Teen innovators use machine learning and chemistry to combat PFAS pollution

When two students noticed that PFAS pollution wasn't getting enough attention in the UK, they decided to do something about it. Christopher Whitfeld and Wenqi (Jonathan) Zhao combined their interests in machine learning and natural sciences to create an affordable solution to detect and treat PFAS. After winning the 2024 Stockholm Junior Water Prize, their mission to protect water is just beginning.

Christopher and Jonathan, now in their final year at Eton College, have been close friends and collaborators for several years. Their shared passion for scientific research and environmental issues led them to investigate PFAS pollutants in water.

PFAS chemicals, often linked to cancer and health issues, are hard to remove from the environment and pose a long-term threat. Christopher and Jonathan’s project, focusing on the Thames Basin in the UK, aimed to provide a practical, scalable solution for predicting PFAS levels in the environment and reducing PFAS concentrations in tap water.

They created a geospatial neural network that could predict PFAS concentrations within 90% accuracy of real-world data. They also designed a simple, cost-effective filtration device to reduce PFAS concentrations by up to 93%. This device could be installed on taps, providing clean, safe drinking water to communities affected by PFAS contamination.

Making Waves recently spoke to Christopher and Jonathan to learn more about their project and their future plans.

Q: What inspired you to work on water pollution and PFAS?

Christopher: I’ve spent so much time around water, doing sports like rowing and sailing, that it’s become a really central part of my identity. In the UK, sewage spills are often in the news, and I wondered how I could combine my love for water with scientific research to address some of the pollution issues we face.

Jonathan: Water issues connect to so many different fields. I’m more into biology and chemistry, while Christopher is into computer science. Water pollution, particularly PFAS, felt like an issue where we could combine our interests to make a positive change. We also wanted to improve awareness of PFAS pollution, which is a challenge in the UK.

Christopher Whitfeld and Wenqi (Jonathan) Zhao

Q: Can you explain what PFAS are and why they’re dangerous?

Jonathan: PFAS, which stands for per- and polyfluoroalkyl substances, are called “forever chemicals” because they don’t break down easily in the environment.

Even just four nanograms of PFAS, which you could compare to a couple of drops in an Olympic-sized swimming pool, are enough to be harmful, causing cancer, immune system problems, and other health issues.

But the problem right now is that people don’t know where PFAS might be because it’s so expensive to test and the concentrations are so low.

Christopher: The UK doesn’t have the same regulations for PFAS as the US, which has a limit of four nanograms per liter in water.  Such a limit, however, is imperative to prevent these harmful, carcinogenic effects. PFAS also aren’t widely reported on in the UK.  So PFAS are a problem both because they are harmful and because not enough people know about them.

Q: How did you make a geospatial neural network for your project?

Christopher: A geospatial neural network is a machine learning algorithm that utilizes geographical data. We took existing data from several sources and added that to our model. This includes data on PFAS sampling in different locations, soil composition, industrial activity, and water quality.

Based on this data, we were able to predict PFAS levels in different parts of the Thames Basin with about 90% accuracy.

Q: How do you see machine learning and natural sciences working together?

Jonathan: When you combine natural sciences with computational sciences, you can tackle problems from two different, complementing angles.

In our project, we used machine learning to predict PFAS levels and then validated those predictions with chemical experiments.

The natural science approach is generally more reliable, but it can be costly and time-consuming, while machine learning can process large amounts of data quickly and affordably.

Christopher: Computers have the power to effectively complete so many more tasks than humans. By using the strengths of both fields, you can save time and get more accurate results, which is key for complex environmental problems like PFAS pollution.

Q: How does your filtration system work, and who is it designed for?

Jonathan: The main goal of our project is to develop a PFAS solution that is equitable, so anyone can afford it. The problem right now is that effective filters, like reverse osmosis and nanofiltration kits, are very expensive for the average person.

Our filtration device, which uses activated carbon to filter out 93% of PFAS from drinking water, costs about £6 to produce.

The idea is that people can check PFAS levels in their area using our model and decide whether they need the filter. It's a very easy point-of-use filter that anyone can install onto a tap, letting safe water pass through.

Q: What’s next for your project?

Christopher: The main obstacle right now is awareness. PFAS pollution is not known widely enough for our solution to be effective. One of the things we are focusing on is outreach, like in youth magazines.

Young people are genuinely concerned about the environment, and if we can educate them about the problem, that information gets passed up to their parents at the dinner table and can proliferate towards those in power.

Our long-term goal is to create enough awareness to get new PFAS regulations in place in the UK. We would also like to improve our model and filter to be more user-friendly.

Photo: Jonas Borg

Q: Why do you think the Stockholm Junior Water Prize is important?

Christopher: I think it’s about enfranchisement. There are loads of people who are interested in environmental causes but don’t really have a clear path to act on it.

The Stockholm Junior Water Prize, with Xylem’s support, builds a bridge for young people to get involved in water research and actually make a difference. It’s so important to provide that pathway, and we saw that firsthand in Stockholm.

Jonathan: I think the Stockholm Junior Water Prize serves two functions: it exposes young people to environmental issues in school. We had a day visit to Xylem’s offices and laboratory, and they talked about opportunities for young people to get involved. The other benefit is that the Stockholm Junior Water Prize can connect students with external resources, like mentors, who can really catalyze your project and make it into something more meaningful.

Q: Is there anything else you would like to add?

Christopher: On the stage in Stockholm, we were a bit shocked and didn’t get a chance to thank everyone properly. At our school, I’d like to thank Dr. Day and Mrs. Stock, who were integral in allowing us to journey around London collecting water samples. Also, our science teachers, Ms. Knowles, Mr. Copsey, and Dr. Edmonds, who helped us with in-school research. Ms. Herbommez, our environmental coordinator, has been fantastic over the years, encouraging us to engage in environmental causes. Finally, my parents – they’ve been incredible in so many ways. They’ve prepared me my entire life for this, driving my interest in current affairs, research and science, and instilling a work ethic to follow through on a project.

Jonathan: This project was done entirely outside of classes and lessons at school. Those teachers who could have just stopped when the lesson ended, they continued to support us and facilitate experiments, that was a main reason why this project was possible. I’d also like to thank my parents. They live in China, and during the award ceremony in Stockholm it was 2 a.m. for them, but they were still watching. That was quite an incredible feeling.