Computational modeling of carbon supercapacitors with the effects of surface curvature included. Credit: Jingsong Huang, ORNL

A fair number of recent posts have been about nanostructured porous materials (PCCMs, diamond aerogel, tunable gold). The interesting characteristics of these materials necessarily depend on the porosity: What is not there makes the material what it is. The massive surfaces, bulk surfaces actually, can act in surprising ways and it is important not to discard seemingly anomalous results, as a recent story out of Oak Ridge National Lab shows.

The ORNL work goes back to a paper published in 2006 by ACerS Fellow Yury Gogotsi, professor of MSE at Drexel University, in Science (doi 0.1126/science.1132195). Gogotsi’s group was studying carbon supercapicitors, and were particularly interested in a carbide-derived carbon class of porous carbon materials. Carbide-derived carbon is synthesized by subjecting metal carbides to high-temperature chlorination, which removes the metals and metalloids as chlorides, leaving behind a nanoporous carbon structure comprising 50-80% open volume. The pore structure of CDCs can be carefully controlled to within a very narrow size disribution; average pore size can be tuned to within 0.05 nm for pores in the 0.5-3.0 nm range.

Supercapacitors (sometimes called electrical double-layer capacitors) are energy storage devices that store charge by adsorbing ions on the surface of highly porous materials taking advantage of the electrostatic separation between electrolyte ions and electrodes with high surface areas to up the capacitance. The ability to synthesize CDCs with huge porosity in a tightly controlled size range, makes them prime candidate materials for energy storage, and the payoff could be significant — typical dielectric capacitors have capacitances in the range of microfarads per gram of active material, while supercapacitors can have capacitance values in the tens of Farads.

Gogotsi’s work took a turn for the unexpected with the discovery that capacitance increased, a lot, for pore sizes smaller than the solvated ions. As the paper reports, “Decreasing the pore size to a value approaching the crystallographic diameter of the ion led to a 100% increase in normalized capacitance.” The anomalous effect was attributed to a distortion of the solvation shell around the ion, similar to the distortion of a balloon when it’s squeezed through an opening smaller opening. The distortion places the ion closer to the electrode, increasing capacitance.

Such a startling anomaly was questioned at the time, according to the ORNL story. A team of computational modelers, computational chemists Bobby Sumpter and Jingsong Huang and computational physicist Vincent Meunier, used ORNL’s supercomputers to model interactions between ions and the porous carbon surface at the nanoscale using density functional theory.

Computational modeling refers to a spectrum of computer modeling and simulation methods that sort out roughly by scale. Density functional theory, more generally categorized as “electronic structure methods,” models on the smallest physical scale — atomic structure up to crystal structure (which, amazingly, spans orders of magnitude of scale). Most engineers and scientists are very familiar with the component-scale computer modeling methods of finite element, finite difference, finite volume analysis, etc. Same idea, different scale. (An emerging field of materials science known as “integrated computational materials engineering,” is generating a lot of interest, and we hope to give it more attention in subsequent posts.)

Through simulation, the ORNL team was able to determine that the ion “easily pops out of its solvation shell and fits into the nanoscale pore.” Sumpter explained that the ion actually desolvates in the bulk, driven by electrostatic potential and van der Waals forces pulling it into the porosity and “…in fact it’s very easy for it to get in.” And, Sumpter says, the model they developed explained all the data.

Because nanoscale porous materials have huge surface areas, they also tend to have a lot of topography with a variety of concave and convex curvatures. The simulation research revealed that the topography can make an enormous difference in capacitance. For example, positive curvature surface mounds can store and release ions much quicker than negative curvature holes.

It’s not just geeky fun, though. The ORNL team had already worked with Rice University to use atom-thick sheets of carbon materials to create a supercapacitor prototype that is transparent, flexible and can be wrapped around a finger. As Sumpter said in the story, “…we’ve gone all the way from modeling electrons to making a functional device that you can hold in your hand.”