Meta and Georgia Tech use AI to advance carbon capture solutions

Tech giant Meta has teamed up with the Georgia Institute of Technology to create a massive open dataset to advance AI solutions for carbon capture, a technology with great potential to help solve the problem of global warming.

Excessive carbon emissions are the main cause of climate change, and we’re already at a place where simply cutting them is not enough – although we should still do that.

Capturing carbon dioxide directly from the air or industries and recycling it is another idea that can sound like a win-win climate solution. This way, greenhouse gases stay out of the atmosphere, allowing us to slow down the current warming that’s beginning to hit dangerous levels.

The challenge, though, is that for direct air capture technology, every sort of environment and location requires a uniquely specific design. The configuration of the technology in Florida would surely be different than one in Finland, for example.

The systems have to be designed with exact parameters for temperature, humidity, and air flows for each location. Such efforts to maximize efficiency take precious time and cost a lot of money – unsurprisingly, most current carbon-capture projects do not pay off.

Now, however, Georgia Tech and Meta say they have collaborated to produce a massive database, potentially making it easier and faster to design and implement direct air capture technologies.

“The open-source database enabled the team to train an AI model that is orders of magnitude faster than existing chemistry simulations. The project, named OpenDAC, could accelerate climate solutions the planet desperately needs,” said the university in a press release.

The team’s research was published in ACS Central Science, a journal of the American Chemical Society.

“For direct air capture, there are many ideas about how best to take advantage of the air flows and temperature swings of a given environment,” said Andrew J. Medford, associate professor in the School of Chemical and Biomolecular Engineering (ChBE) and a lead author of the paper. “But a major problem is finding a material that can capture carbon efficiently under each environment’s specific conditions.”

Their idea was to “create a database and a set of tools to help engineers broadly, who need to find the right material that can work,” Medford said. “We wanted to use computing to take them from not knowing where to start to giving them a robust list of materials to synthesize and try.”

It helped that researchers with Meta’s Fundamental AI Research (FAIR) team were looking for ways to harness their machine-learning prowess to address climate change. They landed on direct air capture as a promising technology and went straight to Georgia Tech.

Anuroop Sriram, research engineering lead at FAIR and first author on the paper, generated the database by running quantum chemistry computations on the inputs provided by the Georgia Tech team.

These calculations used about 400 million CPU hours, which is hundreds of times more computing than the average academic computing lab can do in a year. FAIR also trained machine learning models on the database.

It needs to be said, though, that carbon-capture solutions are not universally championed. Sure, in the best cases, the process leaves less carbon dioxide in the environment, and the captured carbon is, for instance, used to produce construction materials such as concrete.

In other words, carbon is captured and sealed, and a product that has economic value is created.

But in worse cases, capturing carbon can be harmful if it eventually increases the amount of excess carbon dioxide in the environment. A good example is using the captured carbon for enhanced oil recovery.

Finally, even if all that is taken into account, the fact is that no carbon-capture technology works at 100% efficiency, and some carbon dioxide will always escape into the air.