This CAREER award supports theoretical and computational research and education that seeks to develop novel algorithms to infer the architecture of polymer melts from rheological data. Addition of trace amounts of long-chain branching improves the processability of polymer melts. Despite its industrial importance, and advances in synthesis, which enable us to control the amount of branching, standard analytical methods such as chromatography and spectroscopy cannot reliably diagnose these trace levels. Rheology, on the other hand, is extremely sensitive to molecular architecture. This motivates the PI to investigate inverting contemporary models, based on the tube theory and hierarchical relaxation.
The proposed algorithms are based on the idea of Bayesian inference, which is used to transform the inverse problem of inferring structure, into a sampling problem, that is attacked using Markov chain Monte Carlo methods. This approach has four unique advantages: (i) It can be applied to systems with an unknown number of species; (ii) It has a built-in Occam's razor, which prefera less complex solutions, (iii) It can characterize multiple solutions, and (iv) It can incorporate complementary analytical information in a systematic and robust manner.
Contemporary rheological models, however, are less-than-perfect, and effort is directed in this project, at addressing these shortcomings by microscopic studies. In particular, we seek to study blends of cyclic and linear polymers to understand the process of constraint release, and to map physically different microscopic simulation models to understand the role of assumptions in coarse-graining.
This project will promote teaching, training and learning by continuing undergraduate and graduate participation in the research effort. The PI's association with an HBCU will provide a conduit for the participation of minority students in the research. This together with various strategies will help broaden participation. The PI will use computation to emphasize the connection between the microscopic structure and motion and macroscopic properties and phenomena of materials. Educational tools developed will be distributed through the PI's website.
NONTECHNICAL SUMMARY
This CAREER award supports theoretical and computational research and education that seeks to develop novel algorithms to infer the structure of large molecules that have long branched chain-like structures through rheology. Rheology involves measuring how these materials respond to deformation. Addition of trace amounts of long-chain branching improves the processability of these molecules known broadly as polymers. Despite its industrial importance standard analytical experimental methods cannot reliably diagnose these trace levels. Rheology, on the other hand, is extremely sensitive to molecular architecture, and motivates the PI to consider using contemporary models based on sophisticated microscopic theories in a direction reversed from the usual way they are used. This process is called analytical rheology, and is an ill-posed problem, which seriously impairs current methods. The research will develop a method that addresses the most serious shortcomings, which includes the inability to discriminate the number of components and to address the multiplicity of possible structures. This project also involves improving the contemporary models of polymers that would be used in reverse.
This project will promote teaching, training and learning by continuing undergraduate and graduate participation in the research effort. The PI's association with an HBCU will provide a conduit for the participation of minority students in the research. This together with various strategies will help broaden participation. The PI will use computation to emphasize the connection between the microscopic structure and motion and macroscopic properties and phenomena of materials. Educational tools developed will be distributed through the PI's website.
