Kemira and CuspAI have used generative AI to design new PFAS removal materials targeting drinking and process water at trace concentrations.
The collaboration focused on the design of novel materials that could remove selected PFAS, often described as “forever chemicals”, from water using chemistry intended to be stable, sustainable and manufacturable.
According to Kemira, the discovery project explored approximately 300 trillion possible material structures and delivered more than 5,000 novel material designs with full property data for three priority PFAS molecules: GenX, PFBS and PFOS.
AI-designed PFAS removal materials advance to testing
The AI-designed PFAS removal materials have now been narrowed to around 20 priority candidates, which are moving into further development and testing.
Kemira and CuspAI said the project reached this stage in six months, compressing a discovery process that can typically take years.
The programme used CuspAI’s generative materials platform to design metal-organic frameworks, or MOFs, from scratch against industrial performance requirements defined by Kemira.
The target was to identify materials capable of removing specific PFAS molecules from water at sub-parts-per-billion concentrations while remaining water-stable, environmentally compatible, synthesizable and cost-effective.
PFAS treatment remains a major drinking water challenge
PFAS are persistent synthetic chemicals used across a wide range of industrial and consumer applications. They are difficult to break down in the environment and have become a growing concern for drinking water providers, regulators and communities.
Tightening standards, including US Environmental Protection Agency limits announced in 2024 and requirements under the EU Drinking Water Directive, are increasing pressure to develop effective treatment and remediation technologies.
Granular activated carbon is one of the most widely used PFAS treatment approaches today. Kemira said its collaboration with CuspAI was designed to explore whether AI-driven materials discovery could support more selective and longer-lasting alternatives.
Generative AI used to search vast material design space
The design space for metal-organic frameworks is extremely large, with possible candidate structures estimated in the hundreds of trillions.
By using generative AI, CuspAI’s platform searched a material design space of approximately 300 trillion possible structures and produced thousands of candidate designs for further assessment.
The project also identified new functional group chemistries that could have potential for broader adsorption product development.
Dr Chad Edwards, CEO and Co-Founder of CuspAI, said the collaboration demonstrated how AI can compress discovery timelines for environmental challenges.
Sampo Lahtinen, Executive Vice President, Research and Innovation at Kemira, said the project combined Kemira’s water treatment chemistry expertise with CuspAI’s materials design capabilities, with candidates evaluated against industrial requirements.
Antti Salminen, President and CEO of Kemira, said the companies now have “a credible path toward a next-generation PFAS remediation product”.
The project is now moving into further development and testing, while additional programmes across other material classes are being scoped under the partnership’s framework agreement.
Read more PFAS and water treatment news in the Water Quality and Water Treatment sections of H2O Global News.
FAQs
What are AI-designed PFAS removal materials?
AI-designed PFAS removal materials are materials created or optimised using artificial intelligence to capture and remove PFAS chemicals from water.
What PFAS molecules did Kemira and CuspAI target?
The project focused on three priority PFAS molecules: GenX, PFBS and PFOS.
What are metal-organic frameworks?
Metal-organic frameworks, or MOFs, are porous crystalline materials that can be designed with specific structures and chemical properties for filtration, adsorption and separation applications.
Why is PFAS removal difficult?
PFAS chemicals are highly persistent and can be present in water at very low concentrations, making them challenging to detect, capture and remove effectively.
