Applying science and AI to the challenge of carbon capture

In a keynote talk at the congress, Chen Xi, Chair Professor and Dean of Lingnan’s School of Interdisciplinary Studies, spoke about one of the most important issues now facing all of us. His topic was “Addressing Climate Change: Negative Emission and AI-driven Evolution of Advanced Energy Systems” which allowed him to highlight the extent of the challenge and the need for groundbreaking solutions that apply the latest developments in energy materials and technologies to reduce carbon dioxide and greenhouse gas emissions. 

Some of these advances have the potential to promote significantly greater efficiency in energy conversion and storage capacity. Others are more focused on closing the “carbon loop” through direct air capture (DAC) of carbon dioxide or by finding ways to engineer alternative tech-based possibilities.

“It is only the CO2 in the air that can affect our climate,” said Chen, who holds a PhD in solid mechanics from Harvard University and spent 20 years as a professor in the Department of Earth and Environmental Engineering at Columbia University before joining Lingnan last year. “The Earth has the capability of absorbing CO2 naturally through forests, the oceans, and the soil. However, to rely on Mother Nature to absorb the CO2 we are now producing will take at least tens of thousands of years.”

He highlighted how dramatically things had changed since the pre-industrial age, with emissions from transport, chemical and energy plants contributing to a doubling of the PPM (parts per million) in the air in the past 200 or so years. 

Current efforts to achieve carbon neutrality depend on reducing emissions as much as possible by means of innovations, promoting clean energy, and implementing energy-saving initiatives. But that will not be enough. It is also essential to use DAC and methods described as carbon capture utilisation and storage (CCUS). And while statistics show there was some slowdown in total world emissions after the Kyoto agreement, it had to be a major cause for concern that some countries are now increasing their use of fossil fuels, with the result that recent progress is now being eroded. 

“CO2 concentration and global warming are going to be more severe than we have seen in the last ten years,” said Chen, who is known for using a combination of theoretical, experimental and numerical approaches when addressing real-world challenges in areas ranging from energy and the environment to biology and nanotechnology. “The world is starting to see a non-linear trend again. Right now, we are at 420 PPM, but I would say the first critical, quantitative red line is 450 PPM. At that level, the ocean is going to become so acidic that you will see a grand-scale ‘melting’ or dissolution of the coral reefs, which will be a disaster for the ecological chain in the oceans. Many scientists predict that may be the beginning of the sixth mass extinction event on Earth.” 

With the aim of avoiding that, each “good citizen” should emit a net figure of no more than 0.3 tonnes of CO2 per year. At present, though, only two countries – Cambodia and Somalia – are below that target, while the United States is at the annual equivalent of 17.5 tonnes per capita. A switch to clean energy alone cannot solve the problem, while governments in China, India, Africa and elsewhere continue to prioritise economic growth, much of it coming from sectors with high energy consumption. 

“To create carbon balance, we have to find a way to use CO2 and have more incentives for business to do that,” Chen said. 

One possibility, through photosynthesis, is to improve the agricultural yield of farming. Another is using carbon to make cement and other building materials stronger. However, such methods will only slow the pace of increase. The major objective is still to take carbon out of the air through DAC with negative emission technology. 

A so-called “moisture swing” material has been developed to do just that, and a factory in Xian, the first of its kind in Asia, is already in operation. Each of the devices produced can capture up to two tonnes of CO2 annually at very low cost. 

“We have also developed the first moveable CO2 capture devices for flue gas emissions from cars and boilers,” Chen said. “The concentration there is more than 200 times higher than in the air, so we use different techniques. With these efforts, we have established China’s first carbon-negative industrial park, and we turn all the CO2 captured into various types of products, for example building materials. Every brick we fabricate sinks about 200-300 grams of CO2 permanently stored as carbonates, which makes them stronger and cheaper overall with lower lifetime emissions.” 

To help push things forward, AI is also coming into play. In this respect, though, scientists have to start from scratch because ChatGPT does not yet have the specific knowledge in particular areas to update and optimise carbon capture performance. 

This information must first be encoded in long sets of text, making it possible to establish generative AI platforms which can then suggest next-generation materials or, for instance, enhanced use of crystals or polymers. 

“For direct air capture, you have to teach your AI platform about the chemistry, such as entropy, free energy and activation energy, and then the skills to calculate all those energies and how to design materials based on nanostructure, surface modification, or how they react with various polymer chains,” Chen said. “But we have already ended up with several AI suggestions and have fabricated these carbon-capture materials through the first evolution.”