THROUGH A SERIES of nine research articles — each one of which colleagues worldwide consider a ‘landmark’ publication — Howard Curler Distinguished Professor of Chemical and Biological Engineering Juan de Pablo has demonstrated unprecedented advances in developing powerful computational methods that enable researchers to conduct molecular simulations of complex fluids. With his students, de Pablo has invented new simulation methods, algorithms and theoretical formalisms that are key to establishing quantitative relations between atomic-level structure and interactions, processing conditions, macroscopic properties, and performance in applications.
Researchers have a clear understanding of how molecules interact in complex fluids, polymers, electrolytes and biomolecules. Researchers also know what fundamental forces are acting between the molecules, and they can assign equations that govern this behavior. Yet, even with today’s powerful computers, solving these equations can take months—even years, if it is at all possible. That’s where de Pablo and his students have made key strides. “Quicker processors and ever-expanding memory has been rapidly consumed by larger systems and more complicated molecules, particularly synthetic and biological polymers,” says colleague Frank Bates of the University of Minnesota. “Juan de Pablo seems to have figured out how to overcome this limitation through the implementation of revolutionary algorithms.”
The algorithms are mostly stochastic; basically, they generate random realizations of the problem at hand and, using well-defined rules, researchers can assign different weights to those realizations and relate the results to experimental data. Firmly grounded in the principles of statistical mechanics, the algorithms and methods combine elegance, practical usefulness and versatility in applications, says colleague Hans Christian Öttinger of the Swiss Federal Institute of Technology, Zurich. “They are used also by researchers in soft matter physics, biophysics, chemistry, interfacial science ... and they have become standard references for all the leaders in the field of Monte Carlo simulations.”
The research papers, published from 1999 to 2003 either in the Journal of Chemical Physics or Physical Review Letters, demonstrate the feasibility and value of joining several ideas: replica exchange techniques, expanded canonical ensembles, and simulations in a multidimensional space of ensemble variables. The early papers set forth the technical utility of the combined method, while later papers use the method to reveal new insights into several theoretically and practically important systems, including polymers, glasses and asymmetric charged systems. “The attention to both technical detail and physical significance in these works is an essential part of their ultimate impact,” says colleague Gregory Rutledge of the Massachusetts Institute of Technology. “The first is crucial to any ‘early adopters’ of the work; without the confidence and reliability it provides, others such as myself may not take the risk to implement. The second serves the goal of ‘teaching’ the community how and why it may benefit from the adoption of such new methods.”
Researchers have cited the papers on more than 550 occasions, according to citation database Web of Science. More importantly, says colleague Doros Theodorou of the National Technical University of Athens, the papers have influenced young chemical engineers interested in statistical mechanics and multiscale simulations as tools for rational, molecular-level design. “His overall modeling and simulation work points the way to new nanoscale products and processes that will form the focus of the chemical, materials and biomolecular engineers of tomorrow,” says Theodorou.