots. In this can, there's no dirt. It's nothing but moving worms," Larson says.

To form a plastic milk bottle or that thin bag that you load produce into at the grocery store, molten plastic is blown up at extremely high speeds. The branches of the polymer strands are blasted out of place by the stress. As they cool, the polymers relax back into another, different set of entanglements. They leave one messy situation for another. If everything goes right, if there are no unexpected physical stresses, then that new relaxed state won't bulge or be too brittle. It'll be the ideal plastic milk bottle, and so will the other million that are churned out at the factory.

For years, and in some cases still today, plastics and other polymer-based products were designed by trial and error. Manufacturers blew that air in and watched for the bottle to crack or tear. If it did, they went back to the drawing board for a slightly modified polymer with slightly different properties.

Larson and his team at the University of Michigan are perfecting ways of modeling tangles of polymers and their properties at the molecular level—how the polymers fluctuate about, the stresses that influence that behavior, and the barriers and patterns that block the polymers from untangling.

"We're using high-performance computing at NCSA to look at something that was invisible in the past," Larson says.

Polymer melt magic

Polymers that branch into multiple long strands are exceptionally useful for industry. Branches strengthen molten plastic and other "polymer melts," as they're called by engineers. They serve as netting that gives the polymer melt just enough solid properties to avoid ripping or bursting when blown into shape. Manipulating the features of those melts—how long the strands are, and how frequently they split or branch—influences their properties.

Very small changes to the branching can yield very large changes to the polymer melt. Larson's team, for example, has found that changes to just one branch in a million possible branch points can significantly impact the properties of the melt relevant to its strength. This strength can determine whether the melt is suitable for blowing into film or spinning into fiber.

The Larson team is delivering insights into how branch points move as the polymer melt relaxes. Their simulations consider asymmetric star polymers, in which three chains branch from a single central point with one of the chains shorter than the others.

For asymmetric star polymers, the team found that the branch point hops from position to position. With symmetric star polymers, on the other hand, whose branches are all the same length, the branch point is stuck, wiggling in a smaller spherical region. As a result, it must relax its branches by pulling them into the region containing the branch point, like "Houdini dislocating his shoulder to escape a straight jacket," according to Larson.

"This result confirms a very important hypothesis about branch point motion in asymmetric stars," Larson says. "And our methods are very general, so they apply to many types of branched polymers such as 'comb' polymers which have many branches emanating from a linear backbone." The results were published in a 2007 issue of Macromolecules by Larson and Qiang Zhou, a former graduate student who is now at Standard and Poor's doing complex financial analysis for the market intelligence firm. Post-doc Zuwei Wang continues the work today.

From the molecular scale to the real world

By understanding this behavior in both linear chain and branched polymers, researchers can manipulate the features of the polymer melts to match the needs of industry. Larson works with Dow Chemical, for example, to help them figure out how to slow the relaxation of polymers in molten plastic and to transform the way they create their myriad products.

Currently computational power is still a limitation for the team, who run simulations both in house and larger, longer runs at NCSA. Even running on hundreds of processors for months at a time would not give them a clear picture of all the physical characteristics of the polymers over the full amount of time it takes them to relax. "It's impossible to reach real-world conditions, so while we choose to look at long polymers at real-world concentrations, we don't track them until they are fully relaxed," Larson says.

Fortunately, the amount of time that they do simulate reveals key features and properties of the polymers. These can then be fed into another set of simulations, currently being run by grad student Xue Chen, which allows researchers to bridge the molecular scale to the scale relevant for real-world applications.

"We want a model that captures the fundamental properties accurately," Larson says. "Otherwise, the assumptions pile up, and you’re really just guessing."

This research is supported by the National Science Foundation and Dow Chemical.

Team members
Xue Chen
Ron Larson
Zuwei Wang
Qiang Zhou