Are there microplastics in your whisky? What the evidence says about nanoplastics, blending water and the bottling line

Key takeaways

• No one has counted microplastics in whisky. The nearest published data are beer at a mean of 4.05 particles per litre (Kosuth et al., PLOS ONE, 2018) and packaged beverages at around 100 particles per litre in glass bottles (ANSES, 2025). Anyone quoting you a figure for spirits is extrapolating.
• The purification step is itself a documented source. In the study that found ~2.4 × 10⁵ particles per litre in bottled water, roughly 90% of them nanoplastics, polyamide — not the bottle’s PET — was the major contributor by number, and the authors note that “PA is the most popular membrane material used in reverse osmosis” (Qian et al., PNAS, 2024).
• Closures matter more than the container. ANSES traced the contamination in glass-bottled drinks to paint on the metal caps, and reduced it from 287 to 87 particles per litre simply by blowing and rinsing the caps before sealing (ANSES).
• A “not detected” result usually means “not looked for”. The EU’s official method covers 20 µm to 5 mm (Decision (EU) 2024/1441), while 98% of the particles found in one bottled-water survey were smaller than 20 µm (Hagelskjær et al., PLOS Water, 2025).
• Filtration media can shed as well as capture. A point-of-use trial found one carbon-and-ion-exchange device whose effluent carried more plastic fragments than its influent, while a 0.2 µm membrane device removed 90–100% of PVC fragments (Cherian et al., Polymers, 2023).
• The health position is genuinely unsettled, and WHO’s is the moderate one: risk assessed as low on limited evidence, with no routine monitoring recommended (WHO, 2019).

There is no published microplastic count for whisky, gin or rum, and any supplier who offers you one is guessing. What the evidence does support is narrower and more actionable: in packaged drinks, most plastic particles enter after the liquid is made — from purification media, process equipment and closures — which means a spirit’s exposure is concentrated in exactly two places, the water used to reduce it to bottling strength and the bottling line it passes through. Distillation leaves particles behind in the still; everything downstream of the spirit safe is where the question actually lives.

Are microplastics really found in drinks, and in what quantities?

Yes, consistently, though the numbers vary by two orders of magnitude depending on what the method can see.

The most cited figures come from bottled water. Using stimulated Raman scattering microscopy to reach particles as small as 100 nm, a Columbia team reported “about 2.4 ± 1.3 × 10⁵ particles per liter of bottled water”, of which roughly 90% were nanoplastics — between ten and a hundred times more than earlier microplastic-only surveys had found (Qian et al., PNAS, 2024). A separate French survey of ten bottled brands and one tap sample using Raman microspectroscopy down to 1 µm found 19–1,154 microplastics per litre in bottled water (mean 306) and 413 per litre in Toulouse tap water (Hagelskjær et al., PLOS Water, 2025).

For alcoholic drinks, the data are thinner and older. Kosuth and colleagues found anthropogenic particles in beer at a mean of 4.05 per litre after blank correction, with a range of 0 to 14.3, 98.3% of them fibres (PLOS ONE, 2018). In 2025 the French food safety agency ANSES reported around 100 microplastics per litre in cola, lemonade, iced tea and beer sold in glass bottles — five to fifty times more than the same categories in plastic bottles or cans — while water sat far lower at 4.5 per litre in glass and 1.6 in plastic, and wine lower still (ANSES).

Note what that ordering does to the intuitive story. The glass bottle, the format the trade treats as the premium and inert choice, was the worst performer. The plastic bottle was not the problem.

Where do the particles actually come from?

Three sources dominate the published work, and none of them is the ingredient.

Purification media. This is the finding that should interest a spirits producer most, because it inverts the assumption that more treatment means fewer particles. In the PNAS bottled-water study, polyamide rather than PET was “the common major contributor in number among all the three brands”, and the authors’ explanation is direct: “PA is the most popular membrane material used in reverse osmosis, which is a common water purification method shared by all three brands.” They add that polypropylene and polyamide “are widely used as equipment components or coagulant aids in water treatment” (Qian et al., 2024). By mass the picture flips and PET dominates — a reminder that particle counts and particle mass answer different questions.

Closures. ANSES matched the colour and composition of particles in glass-bottled drinks to the paint on the metal caps, and found microscopic scratches on caps caused by friction in storage. Cleaning the caps before capping cut contamination from 287 particles per litre to 106 with an air blow, and to 87 with an air blow plus a rinse (ANSES). That is a two-thirds reduction from a procedural change with no capital cost.

Process contact and the ambient environment. The overwhelming majority of particles in the beer and tap-water survey were fibres (Kosuth et al., 2018) — the signature of airborne and textile shedding rather than of packaging. In a bottling hall that means open tank hatches, hoses, gaskets, filter sheets and clothing.

For a distillery the structural conclusion is the same one that applies to PFAS, reached by a different mechanism. Particles do not survive the still — they stay in the pot. Whatever is in the bottle arrived afterwards, in the reduction water, in storage and transfer, or on the filling line.

Does water treatment remove them — or add them?

Both, depending on the barrier.

Physical barriers work when the pore size is small enough. In a controlled trial of three point-of-use devices, the unit combining microfiltration at 0.2 µm with carbon and ion exchange removed 90–100% of PVC fragments and 55–94% of PET fragments. The unit with a nominal pore size of 1 µm or greater managed 46–78% and 54–86%. And the device with no membrane at all — granular activated carbon plus ion exchange — produced effluent containing more PVC and PET fragments than the water going in, at several points in its service life (Cherian et al., Polymers, 2023).

That last result is the one to carry into a supplier conversation. A carbon bed is a bed of friable granules with plastic components around it; under flow it can release more than it captures. “We filter it” is not an answer to this question unless the filter is a membrane and someone is watching the membrane.

Reverse osmosis is a genuine barrier by pore size — it rejects dissolved ions, so it certainly rejects particles — but the PNAS finding shows the membrane and its housing are also candidate sources of polyamide downstream. Reviews of membrane processes for micro- and nanoplastic removal report efficiencies spanning roughly 37% to over 99% depending on configuration, which is less an endorsement than a warning that the number depends entirely on the installation. The practical answer is the same one that applies to everything RO does not quite finish: a membrane followed by polishing and a defined final filtration, with the final barrier the last thing the water touches. See what reverse osmosis does not remove.

What do the regulations actually require in 2026?

Less than most buyers assume, and the gap is worth understanding before anyone signs a specification promising compliance.

The EU placed microplastics on the watch list of the Drinking Water Directive (2020/2184) rather than setting a parametric value, and the Commission adopted the measurement methodology in Delegated Decision (EU) 2024/1441 of 11 March 2024. That method covers particles from 20 µm to 5 mm and fibres from 20 µm to 15 mm, requires that “a minimum volume of 1 000 (thousand) litres of water shall be sampled”, and reports results as objects per cubic metre (EUR-Lex). It is a monitoring framework, not a limit.

India’s IS 10500 : 2012 sets no value for microplastics (IS 10500 : 2012), consistent with the broader position that no major drinking-water standard yet carries one — see water quality standards for spirits in India. WHO, reviewing the evidence in 2019, concluded that the risk was low, that particles above 150 µm are unlikely to be absorbed, and explicitly did not recommend routine regulatory monitoring, while calling for more research (WHO).

The 20 µm floor is where the regulatory picture collides with the science. In the French bottled-water survey, 98% of all detected microplastics were smaller than 20 µm and 94% smaller than 10 µm; the authors argue the Directive’s detection limit misses precisely the fraction of “greatest risk to human health” and press for the 1–20 µm range to be brought into regulatory guidelines (Hagelskjær et al., 2025). A certificate reporting “no microplastics detected” against the EU method is therefore compatible with hundreds of particles per litre below the method’s floor. The number that matters on any such report is not the result — it is the size limit of detection.

So how worried should a producer or a drinker be?

Honestly: less than the headlines imply, and more than the current paperwork reflects.

The toxicological case is unresolved. WHO’s 2019 assessment remains the cautious mainstream position — low risk on limited evidence — and it has not been overturned, though the discovery that nanoplastics outnumber microplastics by roughly an order of magnitude changes the exposure arithmetic that assessment was built on, since particles small enough to cross biological membranes are the ones that assessment could not count.

For a spirits producer the commercial reality is simpler than the toxicology. No regulator currently requires a microplastic result on a water certificate, and no laboratory can honestly certify a liquid “microplastic free” — the claim is not analytically supportable at the sizes that dominate. What is supportable is provenance of the barriers: knowing what the last filter was, what it is made of, what its rated pore size is, and what touched the liquid after it.

That is the same discipline as the rest of a blending-water specification, applied to a parameter nobody has written a limit for yet. It also happens to be free. Cleaning a cap before you seal a bottle cost ANSES’s laboratory nothing and removed two-thirds of the particles.

What should a buyer actually ask for?

1. The final barrier, in writing. Rated pore size, medium, and its position in the train. A 0.2 µm membrane as the last step is worth more than an elaborate train that ends in a carbon polisher (Cherian et al., 2023).
2. The limit of detection on any microplastic result. Without it the result is uninterpretable, and against the EU method’s 20 µm floor it is nearly uninformative (Hagelskjær et al., 2025).
3. Closure and cap-handling practice for any bottled water arriving at your site, and for your own filling line. This is the largest documented contributor in glass-packed drinks (ANSES).
4. Contact materials downstream of the final filter — hoses, gaskets, liners, filter sheets. Particles cannot be removed by a barrier they never pass through.
5. Membrane change-out and integrity records. An ageing polyamide membrane is both a weaker barrier and a candidate source (Qian et al., 2024).
6. Restraint in your own marketing. “Microplastic free” is not a defensible claim. “Final filtration at 0.2 µm, membrane records available” is.

Glossary

• Microplastic — a plastic particle between 1 µm and 5 mm; the EU’s official drinking-water method counts only from 20 µm upward.
• Nanoplastic — a plastic particle below 1 µm. Small enough to cross biological membranes, and invisible to most published survey methods.
• Polyamide (PA) — nylon. The dominant membrane material in reverse osmosis, and the most abundant polymer by particle count in the PNAS bottled-water study.
• Stimulated Raman scattering (SRS) microscopy — an imaging technique able to identify individual particles down to about 100 nm, used to produce the first nanoplastic counts in bottled water.
• Watch list — under the EU Drinking Water Directive, a substance monitored and reported on without a binding parametric value.
• Limit of detection (LOD) — the smallest particle a method can see. For microplastics it determines the result more than the sample does.
• Point-of-use (POU) device — a final-stage filter at the tap or the tank outlet, as distinct from central treatment.

Frequently asked questions

Are there microplastics in whisky? Nobody has published a count. The nearest measured categories are beer at a mean of 4.05 particles per litre (Kosuth et al., 2018) and glass-bottled soft drinks and beer at around 100 per litre (ANSES, 2025). Distillation leaves solid particles behind, so anything present in a finished spirit entered after the still — principally with the reduction water, the process equipment or the closure.

Does distilling the spirit remove microplastics? It removes them from the distillate, for the same reason it removes minerals: particles are not volatile and stay in the pot. That is precisely why the post-distillation steps carry the whole exposure.

Is demineralised blending water lower in microplastics than tap water? By barrier logic it should be — it has passed membranes that tap water has not. But the PNAS finding that polyamide from RO membranes was the most abundant polymer by count in bottled water (Qian et al., 2024) shows treatment is not automatically subtractive. What decides the outcome is the final barrier and what touches the water after it.

Can a supplier certify water as microplastic free? No. The EU’s own method sees nothing below 20 µm (Decision (EU) 2024/1441), and 98% of the particles found in one bottled-water survey were smaller than that (Hagelskjær et al., 2025). A “not detected” result describes the method’s blind spot, not the water.

Does glass packaging solve the problem? It did not in the only comparative study of packaged beverages: glass-bottled drinks carried five to fifty times more microplastics than the same drinks in plastic bottles or cans, traced to paint on the metal caps (ANSES).

Is any of this a health risk? WHO’s 2019 assessment concluded the risk from microplastics in drinking water was low on the available evidence, noted that particles above 150 µm are unlikely to be absorbed, and did not recommend routine monitoring (WHO). That assessment predates the nanoplastic counts, which is why the honest answer remains unsettled rather than reassuring.

What is the single cheapest thing a bottling hall can do? Clean closures before capping. Blowing caps with air took ANSES’s test from 287 to 106 particles per litre; adding a rinse took it to 87 (ANSES).

Blenders International · Water Science series