Walk into an Apple store and you’ll see that the newest phones now incorporate an extra camera to provide wide-angle capabilities, allowing us to capture both the sand we’re standing on as well as the waves crashing on the rocks in the distance together, all in gorgeous resolution. We value the ability to see as much as possible in as much detail as possible, and as a result our world is filled with similar technological marvels that let us do just that—IMAX theaters, aerial drone photography, 3-D TV, and more.
Scientists, especially those that study systems and phenomena at the nanoscale, are no different. They pursue a combined understanding of how things behave “far away” over long time and length scales and “up close” at the molecular level. To get a sense of how small this is, there are approximately eight septillion (10 to the 24th power) molecules of water in your mug. Achieving this combination of detail and scope is crucial. To design materials, we want to predict how changing certain inputs controls a material’s properties and how it will behave over time, but observing and characterizing the molecular-level behavior of a system is essential to ensure we have the fundamental understanding we need to make those predictions.
Unfortunately for scientists, getting a “wide-angle” view is not as easy as purchasing a new phone. To get a complete and detailed picture, they need to turn to a combination of different methods, since a single technique is often insufficient. This can be tricky, just like photoshopping together different images—sometimes the seams show. However, recently published works from the Center for Integrated Mesoscale Architectures for Sustainable Catalysis (IMASC) and the Center for Bio-Inspired Energy Science (CBES) show how explorations at multiple levels of resolution—termed “multiscale modeling”—can be executed elegantly. Such combinations of methodologies can be powerful for advancing materials design for sustainable energy technologies.
Modeling surface restructuring to design more efficient catalysts
One class of materials that is key for energy sustainability are catalysts, materials that accelerate reactions and thus make the process of converting one substance (such as a waste stream) to another substance (such as a valuable chemical product) more efficient. Especially promising within this field are bimetallic catalysts that comprise two different elements, each of which contributes its own advantages to the combination.
“There’s a tradeoff between selectivity [making a specific product] and activity [making a lot of product] that is addressed with this bimetallic composition,” said Boris Kozinsky, a principal investigator in IMASC. “You can have two metals that each do certain parts in the catalytic chain.”
David (Jin Soo) Lim, the lead graduate student on this project, along with Kozinsky and others, studied a bimetallic surface made of palladium and silver. This catalyst helps accelerate reactions that add hydrogen atoms to a material, such as the reaction that turns an unsaturated fat, like oil, into a saturated fat, like margarine. In industry, these reactions can help make products more valuable or less toxic. Despite the ubiquity of the reaction, not much is known about how the bimetallic alloy behaves, which is problematic, according to Kozinsky. “When you make a catalyst, you want to optimize its properties, but if you don’t know how the structure is changing, it’s very difficult to control its synthesis and reliability,” he says.
The IMASC team set out to study how the atoms on the surface can move over time. These mechanisms are crucial to understand because reactions take place at a catalyst’s surface and the surface may respond to the presence of these reactants and products by restructuring.
“Catalytic surfaces are never bystanders in the reaction, they are always participants,” Kozinsky explained.
Previous studies used a technique called density functional theory (DFT) that models the electron density of the atoms in order to find the energies of various surface arrangements. While modeling the system at the electron level helps to precisely capture which arrangements are more energetically favorable, it cannot access the time evolution of the restructuring process that converts one surface arrangement to another. Discovering these mechanisms through modeling requires many evaluations of the energies of different surface arrangements, a task that is too expensive for DFT.
Instead, Kozinsky and others combined a less precise, but faster model of the energy of the system, with an algorithm that can use those energy calculations to simulate how the system will behave over time. By using this approximate method, the researchers could model the system for long enough to observe restructuring events, which they detected automatically by having the computer flag times at which the atoms’ positions change sharply. The researchers could then focus in on these observed restructuring events, using the more precise and expensive DFT method to refine their calculations.
