Modeling Polymeric Materials

By Anonymous

2006 Danckwerts Lecture

During a one-on-one interview. Doros Theodorou, the Danckwerts Lecturer at AIChE’s 2006 Annual Meeting, explained the more-complex aspects of the hierarchical computations for polymer characterization and the commercial relevance of his promising research. My PhD thesis aimed at predicting molecular packing and elastic constants of glassy polymers and could he described as one of the first successful applications of molecular modeling to polymers. Since then, the field of molecular modeling has exploded, with applications in all kinds of materials and properties, he said.

A professor of chemical engineering in the the Dept. of Materials Science and Engineering, National Technical Univ. of Athens, Greece, Theodorou recently co-edited a new book with fellow researcher Michael Kolelyanskii on simulation methods for polymers*. We simulate materials (mainly polymers and synthetic zeolites) on a computer using statistical mechanics and algorithms, bused on the fundamental molecular sciences, in order to predict material properties from their chemical constitution. Knowledge of these properties is needed in the design of products with prescribed performance characteristics in end-use applications (i.e.. plastics with desired stiffness and strength as structural materials, better processability in the melt state, controlled permeability by atmospheric gases leading to prolonged shelf life of products packaged in them, adhesives that can bond different materials together, polymeric or inorganic membranes that can separate mixtures with high throughput and selectivity. etc.) and in the design of processes to make these products.

In general, predicting properties from chemical constitution is a formidable task. “Our computational approaches are often hierarchical, i.e., operate at various levels of description, from the detailed atomistic to the macroscopic, utilizing systematic coarse-graining (e.g., from atoms to groups of atoms to entanglement networks to continuous media) to capture properties with currently available computational resources.

The input to molecular modeling calculations is the chemical constitution of a material, which consists of building blocks, such as molecules, atoms, ions, etc. These calculations predict how these building blocks arrange themselves in space, and what properties they give rise to. What molecular modeling can do is predict a wide variety of properties starting from the same fundamental input. It can also bring out the mechanisms, i.e., why does a material exhibit the properties that it does?, and how are these properties expected to change if we change something in the constitution of the material.

Theodorou points out a particularly challenging task – computational prediction of physical properties for polymeric materials, because of the extremely broad spectra of length and time scales governing structure and molecular motion in these materials. For instance, a polymer has structure at the level of atoms and bonds (10-^sup 10^ m), at the level of entire polymer chains (10^sup -8^ m), at the level of domains in a semicrystalline material or a phase-separated blend (10^sup -6^ m or higher). A piece of Teflon is white because it contains crystallites that can scatter light, i.e., are of a length scale commensurate with the wavelength of light (10^sup -7^-10^sup -6^m).

At the atomic level we have bond vibrations with periods 10^sup – 14^ s. Individual bonds in a polyethylene melt flip between trails and gauche states every 10^sup -11^ s or longer, depending on the temperature. As a result of these flips, chains can adopt a tremendously large number of conformations. Longer pieces of chains take longer to change their shapes. The time required for a chain to move by a length comparable to its size in a melt and thereby “forget” its previous shape is 10^sup -3^ to 1 s for usual molecular weights. On the other hand, a glassy polymer, such as polystyrene, polyethylene terephthalate (PET), or the polycarbonate in your CDs, undergoes very slow structural changes (physical ageing) with characteristic times of years (10^sup 7^ s). Between vibration of chemical bonds and physical ageing of a polymer glass, we have 21 orders of magnitude in time scale, which is quite impressive.

“This challenge can only be met through the development of hierarchical analysis and simulation strategies encompassing many interconnected levels, each level addressing phenomena from the detailed atomistic to the continuum microscopic over a specific window of time and length scales. Complementary technologies, such as group contribution methods for the estimation of properties, and high-throughput experimentation aimed at the massively parallel synthesis of materials of similar constitution and testing of their properties are valuable for validating modeling methods and also for realizing the design principles reached by the modeling.

Theodoron is working with software developers for materials design that have already incorporated his methods into their commercial products. Examples are Accelrys, Inc. (San Diego, CA, and Cambridge, U.K.), and the recently founded company Scienomics SARL in France. We work closely with industrial researchers in order to develop methods and software appropriate for solving their materials design needs. For example, we have worked with BP in Naperville, IL, on permeability-related problems, with DSM Research BV in the Netherlands on adhesion and interface-related problems, and with Mitsui Chemicals in Japan on polymer equilibration methods. Although we are not a company, we welcome collaborations with industry – industrial problems are a great source of inspiration for us. Our primary need is for able and motivated young scientists and engineers (working as PhD students and post-docs) willing to embark on the development of these methods and their implementation on specific industrial problems. “This is far from routine work; it needs inspiration and dedication. I find it fascinating to enlist one’s basic scientific knowledge and mathematical skills in order to understand and predict why muterials behave the way they do, he concludes.

“It would be a great mistake to think of the content of chemical engineering science as permanently fixed. It is likely to alter greatly over the years, in response to the changing requirements of industry and to new scientific discoveries and ideas for their application.”

P.V. DANCKWERTS, 1966

Doros Theodorou, this year’s Danckwerts Lecturer, grabs the attention of the audience with a discussion of his cutting-edge research – hierarchical computations for polymer characterization that can be used to efficiently predict the polymer’s physical properties.

* Kotelyanskii, M. J., and D.N. Theodorou, Eds., “Simulation Methods for Polymers,” Marcel Dekker, New York, 900 pp. (2004). ISBN 0824702476.

Visit http://comse.chemeng.ntua.gr. for more information. This website contains a general introduction to his group’s research projectsa.

Copyright American Institute of Chemical Engineers Jan 2007

(c) 2007 Chemical Engineering Progress. Provided by ProQuest Information and Learning. All rights Reserved.