For weeks, the whiteboard of the laboratory was crowded with scribbles, diagrams and chemical formulas. A research team through the Olivetti group and MIT concrete Sustainability Hub (CSHUB) worked intensely on a key problem: how can we reduce the amount of concrete cement to save on costs and emissions?
The question was certainly not new; Materials like flying ash, a by-product of coal production and slag, a sub-produced of steel, have long been used to replace part of the cement in concrete mixtures. However, demand for these products exceeds the supply while industry seeks to reduce its climatic impacts by expanding their use, which makes the search for urgent alternatives. The challenge that the team discovered was not a lack of candidates; The problem was that there was too much to sort.
On May 17, the team, managed by Postdoc Soroush Mahjoubi, published open access paper in nature Communication material describing their solution. “We realized that AI was the key to moving forward,” notes Mahjoubi. “There is so much data on potential materials – hundreds of thousands of pages of scientific literature. Trier would have taken many working lives, when other materials have been discovered! ”
With large language models, like chatbots that many of us use daily, the team has built an automatic learning framework that assess and sorts candidate materials according to their physical and chemical properties.
“First, there is hydraulic reactivity. The reason why concrete is strong is this cement – the “glue” which keeps it together – hardens when exposed to water. So, if we replace this glue, we must make sure that the substitute reacts in the same way, ”explains Mahjoubi. “Second, there is pozzolanicity. It is when a material reacts with calcium hydroxide, a by-product created when the cement meets water, to make concrete hard and stronger over time. We must balance the hydraulic and pozzolanic materials in the mixture so that the concrete works at best.”
Analyzing the scientific literature and more than a million rock samples, the team used the framework to sort the candidate materials in 19 types, ranging from biomass to mining by-products to demolished building materials. Mahjoubi and his team found that appropriate materials were available in the world – and, more impressive, many could be incorporated into concrete mixtures simply by grinding them. This means that it is possible to extract emissions and cost savings without too much additional treatment.
“Some of the most interesting materials that could replace part of the cement are ceramics,” notes Mahjoubi. “Old tiles, bricks, pottery – all these materials can have high responsiveness. This is something we observed in the old Roman concrete, where ceramics have been added to help the waterproof structures. I have had many interesting conversations on this subject with Profess Admir Masic, who is carrying out a lot of the old concrete studies here at MIT.
The potential of everyday materials such as ceramic and industrial materials such as mining residues is an example of how materials like concrete can help allow a circular economy. By identifying and reusing materials that would otherwise end in landfills, researchers and industry can help give these materials a second life as part of our buildings and our infrastructure.
For the future, the research team plans to upgrade the framework to be able to assess even more materials, while experimentally validating some of the best candidates. “AI's tools have put this research far in a short time, and we are delighted to see how the latest developments in large -language models allow the next steps,” explains Professor Elsa Olivetti, principal author of work and member of the MIT of MIT of the MIT of the MIT. She is director of the MIT CLIMATE project mission, CSHUB principal investigator and head of the Olivetti group.
“Concrete is the backbone of the built environment,” explains Randolph Kirchain, co-author and director of CSHUB. “By applying data science and AI tools to the design of materials, we hope to support industry efforts to build more permanently, without compromising strength, security or sustainability.
In addition to Mahjoubi, Olivetti and Kirchain, the co-authors of the works include the Postdoc Vineeth Venugopal MIT, Ipek Bensu Manav SM '21, Phd '24; And CSHUB assistant director Hessam Azarijafari.
