Glass bottles: most plastic came at capping

What they found

In 2021, a Technical University of Munich team followed mineral water through four German bottling plants, from the well to the sealed glass bottle. Well water held under 1 plastic particle per litre, and washed bottles were below the detection limit. Capped bottles averaged 317 per litre. Most of that plastic resembled the caps' inner seal; the authors suspect it wore off at capping or opening.

They counted particles 11 µm (thousandths of a millimetre) and larger, identified by infrared imaging. At one plant, three filled bottles taken just before capping held 6 per litre, below the detection limit. Capped bottles held 75 to 700 per litre, and 81% of it was polyethylene (a common soft plastic), like the cap seals. EFSA, the EU food agency, calls this key to how plastic rubs off, and notes the counts are low next to studies reporting thousands. A French agency lab (ANSES), counting 30 µm and up, found water in glass bottles held more particles than water in plastic ones (4.5 vs 1.6 per litre), and in glass-bottled soft drinks and beers most particles matched the caps' paint.

Evidence: Emerging

Why it matters

Paying for bottled water doesn't reliably buy fewer plastic particles. Counting down to 1 µm, a French lab found 19 to 1,154 particles per litre, 98% of them under 20 µm, in ten brands of water in plastic bottles (one sample each). One tap sample from Toulouse, tested the same way, held 413 per litre, more than 8 of the 10 brands.

Evidence: Weak

The swap

Filter your tap water instead

Bottled water isn't reliably cleaner than tap, and filtered tap costs less per litre. If you filter, pick one with a tested particle claim (NSF/ANSI 401 microplastics reduction or NSF/ANSI 42 Class I particulate), a good sign rather than proof. A jug with only carbon and resin did poorly in a lab test. Already filtering? No need to change. Where tap water isn't safe to drink, or a boil-water notice applies, keep using bottled or boiled water as local advice says.

Cost: ~$30 (jug) to $100+ (under-sink) · Swap rating: Likely

What we don't know yet

None of these studies measured water as poured at home, and each counted different sizes, so numbers can't be compared across studies. The German plants used glass bottles with metal caps, only one had filled bottles tested before capping, and the 317 wasn't blank-corrected (blanks averaged 13 per sample). Unfiltered tap isn't shown to have fewer particles, and filter tests on plastic are few.