Magnetic nanoparticles capture microplastics from water

To tackle the growing ecological threat of microplastics, magnetic nanoparticles have been created that can remove plastic fragments from water. Researchers used machine learning to identify the ideal removal conditions for particular microplastics, a strategy that may help to optimize other clean-up methods^[1].

Microplastics are scraps of waste plastic, typically 1 micrometer to 1 millimeter in size, which are now ubiquitous in the environment. The particles adsorb toxic metals and organic pollutants. They are easily ingested by aquatic life, and once microplastics and their toxic payloads are in the food chain, they can accumulate in other species, including humans.

Methods to remove microplastics from wastewater face various drawbacks. Using light to destroy microplastics is effective but expensive and energy-intensive. Certain microbes can break down microplastics, but this generates other molecules that may themselves be toxic.

Magnetic nanoparticles offer a simple, low-cost and environmentally friendly solution. But these nanoparticles are prone to oxidation and may clump together in water, reducing their effectiveness, explains Rifan Hardian of KAUST’s Physical Science and Engineering Division.

A collaboration between KAUST and University Malaysia Terengganu has now developed magnetic nanoparticles that overcome these problems and remove not only microplastics but also the organic pollutants they carry.

First, the researchers prepared iron oxide nanoparticles with a protective porous silica coating. Then, they added linker molecules to the silica, which allowed them to decorate the particles with more molecules called imines. This covered the nanoparticle with molecular strands that capture microplastics, while chemical groups in the imines bind organic pollutants.

The researchers tested these nanoparticles on different sizes of polystyrene microplastics carrying common pollutants. They used a magnet to pull the nanoparticles from the water, along with their microplastic cargo and then washed off the fragments so the nanoparticles could be reused.

But with so many different variables to consider — from the number and size of microplastics in the water, to the concentration of imines on the magnetic nanoparticles — the researchers needed an efficient way to identify which factors offered the best clean-up solution.

So they used a method called ‘design of experiments’ to determine which combinations of variables would produce the most useful data and then fed those data into a machine-learning system that quickly identified any trends. This highlighted ways to maximize microplastic removal while minimizing the amount of imine required.

For example, a relatively low concentration of imines could remove 90 percent of a sample of microplastics roughly 300 micrometers in size. With fewer, smaller microplastics, an extremely low concentration of imines could still achieve about 80 percent removal efficiency.

“With our machine-learning model, we can predict the efficiency of the microplastic removal at various numbers and sizes,” says Hardian. “This prediction can then be used to determine the required adsorbent concentration.”

“The next milestone will be applying our methodology in other microplastic removal technologies so we can compare the efficiency of available technologies,” says Gyorgy Szekely, who led the KAUST researchers. They also hope to develop an autonomous laboratory system that could continuously optimize the experimental conditions.

Reference

1. Rushdi, I. W., Hardian, R., Rusidi, R. S., Khairul, W. M., Hamzah, S., Khalik, W. M. A. W. M., Abdullah, N. S., Yahaya, N. K. E. M., Szekely, G. and Azmi, A. A. Microplastic and organic pollutant removal using imine-functionalized mesoporous magnetic silica nanoparticles enhanced by machine learning. Chemical Engineering Journal 510, 161595 (2025).| article

---

---