Microplastics in bottled water: what the studies actually counted

Search for microplastics in bottled water and you will meet two numbers. One says a litre holds about 325 plastic particles. The other says about 240,000. They are often quoted side by side as if one study had simply been redone more carefully.

That is not quite what happened. The two studies used different tools and looked for particles of very different sizes. Read together, they say less about which brand to avoid and more about how young this kind of measurement still is.

The 2018 study: 259 bottles, 11 brands

The first study that put bottled water microplastics in the headlines was published in Frontiers in Chemistry in 2018 by Sherri Mason and colleagues. They bought 259 bottles of 11 brands in 19 locations across nine countries and stained the water with a dye called Nile Red, which sticks to plastic and glows under the right light.^[1]

The brands included several that are on Waterphile: Evian, S.Pellegrino, Gerolsteiner, Aquafina and Dasani, alongside large national brands from India, Indonesia, Mexico, Brazil and China.^[1]

What they reported:^[1]

• 93% of the bottles showed some sign of microplastic contamination.
• After subtracting possible lab contamination, the average was 10.4 particles larger than 100 micrometres per litre.
• Counting smaller particles too, down to 6.5 micrometres, the average rose to 325 particles per litre.
• Individual results ranged from zero to over 10,000 particles per litre.

Only about half of the larger particles could be confirmed as plastic with infrared spectroscopy, and the most common polymer among those was polypropylene, at 54%. The authors point out that polypropylene is a common material for bottle caps, and conclude that the contamination is at least partly coming from the packaging or the bottling process.^[1] For the particles under 100 micrometres, the study relied on the dye alone; the authors describe them as "most probably plastic".^[1]

One detail is easy to miss. One of the 11 lots was sold in glass: Gerolsteiner, in 750 mL bottles. Every bottle in the study, glass or not, had a plastic cap.^[1] So the study was not set up to compare glass with plastic, and it should not be read that way.

The 2024 study: down to the nanoscale

In January 2024 a Columbia University team published a very different kind of measurement in PNAS. Instead of a dye, they used a laser imaging method called stimulated Raman scattering, paired with an algorithm that identifies the chemical signature of seven common plastics, one particle at a time. It could pick out particles below 100 nanometres.^[2]

They tested three popular brands sold in the US and did not name them.^[3] Their estimate was about 240,000 plastic particles per litre on average, and roughly 90% of those were nanoplastics, meaning smaller than one micrometre.^[2] Columbia's announcement gives the spread as 110,000 to 370,000 particles per litre.^[3]

The NIH summary of the work adds two useful details. The most common plastic was polyamide, a type of nylon that is often used in filters that purify water. PET, the plastic most water bottles are made of, was also abundant. The method also flagged millions of other particles that did not match any of the seven plastics, and it is not yet clear what those are.^[4]

So the jump from hundreds to hundreds of thousands is mostly a jump in size. The 2018 study stopped at 6.5 micrometres; the 2024 study went more than 60 times smaller. NIH describes the new total as 10 to 100 times more plastic particles than earlier studies, which mostly focused on larger microplastics.^[4]

The argument about blanks

Counting particles this small is hard because plastic is everywhere, including in labs. Later in 2024, PNAS published a letter from Dušan Materić arguing that nanoplastic measurements must use appropriate blank samples to rule out contamination introduced during the analysis.^[5] The Columbia authors replied in the same journal, defending how they had set up their blanks.^[6]

Exchanges like this are a normal part of science and do not mean either study is wrong. They do mean the 240,000 figure is best treated as an early estimate from three unnamed brands, not as a settled property of bottled water.

What regulators say

The World Health Organization reviewed the evidence on microplastics in drinking water, bottled and tap, in 2019. Its key messages:^[7]

• Data on microplastics in drinking water were limited, with few fully reliable studies.
• Chemicals and microbial films associated with microplastics posed a low concern for human health, based on the limited evidence available.
• For the particles themselves, there was not enough information to draw firm conclusions, particularly for nanoplastics, but no reliable information suggested a concern.
• Routine monitoring of microplastics in drinking water was not necessary at that time, and more well-designed research was needed.

The US Food and Drug Administration's current page on microplastics in food says that the scientific evidence does not demonstrate that the levels detected in foods pose a risk to human health. It also says there is not sufficient evidence to show that particles from plastic packaging migrate into foods and beverages.^[8] That second statement sits a little awkwardly next to the 2018 study's bottle-cap finding, which is one more sign of how unsettled the field is.

Neither body has set a limit. The FDA's identity and quality standard for bottled water sets allowable levels for minerals, metals and other contaminants, and does not mention plastic particles.^[9] California has gone furthest for drinking water: acting on a 2018 state law, its State Water Resources Control Board adopted a definition of microplastics in drinking water in 2020, followed by a standard method for testing and reporting them.^[10]

Why Waterphile does not show a microplastics number

Waterphile's data comes from each brand's own water quality report or label. Those reports are built around the regulated list of minerals and contaminants, and we have not found a microplastics count in any of them. Without a number from the brand, or from a study that names brands and tests them the same way, we would have to guess. We do not guess.

What the reports do show is where a water starts and how it is treated. The four brands from the 2018 study that are on Waterphile make the point:

Water	TDS mg/L	Calcium mg/L	Magnesium mg/L	Sodium mg/L	Score
Evian	345	80	26	6.5	9.8
S.Pellegrino	901	164	46.4	29.9	6.3
Aquafina	ND	—	—	—	—
Dasani	—	—	—	—	—

Evian and S.Pellegrino are natural mineral waters bottled from their own sources in the French Alps and in San Pellegrino Terme, Italy.^[11][12] Aquafina and Dasani are purified waters. Their makers' reports describe public or municipal water supplies treated with steps that include reverse osmosis.^[13][14] Our guide to mineral, spring and purified water explains the legal difference. None of that tells you how many plastic particles are in a given bottle. It does tell you what you are paying for.

The short version

• Both headline numbers come from real, peer-reviewed studies. They differ mainly because the 2024 study counted far smaller particles.
• The 2018 study found contamination in 93% of bottles, and its polymer findings point partly at caps and bottling.
• The 2024 study named no brands, so no one can use it to rank them.
• WHO and the FDA say current evidence does not show a health risk, and both say more research is needed.
• No bottled water standard sets a limit, and the brand reports we use do not report a count.

If a brand publishes microplastics testing, or a study tests named brands with a method that holds up, we will add it to the brand pages and update this post.