Collision chemistry

Some scientists look to the outer reaches of the universe to find new knowledge. For William Hase, a chemistry researcher at Texas Tech University, essential questions remain closer to home: in the chemical bonds, and molecule and surface collisions that occur all around us.

While classical notions of chemical reactions may hold for some molecules, mounting evidence shows that everyday reactions are more complicated than previously thought. Different kinds of energies — vibrational, rotational and translational — may be required to activate a reaction. Moreover, the rupture and formation of bonds during chemical reactions involve electron-electron and electron-nuclei interactions, which complicate matters.

Laboratory experiments can highlight the broad strokes of chemical reactions. Imaging techniques like atomic force microscopy help identify the three dimensional characteristics of the molecules. But neither method has the resolution required to explore the atomistic dynamics at work when two reactants come together to form something new.

Supercomputers, however, have allowed Hase to reproduce laboratory results and add to the canon of knowledge about chemical reactions: specifically, the dynamics of individual molecules when they collide.

"For molecules to react, they need specific types of geometries, specific types of energies, and certain types of motion," Hase explained. "Reactions are often not random. They're very specific in terms of what's going to react."

Partnering with researchers in Germany, Washington, Iowa, and at his own institution in Lubbock, Hase uncovered the "roundabout" motion of oxygen atoms bonding in SN2 reactions, modeled heat moving between interfaces of materials, similar to those in the tips of space shuttles, and revealed how ions get trapped on the surface of biological materials.

This wealth of collaborations stems, in part, from Hase's decades of experience with computational chemistry. It also derives from the fact that he developed the computer program, VENUS, that researchers use to perform classical chemical dynamics simulations, as well as software packages that interface VENUS with computer programs (such as NWChem) to allow the direct use of quantum chemistry for chemical dynamics simulations.

"We often do not know what we're going to find before we do the simulations," Hase confessed. "We're trying to recover some experimental observation, but there's so much more detail in the study."

Atoms can be in various orientations when they come across a potential bonding partner and some aspects of bonding are fundamentally random. Because of this randomness, it is necessary to repeat simulations hundreds to thousands of times, sampling all possible reactant conditions, to derive a statistically valid answer.

This requires serious computing horsepower.

Hase uses the Ranger and Lonestar supercomputers at the Texas Advanced Computing Center, and his collaborators at Iowa State University use Kraken at the National Institute for Computational Sciences in Tennessee, another high-performance system that, like Lonestar and Ranger, is part of the Extreme Science and Engineering Discovery Environment (XSEDE).

Hase's work spawned recent articles in Nature Chemistry, the Journal of Physical Chemistry, the Journal of Chemical Physics, and the Journal of the American Chemical Society (JACS). The article in JACS describes efforts to understand the natural synthesis of abietic acid (pine resin) by trees. Simulations of how resin is produced revealed that the reaction could yield two potential products, but only one occurs in nature.

Why is this the case?

After articulating the behavior of the constituent atoms, the researchers concluded that an enzyme involved in the process is "steering this reaction to avoid the generation of byproducts with different molecular architectures."

Similarly, in the case of the SN2 reaction — a fundamental class of reactions where a pair of electrons are donated to an electron-deficient center, bonding to it and expelling another group — Hase discovered that the exchange can take a few forms. Sometimes a simple handoff occurs. In other cases, the two molecules spin around each other before they exchange electrons. The difference depends on the orientation of the reactants as they collide.

"Nobody would've imagined some of the mechanisms that we discovered," Hase remarked. "It looks like one of the atoms roams around a molecule like a car moving around a roundabout in Europe."

It may seem esoteric to worry about the traffic conditions that atoms face as they trade electrons, but as Hase pointed out in a Nature Chemistry review article, by understanding these multiple pathways, "it may be possible to ‘surgically' direct the reactions towards only one pathway by their selective excitation."

Better, faster reactions to create larger, more complex molecules could revolutionize materials science and industry. Simulations to uncover the hidden dynamics of reactants are the first step.

Recently, Hase began exploring ion implantation and its possible applications for molecular recognition in medicine. A whole world of new applications lies hidden in this area.

"Each peptide molecule has properties in which it interacts strongly with a small subset of other types of molecules, a process known as molecular recognition," Hase explained. "Molecules have certain types of shapes or functional groups on them, or certain types of interactions where they preferentially interact with other types of molecules. We're colliding peptide ions with surfaces, and the collisions are used to template an organic surface with a specific type of peptide on it."

This ability to create molecular templates may one day allow scientists to design small molecules that can control cells during disease or slow aging.

Julie Laskin, a collaborator of Hase's based at the Pacific Northwest National Laboratory, said Hase's efforts to understand how readily those molecules can bind to the surface, or how well they are trapped in the surfaces, is important for experimentalists. The information helps them tune their experiments and suggests new areas of study.

"Our interest is in the preparation of biological surfaces, where you want to covalently tether some molecules to the surface to make a good biological surface on which cells can grow," Laskin said. "The work that Hase's group is doing is wonderful. It provides us with some guidance and some insights that we can compare with our experimental data."

With his focus on computer simulations, Hase may seem like an atypical chemist, but computational chemistry is becoming more common every day.

"Many people think of chemists in the lab, doing synthesis with beakers, but I never liked that side of chemistry," Hase said. "I always liked math. I always liked equations, being able to describe things, and I always liked computer programming."

After publishing more than 200 papers and several books, it's fair to say Hase's interest in computational chemistry has paid off.

"At the very beginning, I was a new kid on the block, but now I'm an old timer," Hase reminisced. "It's amazing…as the simulation models get better, this approach can only become more and more powerful."

A version of this story first appeared on the TACC website.