AI Helps In Search For ‘Green Bullet Catalyst’ In Hydrogen Fuel Production
Using supercomputers, a research team from Trinity College Dublin (TCD) has taken a key step toward the production of non-polluting hydrogen fuel cheaply, efficiently and in industrial quantities.
Hydrogen fuel has long been seen an alternative and major non-polluting fuel source that can be especially useful in some of the worst polluting sectors such as transportation and industry. Producing hydrogen fuel in commercial quantities, however, is complicated and expensive.
Hydrogen fuel is produced by splitting water into oxygen and energy-rich hydrogen, which is then collected and stored in fuel cells. The only problem is this process demands a huge amount of electricity. As a consequence, hydrogen fuel has remained commercially unviable. This drawback has left natural gas, which is far more polluting, as the favorite alternative to replace crude oil.
TCD researchers pointed out the basic problem facing them, and every other team that has had a go at this problem, is that water is very stable. Water needs a great deal of energy to split into hydrogen and oxygen.
Also, a tough barrier that has to be mastered is the energy, or "overpotential," associated with the production of oxygen. This overpotential is the bottleneck hindering the splitting water to produce hydrogen.
In an electrolytic cell, the existence of overpotential means the cell requires more energy than thermodynamically expected to drive a reaction such as electrolysis. The extra energy produced is lost as heat.
For hydrogen fuel to become a zero-emission fuel, its production needs to be powered by renewable electricity close to wind farms or solar farms. Production will only become feasible on an industrial scale if affordable catalysts can be found, which limit the overpotential of oxygen in the production process.
Combining powerful supercomputers and leading chemistry research, TCD researchers have taken a major step towards attaining the holy grail of catalysis. As a result, the dreaded overpotential problem now seems easier to overcome. Researchers have also made it easier to search for the elusive "green bullet" catalyst.
"We know what we need to optimize now, so it is just a case of finding the right combinations," Michael Craig from TCD, lead author of the study published in Nature Communications, said.
The team will now use AI (artificial intelligence) to put a large number of earth-abundant metals and ligands (which glue them together to generate the catalysts) in a melting pot before assessing which of the near-infinite combinations yield the greatest promise. The team has also established fundamental principles for the design of ideal catalysts.
"Given the increasingly pressing need to find green energy solutions, it's no surprise scientists have been hunting for a magical catalyst that will allow them to split water electrochemically in a cost-effective, reliable way," Prof. Max García-Melchor, who was part of the team, said.
"However, it is no exaggeration to say that before now such a hunt was akin to looking for a needle in a haystack. We are not over the finishing line yet, but we have significantly reduced the size of the haystack and we are convinced that artificial intelligence will help us hoover up plenty of the remaining hay."
He also added this project shows what can happen when researchers from different disciplines work together to apply their expertise to try to solve a problem that affects each and every one of us.