Kemira Uses AI to Develop Materials for PFAS Removal from Water
Kemira and CuspAI are developing materials designed to capture per- and polyfluoroalkyl substances, known as PFAS, at trace concentrations in water. The collaboration was featured in an October 5 interview published by BusinessGreen, which highlighted the use of AI to navigate an exceptionally large materials design space.
According to Kemira’s May announcement, the project explored approximately 300 trillion possible structures, generated more than 5,000 material designs and narrowed these to around 20 priority candidates in six months. The work targeted three PFAS compounds: GenX, PFBS and PFOS. These figures describe the discovery process, rather than thousands of materials manufactured and tested in treatment plants.
The candidates are metal-organic frameworks, or MOFs, porous crystalline materials whose structures can be tailored for adsorption. Kemira specified requirements including stability in water, manufacturability, environmental compatibility and cost. The partners are progressing to further development and testing, with the aim of finding more selective, longer-lasting alternatives to existing adsorbents.
For prospective users, the distinction between discovery and deployment matters. A promising computer-generated structure provides a starting point for experiments. Investment decisions require evidence that the resulting material can be produced consistently and deliver reliable treatment under the conditions encountered at a particular facility.
Why Existing Treatment Presents Challenges
PFAS are persistent synthetic chemicals used in applications including water-resistant products, non-stick materials and firefighting foams. Removing them requires treatment suited to their chemical properties. The U.S. Environmental Protection Agency (EPA) identifies activated carbon, ion exchange resins and high-pressure membranes among the technologies capable of reducing PFAS in drinking water.
Activated carbon captures contaminants on its extensive internal surface. However, performance depends on factors including the PFAS present, water composition, flow rate and filter configuration. The EPA notes that shorter-chain substances such as PFBS do not adsorb as effectively as longer-chain compounds such as PFOS. Membrane systems can remove a broad range of PFAS, but generate a concentrated waste stream requiring management.
These limitations explain the practical interest in selective materials. For a utility comparing treatment options, a useful evaluation would ask how much compliant water a material can produce before replacement, which compounds pass through first and how performance changes as source water quality varies.
Removal must also be distinguished from destruction. Capturing a pollutant creates a separate question about its subsequent handling. Any assessment of a new adsorbent should therefore cover the spent material, potential regeneration and the treatment or disposal of resulting residues.
European Rules Increase the Need for Reliable Treatment
The development comes as European drinking water requirements move into implementation. The European Commission announced that harmonized PFAS monitoring requirements under the recast Drinking Water Directive began applying on January 12, 2026. Member States must report monitoring results, including exceedances, incidents and granted derogations.
Where limits are exceeded, authorities must take action to reduce PFAS levels and protect public health, while informing the public. The Commission identifies possible responses including closing contaminated wells, introducing additional treatment and restricting the use of affected drinking water supplies.
For water suppliers, these obligations create an immediate operational consideration: compliance planning cannot depend on the future availability of a material still undergoing development. Emerging technologies can inform longer-term procurement and pilot programs, while current treatment decisions need to rest on demonstrated performance and applicable requirements.
What Utilities and Industrial Users Should Assess
The next meaningful evidence for prospective customers would include comparative testing against established treatment media. Relevant questions include removal at realistic concentrations, performance in mixed-contaminant water, operating life and whether any material components enter the treated water.
Manufacturing deserves equal attention. A candidate that performs well in small experiments still needs a practical production route, dependable quality control and a form suitable for treatment equipment. Procurement assessments should also examine supply continuity and the availability of technical support.
For sustainability teams, potential environmental benefits require a broader comparison than removal efficiency alone. A longer operating life could reduce replacement frequency, but a credible assessment would also consider manufacturing inputs, pumping energy, transport, regeneration and final waste management. These are evaluation criteria, rather than benefits established by the discovery announcement.
The project’s significance is therefore its use of AI to accelerate the search for candidate materials. Whether that progress translates into a competitive water treatment product will depend on experimental results and operating evidence. For utilities and industrial users, the decisive outcome would be dependable PFAS control at an acceptable total cost and environmental burden.
Sources:
Cut through the green tape
We don't push agendas. At Net Zero Compare, we cut through the hype and fear to deliver the straightforward facts you need for making informed decisions on green products and services. Whether motivated by compliance, customer demands, or a real passion for the environment, you’re welcome here. We provide reliable information. Why you seek it is not our concern.