Ethan A. Mastny
B.S. Ch.E., Brigham Young University, 2002
Ph.D., University of Wisconsin-Madison, 2007
Crystallization applications range from small volume, high-value added products such as pharmaceuticals and photomaterials, to continuous high-throughput products such as fertilizers. In order to make high-quality products crystal producers must be able to control crystal size, habit, polymorph, and purity. Traditionally crystal engineering has relied upon empirical practices or models generated from experimental data.
In contrast to the traditional approach, this research develops the tools necessary for a bottom-up modeling approach of crystallization. This approach includes modeling the important phenomena of crystallization on molecular, microscopic, and macroscopic time and length scales. Techniques for communication between the models on these different time scales will be developed.
The overall objectives of this research are:
The current focus of the project is to use molecular simulations to predict melting temperatures and pressures. A newly developed Monte Carlo algorithm which can directly calculate the Density of States (the number of possible configurations at a given energy) of a system is begin used to calculate the melting temperature of the Lennard Jones system and Sodium Chloride.
Density of states of the melting of an FCC Lennard Jones Crystal (left)
and melting curve of the FCC Lennard Jones Crystal (right).
Another current focus of research is modeling and control of gas reactions on a solid lattice. A well established method of simulating reactions on a surface is to simulate the individual reaction events that occur according to weighted probabilities. Under this framework, both macroscopic concentrations and the orientation of particles on the lattice affect the system behavior. During a gas reaction that occurs on a surface, adsorption, desorption, and reaction occur on one time scale, and diffusion occurs on a much faster time scale. Studying the transient response of a lattice gas becomes computationally intensive because the bulk of the computation time is spent simulating diffusion events. One focus of this research is to develop methods for quickly simulating these diffusion events, thereby enabling reaction events to occur.
These stochastic lattice gas models are also being used to study the control of inherently stochastic systems. The performance of Proportional Integral Derivative, Model Predictive Controllers, and Nonlinear Model Predictive Controllers on this stochastic lattice gas will be compared.
|Personal Web Page: www.che.wisc.edu/~eamastny|
|[ Home | People | Projects | Recent Presentations | Publications | Tech Reports | Contact Us ]|
University of Wisconsin
Department of Chemical Engineering
Madison WI 53706