AI helps with material substitution
There are many reasons why companies have to replace individual raw materials or materials with others. The search for alternatives is time-consuming and often ends in failure. A research team from Fraunhofer IPA has now developed an AI-supported tool for material substitution.
Cobalt is used in lithium-ion batteries for electric vehicles and therefore plays an important role in the energy transition. However, the silver-grey metal is a critical raw material for several reasons: Cobalt is rare. It makes up just 0.004 percent of the earth's crust. The world's known cobalt reserves are estimated at 7.2 million tons. More than half of these, around four million tons, are located on the territory of the Democratic Republic of the Congo. Working conditions in the mines of this unstable Central African country are often poor and the environmental damage caused by ore mining is extensive.
Whether it's a lack of security of supply, excessively high prices on the global market, ethical concerns, bans or product innovations with better material properties: There are many reasons for companies to look for alternative materials. "There are databases that product developers can use for research. But they often don't provide any useful results because they don't take into account the exact application in the company," says Charlotte Schmidt from the Sustainability and Material Compliance Management research team at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA in Stuttgart.
AI sifts through scientific publications
To make the search easier and achieve more suitable results, Schmidt and two colleagues have developed an AI-supported tool for material substitution. Using an input screen, users first have to enter specific details about the material or raw material they want to replace and then specify the required properties of the alternative material as well as information about the context of the desired material use. This is followed by an AI search, which searches the "Semantic Scholar" database using the specific data and user requirements. By comparing the user input with the information available in the database, the AI identifies suitable alternative materials.
The AI connection for material substitution is just one of several components with which the researchers support companies in their search for alternative raw materials, materials or chemical substances. Once the AI has completed its task, it subjects the proposed substitutes and the source materials to a comprehensive assessment, taking into account legal, ecological and social aspects as well as security of supply. In close consultation with the company in question, the scientists then check how exactly the proposed materials meet the specific requirements. A report is produced at the end of the process. This presents the most suitable substitutes and the assessment of the various criteria. The scientists thus provide companies with a sound basis for decision-making.
Initial tests show: AI connection is promising
One of the alternatives to cobalt suggested by the AI-supported material substitution tool is iron. "It is not a new discovery that lithium iron phosphate can be used for the cathodes of batteries instead of lithium nickel manganese cobalt oxide," says Schmidt. "But this and other results have shown us that the AI connection is promising in the search for alternative materials."
About the research project
The AI-supported material substitution tool was developed as part of the "Ultra-Efficiency Factory - Deep Dive" research project, which was launched in April 2024 and will run until the end of August 2025. The research project is supported by the Baden-Württemberg Ministry for the Environment, Climate Protection and the Energy Sector with a total of 1.4 million euros. Other project partners from industry and science are also involved.
