News Archive

San Diego Supercomputer Expanse Used to Advance Fuel Cell Tech

Published November 1, 2023

Diagram showing the movement of atoms through a fuel cell.

Toyota Research Institute and Georgia Tech used Expanse to make progress with fuel cell technology.  Credit: Georgia Tech

By Thomas Frost and Kimberly Mann Bruch, SDSC Communications

Fuel cells, which convert chemical energy from reactants like hydrogen and oxygen to electrical energy, have many benefits over traditional combustion-based technologies. Since 1839, fuel cell technology has been greatly advanced by research teams working in fields like transportation, materials handling and energy storage. But the technology is difficult to implement due to an inability to effectively support hydrogen distribution affordably. Because of this, scientists are constantly looking for a way to make fuel cells more cost-efficient and environmentally friendly.

Recently, a team of researchers from Toyota Research Institute and Georgia Institute of Technology (Georgia Tech) used Expanse at San Diego Supercomputer Center (SDSC) at UC San Diego to discover 10 novel candidates for fuel cell technology.

“The energy conversion method used in fuel cells has high efficiencies while resulting in clean by-products – such as water vapor – as the primary emissions from fuel cells,” explained Rampi Ramprasad, a professor of materials science and engineering at Georgia Tech. “For this reason, we have long been examining the best materials to create fuel cells materials and thanks to Expanse we made significant progress earlier this year.”

Specifically, Ramprasad led a study entitled Informatics-Driven Selection of Polymers for Fuel-Cell Applications alongside Georgia Tech colleagues Huan Tran, Kuan-Hsuan Shen and Shivank Shukla as well as Ha-Kyung Kwon from Toyota Research Institute. The study was published in The Journal of Chemistry C.

“Our study used new, innovative methods based on machine learning and informatics to help design potential proton exchange membranes (PEMs) for fuel cells that would be more effective than Nafion, a polymer most commonly used with current technology,” Ramprasad said.

According to Ramprasad, PEMs are membrane materials that are permeable to protons but not much else. When looking at the polymer Nafion, there are barriers when applied as the PEM for fuel cells because of its high cost, water dependency and oxygen permeability. These issues spearhead research to optimize the conductivity and other important materials attributes by finding an alternative for Nafion. A polymer with optimal properties would improve costs and energy management for fuel cells. 

“We used Expanse to design 48 new candidates for PEM, 10 new candidates for the fuel electrode ionomer and two candidates for the oxidant electrode ionomer – next we filtered through 30,624 known polymers with predictive models to identify prospective solutions,” Tran said. “These discoveries resulted in a decrease in the time and cost relative to conventional forms of testing.”

The variables considered were water uptake, gas permeability, electron obstruction and material robustness because these properties can make for effective materials for fuel cells.

“We were able to identify several new materials that would surpass presently used materials for this application (i.e., Nafion),” Ramprasad said. “Thanks to Expanse, our study resulted in new alternatives for Nafion, for in-depth experimental investigation in the future by the community to evaluate their suitability and efficacy for new fuel cell designs.”

The study was funded by Toyota Research Insitute with computational allocations funded by the National Science Foundation Extreme Science and Engineering Discovery Environment (XSEDE) (allocation no. DMR080044).