An engineering team has built a machine-learning framework to evaluate and sort candidate materials for cleaner concrete, based on physical and chemical properties.
Led by civil and environmental engineering researcher, Soroush Mahjoubi at MIT, the team found some of the “most interesting materials that could replace a portion of cement are ceramics”.
The question the Olivetti Group/MIT Concrete Sustainability Hub team was working on was: How can we reduce the amount of cement in concrete to save on costs and emissions? However, the challenge wasn’t a lack of candidate materials – they said there were too many to sort through.
So, the team turned to artificial intelligence for help.
“There is so much data out there on potential materials — hundreds of thousands of pages of scientific literature,” Mahjoubi said. “Sorting through them would have taken many lifetimes of work, by which time more materials would have been discovered!”
Analysing scientific literature and over one million rock samples, the team used the framework to sort candidate materials into 19 types, ranging from biomass to mining byproducts and demolished construction materials. These materials were available globally and the team found that many could be incorporated into concrete mixes just by grinding them – offering cost savings without much additional processing.
Some of the most interesting materials were ceramics – old tiles, bricks, pottery – with the potential for high reactivity, something the team had observed in ancient Roman concrete where ceramics were added to help waterproof structures.
The potential for ceramics and industrial materials like mine tailings is an example of how materials like concrete can help enable a circular economy, the team said. Their research was published in anopen-access paper in Communications Materials in May.

