Simulating the Future of Fuel
Fossil fuels are a staple of modern civilization, proving themselves a tried and true energy source available to the masses at reasonable prices. Yet the full price is not paid in human currency, but in environmental damage to the planet.
Cleaner alternatives, like fuel cell systems, have existed for decades but their impact has been limited due to expensive components and complicated design. Professor Gyeong Hwang’s research team in the McKetta Department of Chemical Engineering is now improving the efficacy of fuel cells using first-principles-based, atomic-level computer simulations at the Texas Advanced Computing Center (TACC).
Fuel cells are similar to batteries, consisting of two electrodes (one negative, one positive) and an electrolyte between them. Conventional batteries work by transforming stored chemical energy into electrical energy, which is then transmitted via a circuit to carry out work. Unlike batteries, fuel cells can never be depleted because they depend on a steady inflow of fuel rather than an internal reserve.
“Fuel cells can power almost any portable electronics that use batteries, including cell phones and laptops,” said Professor Hwang. “Moreover, fuel cell vehicles have the potential to significantly lower greenhouse gas emissions and reduce our dependence on oil.”
Hydrogen is a popular candidate to fuel this technology. When hydrogen is pumped into a fuel cell’s negative end (anode), it splits into electrons and protons. The electrons are carried through an external circuit where they can do useful work, arriving eventually at the positive end (cathode). The protons travel through the electrolyte to the cathode where they combine with the electrons and oxygen to form water as a byproduct, exiting the fuel cell. The only byproduct of a hydrogen fuel cell is unpolluted pure water, making this technology a source of clean energy.
“Through this simple electrochemical process we can directly convert the chemical energy in hydrogen to electricity, but we need catalysts to facilitate the reactions involved, particularly the sluggish oxygen reduction reaction at the cathode, which is a major obstacle to overcome,” Hwang said.
Catalysts speed up chemical reactions by lowering energy barriers that need to be overcome. Unfortunately, current electrocatalysts for hydrogen fuel cells are mostly made from platinum, a rare and costly element whose efficiency and stability have proven difficult to match.
Until now, that is.
Hwang’s research team carried out first-principles-based, atomistic simulations on the TACC’s Ranger and Lonestar supercomputers to design platinum-free alloy electrocatalysts that can outperform platinum-based catalysts and are less expensive and more abundant.
“It can be very time-consuming and cost-ineffective to find the right combination of alloying elements using only the trial-and-error experimental approach,” Hwang said. “That’s why we wanted to use the computational approach to provide guidelines for screening potential candidates.”
Hwang and his team focus on developing first-principles-based computational schemes capable of predicting the structure and catalytic activity of metal alloy nanoparticles. Their program couples various state-of-the-art computational methods including quantum chemistry, molecular mechanics and statistical theories.
“We first identify possible atomic configurations for chosen alloy nanoparticles and then see how chemical reactions occur on their surfaces using computer simulations,” Hwang said.
By mixing different elements, other researchers have identified alloy materials that can outperform platinum’s efficiency as a catalyst. In many cases, however, these alloy catalysts prove unstable and do not last long in their useful state.
For example, mixing platinum with cobalt works very well, but before long cobalt leaks out and the catalyst stops working.
“A significant number of multimetallic alloy nanoparticles with different atomic configurations are possible, and we need to screen what chemical composition, size, and shape of alloy nanoparticles work well for catalysis of the electrochemical reactions in fuel cells,” Hwang said.
Although Hwang’s team has been focusing mostly on binary alloys (two element mixtures), their recent simulations show that certain ternary alloys (three element mixtures) can be promising catalyst candidates for hydrogen fuel cells.
One successful design involves iridium (Ir) doped palladium (Pd)-cobalt (Co) catalysts, proving useful in speeding up the cathodic oxygen reduction reaction. By studying the factors involved in improving the electrochemical reaction, the team can refine their knowledge and predict combinations that might work even better.
Hwang’s team also found ways of mitigating activity loss in catalysts caused by carbon monoxide (CO) poisoning. This happens when CO molecules strongly bind to the active sites of catalysts, preventing them from doing useful work.
The researchers found that palladium (Pd)-gold (Au) alloys can boost CO oxidation to become more tolerant to CO poisoning. Alloys with small palladium ensembles like dimers and compact trimers provide more active sites for coadsorption of CO and O2, significantly reducing CO poisoning by facilitating the oxidation of adsorbed CO molecules. These findings were recently published in the February 2012 issue of The Journal of Physical Chemistry Letters.
The researchers share these promising designs with experimental collaborators who are able to synthesize the materials in a lab and confirm whether or not the computationally-designed catalysts actually work as expected outside of the simulations.
“If the computational and experimental programs are integrated in the right way, they can have a significant effect in solving some of the persistent problems, advance the field rapidly and enhance the commercialization feasibility,” said Amurugam Manthiram, director of the Texas Materials Institute at The University of Texas at Austin.
Like Hwang, Manthiram and his team focus on developing new materials for fuel cells and rechargeable batteries that are durable, low-cost and more efficient. However, Manthiram focuses on the experimental aspect of research, where the properties of new materials are tested beyond simulations. He believes computational work can help guide experimental programs and that the two aspects of the research are complementary.
Graeme Henkelman, an associate professor at the university’s Department of Chemistry and Biochemistry, has an ongoing project to understand nanoparticle catalysts through atomic scale simulations.
“Experimentalists can synthesize a particle, but you have to know exactly what the surface looks like and what the shape is to understand the chemical reactions that take place on the surface,” he said.
The problem is that simulations are simplified versions of reality.
“The underlying theory, quantum mechanics, is extremely hard to solve, so if you do it exactly you can only solve for maybe 10 electrons at most, and in a nanoparticle catalysts there can be thousands,” he said.
Although the difficulty of scaling complex calculations can be helped with highly-parallel computers like those at TACC, Henkelman believes it is important to complement simulations with experimental tools.
“The most exciting thing is trying to understand real experiments and using them to really test our theories,” he said.
With collaborations like these and the growing processing power of computers, the likelihood of creating a viable fuel cell system is greater than it has ever been. However, the era when our day-to-day technology depends on this energy source is still relatively distant.
“When hydrogen fuel cells become the main producers of clean energy our lives will change drastically, as we will run almost all our power with clean and efficient power sources,” Hwang said. “It may take another decade or more, but we still need to continue developing new knowledge and technologies that make that happen. We cannot just wait.”
*Diego Joaquin Cruz Ramirez, Science and Technology Writer