Past Winner
2006 NSERC André Hamer Postgraduate Prize

Erin Johnson

Doctoral Level

Dalhousie University

Computer-driven simulation models have become an indispensable part of every chemist's tool kit. But, even though they provide an inexpensive way for researchers to predict the chemical properties of a system before heading into the laboratory, existing techniques still have a long way to go before they produce accurate results in every situation.

Computational chemist Erin Johnson's doctoral research proposal, for which she earned an NSERC André Hamer Prize, focuses on refining existing modeling methods and developing new ones in an effort to increase their accuracy and efficiency.

Just a few decades ago, simulation methods were dismissed by many chemists as being of little use. Before the 1990s, Johnson says, "It would take excessively long periods to treat any molecule that was relevant, or the methods were so crude that you couldn't come up with any quantitative predictions."

Even today, the chemical interactions in systems larger than a few atoms are too complex for any computer to calculate exactly, so computer models work by developing approximations. The better the modeling method, the closer the approximation comes to representing reality. "Our hope is to come up with a new method that is able to give a higher level of accuracy than the methods that are currently available so we can have an improved description of the chemistry," says Johnson.

She works primarily with density functional theory (DFT) methods, which model electron density in a substance. DFT methods use far less computer power than earlier simulation methods, but lack accuracy under certain circumstances. Johnson's doctoral research is on developing a DFT approach that yields accurate results in systems characterized by dispersion interactions (weak attractions between molecules).

Successful results would expand the range of chemical processes that can be modeled by DFT and constitute a major development in the field. The ultimate vote of confidence in any new methods would come if they get incorporated into commercial computational chemistry software packages, such as the very popular Gaussian software.

The methods being developed by Johnson and her thesis supervisor, internationally renowned computational chemist Axel Becke, have potential uses in virtually any chemical process, including drug design, developing engine coatings and more. And, as methods improve and computer technology advances, more possibilities open up. "If we could come up with a fast enough method, we could start modeling proteins, even though they are very large," predicts Johnson.

Even with the rapid rise in popularity and increasing sophistication of simulation methods, laboratories won't become obsolete any time soon. Johnson says that the two complement one another. "Hopefully you get to the point where before you go into the lab you can model something theoretically and get an idea of whether it's going to work. It's always a partnership of using theory to predict experiment and using experiment to test theory."