AI stirs up the recipe for concrete in MIT study
Cambridge, MA – In a groundbreaking stride towards sustainable infrastructure, researchers at the Massachusetts Institute of Technology (MIT) have unveiled an innovative AI-driven framework designed to revolutionize concrete production. This pioneering study, published on May 17 in Nature’s *Communications Materials*, addresses the urgent need to reduce the environmental impact and cost associated with traditional cement, the primary binding agent in concrete.
For decades, materials like fly ash and slag have served as partial cement substitutes, but their supply is now struggling to keep pace with the increasing demand for eco-friendly construction. This challenge left scientists with a monumental task: sifting through hundreds of thousands of pages of scientific literature and vast data on potential alternative materials. “We realized that AI was the key to moving forward,” explains postdoc Soroush Mahjoubi, who led the research team from the Olivetti Group and the MIT Concrete Sustainability Hub (CSHub). “There is so much data out there on potential materials — hundreds of thousands of pages of scientific literature. Sorting through them would have taken many lifetimes of work.”
Leveraging large language models, similar to the AI chatbots commonly used today, the MIT team developed a sophisticated machine-learning framework. This framework intelligently evaluates and categorizes candidate materials based on two crucial properties vital for concrete’s performance: hydraulic reactivity and pozzolanicity.
Mahjoubi elaborates: “First, there is hydraulic reactivity. The reason that concrete is strong is that cement — the ‘glue’ that holds it together — hardens when exposed to water. So, if we replace this glue, we need to make sure the substitute reacts similarly.” He continues, “Second, there is pozzolanicity. This is when a material reacts with calcium hydroxide, a byproduct created when cement meets water, to make the concrete harder and stronger over time. We need to balance the hydraulic and pozzolanic materials in the mix so the concrete performs at its best.”
Analyzing scientific literature alongside over a million rock samples, the AI framework successfully classified candidate materials into 19 distinct types. These ranged from biomass and mining byproducts to even demolished construction materials. The most remarkable discovery was that many of these suitable alternatives are globally available and can be seamlessly incorporated into concrete mixes with minimal processing, often just by grinding. This significantly reduces both emissions and costs associated with new material production.
A particularly intriguing finding highlighted by Mahjoubi was the potential of ceramics. “Some of the most interesting materials that could replace a portion of cement are ceramics,” he notes, citing “Old tiles, bricks, pottery — all these materials may have high reactivity.” This echoes observations in ancient Roman concrete, where ceramics were used for waterproofing, a topic Mahjoubi discussed with Professor Admir Masic, an expert in ancient concrete studies at MIT.
This innovative research embodies the principles of a circular economy, transforming materials that would typically end up in landfills into valuable components for our future infrastructure. By identifying and repurposing these “waste” materials, the study paves the way for a more sustainable and resource-efficient construction industry.
Looking ahead, the research team plans to enhance their AI framework to assess an even broader spectrum of materials, while simultaneously conducting experimental validations on the most promising candidates. “AI tools have gotten this research far in a short time, and we are excited to see how the latest developments in large language models enable the next steps,” states Professor Elsa Olivetti, senior author on the work and a leader in sustainable materials research at MIT.
Randolph Kirchain, co-author and CSHub director, underscores the broader impact: “Concrete is the backbone of the built environment. By applying data science and AI tools to material design, we hope to support industry efforts to build more sustainably, without compromising on strength, safety, or durability.”
The distinguished team also includes MIT postdoc Vineeth Venugopal, Ipek Bensu Manav SM ’21, PhD ’24, and CSHub Deputy Director Hessam AzariJafari. This groundbreaking work was generously supported by the MIT Concrete Sustainability Hub, funded by the Concrete Advancement Foundation, and received additional funding from the MIT-IBM Watson AI Lab.
