Novel Methods for Simulating the Self-Assembly of Complex Particles

Sunday, October 16, 2011
Exhibit Hall B (Minneapolis Convention Center)
Eric Jankowski, Chemical Engineering, University of Michigan, Ann Arbor, MI

The self-assembly of nanoparticles is a ubiquitous phenomenon that underlies problems of crucial importance in medicine and biology and which can be exploited for the fabrication of customized materials and smart devices.   The link that connects colloid crystal formation to virus capsid assembly to enzyme catalysis to ion transport in fuel cells is the minimization of free energy between interacting thermal particles with complicated shapes.  Given the recent proliferation of synthetic techniques that enable the fabrication and characterization of mondisperse anisotropic particles on nanometer length scales we have unprecedented opportunities to utilize these building blocks to solve practical problems.  My research program focuses on the design and implementation of simulation techniques for studying complex self-assembling systems, providing a conduit between theory and experiment.

Of particular utility in the design of building blocks for robust assembly of target patterns are screening methods that can identify potential barriers to assembly.  I demonstrate the efficacy of a pathway-based approach I have developed for the prediction of strong self-assembly candidates for models of patchy anisotropic colloids, CdTe/CdS tetrahedra, and sticky spherical clusters.  This technique depends upon the calculation of vibrational and rotational entropic contributions of clusters to partition functions and I discuss the details of these algorithms.

Dissipative systems such as ATP-fueled enzymes or self-propelled particles are examples of complicated systems that can require special modeling considerations that have only recently become straightforward to implement.  Using a massively multithreaded approach on consumer graphics hardware I study a model of self-propelled colloids and virus capsomer analogues.  In doing so, behaviors are observed that are inaccessible on node-based computing architectures without considerable investments in capital and development time.  Control over structural morphology is demonstrated via tuning of thermodynamic parameters and the development of new theory to describe these systems is discussed.

Extended Abstract: File Not Uploaded
See more of this Session: Meet the Faculty Candidate Poster Session
See more of this Group/Topical: Education