Liquid hydrogen vectors, AI finds the 41 best molecules among 160 billion

In search of the perfect liquid hydrogen vector

The most promising liquid hydrogen carriers? What in jargon are called LOCH, the acronym of “liquid organic hydrogen carriers“, ie liquid organic hydrogen transporters? These molecules offer an interesting alternative in the field of chemical storage compared to the use of ammonia, methanol or formic acid. Unlike the latter compounds, in fact, LOCH can chemically trap (hydrogenation) and then release (dehydrogenation) pure H2 molecules, through completely reversible cycles.

To find the best liquid vectors, a group of scientists from the Argonne National Laboratory screened 160 billion molecules. A number that exceeds that of people born throughout human history. It is therefore not surprising to know that the company has been helped by the new technologies of artificial intelligence. In detail, the team employed chemoinformatic methods – combining the use of chemical theory with computer and computational techniques – and an AI-based screening approach to accelerate calculations.

Liquid organic hydrogen carriers, the advantages of LOCH

LOHCs offer several advantages to the hydrogen industry. They can, in fact, have a high storage capacity and therefore a high energy density, making them suitable for direct use in fuel cells and other portable applications. Compared to pure hydrogen, they also have a higher safety profile which makes them an interesting option in the field of transport. Different classes of LOHC potentials are currently known, such as the benzene/cyclohexane system or the toluene/methylcyclohexane system.

But the main question remains only one: how is it possible to identify molecules that have the right physical and chemical properties? To answer the Argonne scientists first established a selection criterion based on 4 factors: structural similarity with the known liquid hydrogen vectors; precise melting and boiling points (the carrier must remain liquid when hydrogen is added or extracted); the ability to store a large amount of hydrogen per unit volume; a low energy demand for H2 release.

“We were looking for organic liquid molecules that retain hydrogen for a long time, but not so strong that they couldn’t be easily removed on demand,” explains researcher Hassan Harb and co-author of the study. “They must also have the capacity to store enough hydrogen for practical use“.

Supercomputer, quantum chemistry and AI

The group took action by accessing some chemical databases with data on organic molecules, finding over 160 billion molecule candidates. At this point the selection began, exploiting the supercomputer possessed at the Argonne and calculations of quantum chemistry. But even with these powerful resources, it would take scientists five years to analyze all 160 billion molecules. For this reason, the group has developed a unique AI-based screening approach that has reduced the computing time to just 14 hours, selecting 41 molecules of new liquid hydrogen transporters.

“With the help of artificial intelligence, we looked for organic liquid molecules that, through a low-cost chemical reaction with a catalyst, could alternatively add or release hydrogen to be used as fuel.”, said computational scientist Logan Ward. The work now passes into the hands of the experimenters who will have to test the promising ones. The research was published on Digital Discovery.