鈥淚n Sweden, we are a little spoiled to have clean water, but there are also problems with for example PFAS contamination. Take Ronneby for instance, where fire foam has contaminated the environment and led to long-term consequences. But PFAS is a global problem 鈥� it鈥檚 everywhere. And there鈥檚 more and more evidence that it has a negative impact on our health,鈥� says Johanna Ros茅n, Professor of Materials Science at LiU.
PFAS is a large group of compounds with more than 10,000 identified substances. They have varying properties and are used in everything from fire-fighting foam and cosmetics to food packaging and textile impregnation, to name a few.
Forever chemicals
The problem with PFAS substances is that they are extremely difficult to break down and are therefore also known as 鈥渇orever chemicals鈥�. In addition, they are water-soluble and move easily through the ground, which affects the groundwater. Some PFAS substances can be very harmful to humans, animals and nature, and there are currently no efficient means to render PFAS harmless.
Anna Nilsen
To capture and break down harmful chemicals, the research team intends to use so-called low-dimensional materials, i.e. extremely thin structures, only one or two atoms thick, which have a very large active surface that can react with the pollutants. In addition, their properties can be tailored at the atomic level.
鈥淏reaking down the long PFAS molecules is about being able to break the carbon-fluorine bond, which is very strong. What often happens is that the molecule is split, but what remains can be at least as dangerous as the original contamination.鈥�
Help from AI
According to Johanna Ros茅n, an important part of the project is to be able to analyse how to break down PFAS and above all understand what it will become in the end.
鈥淲e may be able to break down PFAS, but if we only create a shorter and equally dangerous molecule, we haven鈥檛 solved the problem,鈥� says Johanna Ros茅n.
The research group already has preliminary results for some 2D materials with promising properties for purifying water. To find more materials, the researchers have taken the first steps towards an AI-based language model trained on over 5,000 scientific articles in the field. It should be able to identify knowledge gaps and promising strategies that are otherwise difficult to detect.
Jenny Wid茅n
鈥淢y PhD supervisor used to say, 鈥榓 week in the library saves you half a year in the lab鈥� and that鈥檚 a message I鈥檝e carried with me and passed on. The problem is that we don鈥檛 have time to read all the literature today; there鈥檚 a huge amount of data. This is where language models are something that can help us quickly find the information most relevant to the project,鈥� says Johanna Ros茅n.
Interdisciplinary research group
For this reason, Professor Marco Kuhlmann at LiU, who researches and develops large-scale language models, is part of the project group. Anneli Kruve, professor at Stockholm University, is also a co-applicant and adds expertise in the analysis of PFAS and other contaminants. Expertise in material analysis is also a must, and this is where Professor Per Persson at LiU will contribute advanced electron microscopy.
鈥淚t鈥檚 so much fun to start working in new collaborations! We will also bring in international experts who can further accelerate the research process.鈥�
A total of 30 research projects were awarded grants from KAW, and will share SEK 900 million. The projects are mainly focused on basic research in medicine, science and technology considered to have the potential to lead to future scientific breakthroughs. In addition to Johanna Ros茅n, Senior Associate Professor Eleni Stavrinidou also received grants from KAW.
