Can Microplastics Be Removed From Water? What Three Recent Claims Actually Show

Two stories have been circulating: a specialised powder that pulls microplastics out of water with a magnet, and a plant-based plastic that releases enzymes as it degrades, breaking down microplastic already in the soil and water around it. Both sound like the beginning of a cleanup.

They sit at very different points on the evidence scale, and the gap between them is a useful thing to be able to see. One is a peer-reviewed result with a DOI and a candid list of limitations. The others are prototypes that have won prizes and press coverage but have not published a number anyone can check.

Here is what each actually shows, and what none of them changes.

First, the baseline nobody mentions

A headline saying a new method removes 96% of microplastics only means something if you know what the current figure is. It is higher than most people assume.

A systematic review of wastewater treatment plant performance found microplastic removal efficiencies ranging from 58% to 98.7%, with most plants achieving greater than 90% (Cristaldi et al., 2020, International Journal of Environmental Research and Public Health). The work is done in the unglamorous early stages: primary treatment removed roughly 78–98%, secondary a further 7–20%.

The same review found something counterintuitive that tends to get lost in summaries of it. Tertiary treatment — the advanced stage you would expect to be the most thorough — "does not seem to have significant effects on reducing the concentration of microplastics." The heavy lifting happens early, through settling and skimming, not through sophisticated polishing at the end.

And the authors are careful to close the loop most coverage leaves open: despite those high percentages, "millions of microplastics continue to be released every day into the aquatic environment." A 95% removal rate applied to a colossal input still leaves a large absolute number. Percentages flatter; totals do not.

That is the context any new removal technology has to be judged against.

The magnetic powder: a real, published result

This one clears the bar.

Researchers at Durban University of Technology in South Africa developed magnetic nanocomposites — sheets of carbon and boron nitride coated with iron nanoparticles — that bind microplastic particles and are then pulled back out of the water with a magnet, taking the plastic with them (study in Journal of Environmental Chemical Engineering; the researchers, Riona Indhur and Sheena Kumari, describe the work themselves in The Conversation).

The reported results:

• 96% removal of polyethylene and 92% of polystyrene from purified water
• ~94% from drinking water
• Up to 92% from treated wastewater
• Still around 80% effective after three reuses

That last figure matters more than the headline one. A material that survives reuse is a material that could plausibly scale; one that works brilliantly once is a chemistry demonstration.

The limitations are stated plainly, which is a good sign about the work rather than a mark against it. The powders performed well on small microplastics measured in hundreds of micrometres, and substantially worse on larger fragments in the 3–5mm range. Extraction used a 250kg magnetic force. The authors describe the study as early work and say the next steps are scaling it up, testing across more varied water conditions, and designing affordable devices.

That last point is the one to hold onto. The cost works out at roughly US$41 per 1,000 litres — about four cents a litre, which is in the same range as replacement pitcher cartridges, so cost is not the obstacle here. The obstacle is the apparatus. Nothing about a process requiring a quarter-tonne of magnetic force is heading for a kitchen countertop. This is municipal water treatment research, and it should be read as such.

The prototypes: genuinely interesting, not yet evidence

Two student projects have drawn wide coverage this year.

Eco Purge, developed by 18-year-old Arya Satheesh in Ireland, won the European category of the 2026 Earth Prize. It is a plant-based biodegradable polymer with enzymes embedded in its structure; as the material degrades over about two weeks, the enzymes are released and — the claim goes — break down microplastic already present in the surrounding soil or water. Satheesh tested it by placing microplastics alongside the material in soil, salt water and fresh water, and observing over six weeks. She worked with researchers at University College Dublin, Atlantic Technological University, and the BiOrbic Bioeconomy Research Centre, and received $12,500 to develop the prototype further.

Separately, Mia Heller, a high school student in Virginia, built a ferrofluid-based filtration system presented at the Regeneron International Science and Engineering Fair, reporting roughly 96% microplastic removal while reclaiming about 87% of the ferrofluid.

Both are real work by capable people, and neither deserves the reflexive dismissal that "science fair project" tends to invite — Satheesh's collaboration with three research institutions is more than most prototypes have behind them.

But neither has been published in a peer-reviewed journal, and no measured removal rates have been released that anyone outside can examine. What exists is a prize, a prototype, and a description. That is a reasonable place for a promising idea to be. It is not a place from which conclusions can be drawn, and the coverage has consistently outrun what has been demonstrated — particularly for Eco Purge, where "releases enzymes that break down microplastics in the surrounding environment" is a considerable claim about how enzymes behave once they leave a controlled setting.

The honest position is that these are worth watching and not yet worth believing.

Why "removes 96%" is slipperier than it sounds

Across all three, the same caveats apply to any removal percentage:

• Spiked laboratory water is not tap water. Researchers add a known quantity of specific polymers at controlled sizes. Real water carries a messier mix of shapes, polymers, ages and surface coatings.
• Removal rates are size-dependent. The Durban powder's own results demonstrate this — strong on small particles, weak on large ones. A single percentage figure conceals which particles it applies to.
• Most testing ignores nanoplastics entirely, the size class below one micrometre that is hardest to detect and, on current evidence, potentially the most biologically relevant.
• "Removed" is not "destroyed." The magnetic powder concentrates plastic onto a material that then has to be dealt with — the same accounting problem as boiling water, which traps particles in limescale rather than eliminating them.

What this changes for you

Nothing, for now — and that is a real answer rather than a disappointing one.

Every technology here operates at the wrong scale for a household. The powder is a treatment-plant process. The enzyme plastic, if it works as described, is a materials substitution that would play out over years across supply chains. Neither is something to wait for before acting.

The things that do work at home have not changed, and are covered in our room-by-room reduction guide: a filter certified to NSF/ANSI 401 or a reverse osmosis system for drinking water, and an awareness that carbon-only pitchers can add particles rather than remove them. Our water filter guide covers the certified options.

What is genuinely worth taking from the research is the point buried in that systematic review: treatment already removes most microplastic, and the volume reaching the environment is still enormous. That is an argument about how much plastic enters the system in the first place, not about better filters at the end of it — which is why source reduction keeps turning out to be the part that matters.

This article describes research findings and general information, not personalized health advice.