by Bijan Jafari, Eurofins Environment Testing

Most public conversation about microplastics still centers on the environment: sediment cores, surface trawls, and the slow accumulation of fragments in places far from daily life. That framing isn't wrong, but it's incomplete. Plastic packaging, plastic liners, and plastic containers sit in direct, sustained contact with the food and drink people consume every day, and that contact is a plausible route of exposure in its own right. Over four small studies, we tested that idea against real products: berries in a plastic clamshell, milk in two different container types, two brands of yogurt, and a plastic water bottle put through ordinary handling, to see what a consistent analytical method would actually turn up.

Each sample went through the same core workflow: oxidative digestion and density separation to free particles from the surrounding matrix, vacuum filtration onto a silicon membrane, and Raman spectroscopy to confirm polymer identity and size particles between 20 and 500 micrometers. That range matters as much as the method itself: across all four studies, only particles at or above 20 micrometers were counted, and anything smaller simply falls outside what this method can detect. That consistency in method and size range means the differences described below are differences between products, not artifacts of switching methods midway through.

A caveat worth stating plainly before the results: these are case studies, not surveys. Sample sizes are small, one or two products per comparison, single runs in most cases, and the goal was to demonstrate what this kind of testing can reveal, not to characterize an entire product category. Read the numbers as data points, not verdicts.

Fresh produce picks up more than it's given credit for

Raspberries and strawberries are sold almost universally in rigid plastic clamshells, a packaging format so routine it rarely invites scrutiny. We rinsed roughly 40 grams of each fruit in microplastics analysis grade water, then processed the tissue itself through digestion and density separation to recover any embedded particles.

The raspberries yielded 16 particles, split mainly between polycarbonate and polyethylene. The strawberries yielded nearly twice as many, 30 particles, with polyethylene as the largest single fraction, followed by PET, polycarbonate, and polypropylene. A method blank processed alongside the samples returned 7 particles, which gives a sense of the background against which these counts should be read. In both fruits, the majority of particles fell between 20 and 80 micrometers, well below the threshold of visibility and well within the range that would pass through a rinse under a kitchen faucet.

We can't attribute these particles to the clamshell specifically without further work isolating the packaging from other possible sources, such as agricultural handling or processing equipment. But the presence of polymers matching common packaging plastics, at this concentration, in fruit tissue that never left its original container until testing, is a reasonable basis for asking the question further.

 

Two milk containers, similar totals, different chemistry

Milk packaging varies more than most shoppers register: HDPE jugs, polyethylene-lined paperboard cartons, occasionally glass. We compared an HDPE jug against a PE-lined carton, testing only the first pour from each to isolate whatever the container contributes on its own, before any consumer handling such as shaking, repeated opening, or storage enters the picture.

The two formats produced comparable particle counts: 22 in the jug, 26 in the carton. On magnitude alone, the packaging choice looks close to irrelevant. The polymer breakdown says otherwise. The carton sample was strongly polyethylene-dominant, consistent with its PE liner. The jug showed a more heterogeneous profile, with polycarbonate slightly ahead of polyethylene and a minor PET contribution, pointing to a broader set of contact surfaces contributing particles, from the bottle wall to the cap.

The lesson here is a methodological one as much as a practical one: particle count alone can obscure real differences between products. Two containers releasing similar amounts of microplastic may still be releasing meaningfully different material. 

Two yogurts in the same category, a threefold gap between them

If the milk comparison suggested that packaging format mainly reshuffles composition, the yogurt comparison complicated that picture. We processed the entire contents of two single-serve cups, no subsampling, through the same workflow.

The results diverged sharply. One brand contained 74 particles, 81 percent of them polyethylene. The other contained 27, with a more even split across polyethylene, polypropylene, and PET. That's close to a threefold difference in total particle count between two products occupying the same shelf category, sold in broadly similar single-serve packaging. Particle size was the one dimension that held steady: both brands clustered between 20 and 70 micrometers, with a handful of larger outliers above 150 micrometers in each.

Because both magnitude and composition differed here, packaging format alone doesn't explain the gap. Formulation, processing equipment, or supply chain differences upstream of the retail package are all plausible contributors, and distinguishing between them would require a study designed for that purpose. What this comparison does establish is that product category is a poor predictor of microplastic load. "Yogurt" is not a single answer.

 

What a water bottle releases depends entirely on what you do to it

The first three studies asked what's already present in a product at the point of purchase. The fourth asked what happens during ordinary use. Bottled water is squeezed, reopened, shaken, left in cars and gym bags more than almost anything else people buy, so we built a small pilot around that reality.

We filtered water at 5 micrometers, well below the 20-micrometer threshold used throughout this analysis, to establish a clean starting point. That water was returned to its original bottle, which itself required one round of opening and closing the cap, then a designated stress was applied, and the water was filtered again with a fresh filter, so any particles recovered afterward could be traced to the bottle or cap itself rather than to the water supply. Five conditions were tested: hand-pressing the bottle ten times, opening and closing the cap ten additional times, shaking it a hundred times to approximate a workout, two days of direct sunlight exposure outdoors, and two days in a hot car, with both outdoor exposures occurring in ambient temperatures in the upper 90s°F.

 

The mechanical stresses produced the clearest signal. Hand-pressing raised particle counts from 14 to 203. Shaking had the largest effect of any condition tested, raising polyethylene counts from 29 to 224.

The capping condition is worth walking through in more detail. Before the ten designated open-close cycles were applied, the bottle already showed a PE count of roughly 137. We can't say with confidence where that came from. It's worth noting that every bottle in this study needed at least one open-close cycle just to have the pre-filtered water returned to it, yet the other conditions didn't show a comparable PE level at their own "before" readings, so the single refill cycle alone doesn't obviously explain it. What we can say is what happened next: after the ten additional open-close cycles, PE stayed at roughly that same level, while PET rose from about 7 to 17 over those same cycles. So the designated capping stress corresponded with a measurable increase in PET, but not in PE.

Heat and sunlight behaved differently, and the difference is worth sitting with. Rather than increasing detectable particle counts, both conditions saw them fall: heat from 16 to 7, sunlight exposure from 217 to 128. That drop almost certainly does not mean fewer particles were generated. It more likely means particles were fragmented below our 20-micrometer detection limit, moving out of the size range this method can see rather than out of existence. If that interpretation holds, environmental exposure may be a more significant driver of microplastic generation than these numbers alone suggest. We simply lack the resolution here to confirm it.

This was, again, a pilot: one bottle brand, one run per condition. Baseline particle counts also varied noticeably from one bottle to the next, which is worth flagging on its own. Bottle position wasn't tracked going into the study, but these bottles came from a wrapped multi-pack, and it's reasonable to expect that bottles sitting at the edges or corners of that pack absorbed more mechanical stress in transport than bottles protected in the middle. That kind of unrecorded handling history is a plausible explanation for why the "before" readings differ as much as they do across conditions, and it's a variable a larger follow-up study would want to control for. The pattern within each condition, before versus after, is still worth taking seriously. It just shouldn't be treated as settled.

Reading the four studies together

No single factor explains microplastic contamination across all four products. Packaging format changed composition without changing magnitude in the milk comparison. Something upstream of packaging, whether formulation, processing, or supply chain, appears to drive both magnitude and composition in the yogurt comparison. And in the water bottle study, the product itself barely mattered; what changed the outcome was how it was handled after purchase.

That variability is, in a sense, the actual finding. It argues against treating microplastic exposure as a single, uniform risk tied to one culprit, whether a particular plastic or a particular packaging type, and toward treating it as a property of specific products, specific formats, and specific handling conditions, each of which needs to be measured rather than assumed. Environmental matrices like water and soil remain important, but this data suggests they're an incomplete picture of where exposure actually happens. The refrigerator and the pantry deserve the same scrutiny.

The Author:

Bijan, an expert in microplastics analysis, leads the Microplastics Department at Eurofins Environment Testing, where he and his team pioneered the use of Raman spectroscopy to develop commercial-scale microplastics analysis. With years of experience in environmental testing and analytical method development, he has worked alongside his team to advance microplastic contamination research, ensuring high-quality, regulatory-compliant data for scientific and industrial applications. Under his leadership, Eurofins has expanded its capabilities, supporting private clients, government agencies and academics to further the field and support the industry by improving turn-around-times and expanding testing options.