What happens when food and drinks are heated in plastic

Heating food or drinks in plastic is easy to overlook. It feels normal. It feels harmless. But heat changes how plastic behaves, and when plastic is in direct contact with what you eat or drink, that quiet chemistry deserves a closer look.

This is not just a microwave story. Microwaves get the headlines, but in most kitchens plastic meets heat in far quieter ways: the lid on a takeout container, the lining or components inside an electric kettle, a baby bottle warmed in hot water, a paper cup of tea, a scratched nonstick pan, or a water bottle left in a sun-warmed car. Much of the literature on heat and plastic food contact has emerged only since 2020 ^[1,2,84], and the field is still working through how best to measure what it is finding.

The goal here is to slow down. To look carefully at what the science says, what it does not yet say, and where the methodological questions are still unsettled.

Why heat matters: what changes inside the plastic

Plastic is not inert. It is a constructed material made of long, tangled polymer chains, the repeating carbon backbones that define each plastic, along with a mix of plasticizers, antioxidants, stabilizers, and other additives. At room temperature, those structures remain relatively stable. Under heat, they become more mobile, more reactive, and in some cases more fragile.

One of the first changes is softening. Every polymer has a temperature range at which its chains begin to loosen and slide more easily past one another. For some food contact plastics, including polylactic acid (PLA) used in certain compostable cups, that threshold can be approached by something as ordinary as hot tea ^[3-5]. Once a surface softens, routine contact begins to matter more. A spoon stirring, a lid snapping into place, a cap twisting shut, motions that would seem trivial at room temperature, can become enough to release surface particles ^[6,7].

Heat also works more slowly and chemically. In the presence of water, it can accelerate hydrolysis, gradually chipping away at ester and amide bonds in polymers such as PET, PLA, and nylon ^[8,9]. At the same time, higher temperatures can increase oxidative reactions at the surface, damaging carbon bonds and generating smaller oxidized fragments ^[10]. These are not dramatic transformations in the everyday sense, but they are part of how materials age, weaken, and shed over time.

Just as important, heat can mobilize the smaller molecules embedded within plastic itself. These include compounds such as phthalate plasticizers, antimony residues, cyclic siloxanes, triclosan, and photoinitiators from printing inks ^[11-14]. Unlike particles, these substances can migrate into food or drink without any visible flaking or breakdown. In many ways, this is the more established side of the heat and plastic story: long before microplastic release became a major focus, researchers were already studying how additives move out of packaging and into what we consume.

What emerges from the literature is that heat rarely acts alone. It changes the material, but often in ways that make other forces matter more. The European Food Safety Authority’s 2025 review of 122 release studies concluded that abrasion and mechanical stress, more than heat by itself, are the primary drivers of particle release from food contact materials ^[15]. Heat softens the surface. Friction, pressure, twisting, scraping, and repeated use often complete the process. As with many things in chemistry, heat increases what can move. In plastics, that can mean more migration into food and drink.

That is why the question is not simply whether plastic gets hot. It is what kind of plastic, at what temperature, in contact with what food or drink, and under what kind of repeated stress. Heat matters because it shifts the material into a more vulnerable state. From there, ordinary use can do more than it seems.

The microwave specifically

The microwave is usually the first place people look, and the honest answer requires some care.

The most widely cited headline number comes from a 2023 paper by Kazi Albab Hussain and colleagues at the University of Nebraska–Lincoln. In that study, polypropylene baby food containers and reusable food pouches were microwaved in water for three minutes. The authors reported release levels of roughly 4.22 million microplastic particles and 2.11 billion nanoplastic particles per square centimeter of container surface ^[16]. They also exposed kidney cells to those particles at 1,000 micrograms per milliliter and observed a roughly 77 percent drop in cell viability over 48 to 72 hours.

Those are striking numbers, and they spread quickly. They have also drawn an unusually open methodological exchange. A correspondence by Sun and colleagues argued that the analytical approach may have overcounted nanoparticles by not fully separating true polymer fragments from organic colloids and oligomer aggregates ^[17]. The Hussain group published a thorough rebuttal in defense of its method ^[18]. Around the same time, the European Food Safety Authority concluded that, across the broader literature, “the actual release is much lower than the results presented in many publications,” and flagged misidentification and miscounting as recurring issues ^[15]. None of that erases the underlying signal. It frames it. The headline number sits at the high end of a range the field is still learning how to measure, not a finding to wave off.

A more grounded microwave literature exists beyond the most extreme counts. A 2025 meta analysis pooling 237 observations across 30 studies found a clear relationship between temperature and release across PE, PET, PP, and PS containers, but with absolute counts that varied by several orders of magnitude depending on the polymer and the measurement method ^[19,20]. A 2024 study reported that microwaving released fewer particles than simple hot water immersion in the same containers, a result that runs against standard intuition ^[21]. A survey of takeout containers from five Chinese cities found that expanded polystyrene shed about 148 times more particles than polypropylene under hot conditions ^[22]. A 2024 Food Chemistry study reported no detectable microplastics from polypropylene containers heated to 100 °C, while PS and PET did shed, another contradiction worth noticing ^[23]. By contrast, oil contact appears to make some materials worse: polystyrene noodle cups in oily conditions released roughly ten times more nanoplastics than non oil containers as foam unit cells collapsed above 70 °C ^[24].

The most replicated conclusion is not the one most often quoted. Microwaving plastic in contact with food or water releases particles. How much depends on the polymer, the food matrix, the temperature, and the method used to measure it. Different measurement methods are good at different things. Some excel at micron scale counting, others at nano scale resolution, others at controlling for matrix interference. The variance across studies is partly the variance across techniques in a young field still finding its standards. The signal across all of it is consistent enough to act on. Our reading of the literature is that the case for keeping plastic out of the microwave is strong, even as the specific “billions per square centimeter” figure remains one estimate among several rather than a settled number.

Baby bottles and infant formula preparation

The strongest concentrated body of evidence in the heat literature is on polypropylene infant feeding bottles. Much of the foundational work has come from one laboratory, and a growing set of independent studies has begun to fill in around it.

In 2020, Dunzhu Li and colleagues at Trinity College Dublin published the paper that brought this category into focus. Following World Health Organization guidance to sterilize bottles and prepare formula with water at 70 °C or hotter, they measured particle release from polypropylene bottles into prepared formula at an average of 16.2 million particles per liter across ten products, rising to 55 million at 95 °C ^[25]. Modeled global infant exposures ranged from 14,600 to 4.55 million particles per day, with large regional variation.

Subsequent work has largely supported the same directional finding while adding more detail. A 2025 laser direct infrared imaging study found that thermal aging concentrated release in the smallest size fractions ^[26]. Another study reported that brand and bottle design explained more of the variation in the Chinese market than polymer family alone ^[27]. An infant fecal biomarker study found PET concentrations in infants roughly ten times those seen in adults ^[28]. Polypropylene particles shed from infant bottles have also been shown to activate oxidative stress and inflammatory signaling in cultured intestinal cells ^[9], though at concentrations well above realistic infant exposure. Breastmilk storage bags may add an estimated 0.61 to 0.89 milligrams per day for infants whose milk is stored exclusively in bags ^[29], and scoping reviews now document microplastic detection across placenta, amniotic fluid, and meconium ^[30].

A transparency note. The 2020 Li paper and several of the highest profile follow ups ^[7,31] cluster around Trinity College Dublin. That is worth knowing, not because it weakens the finding, but because any field benefits from independent replication. Recent work is starting to deliver it. An independent 2025 assessment of polypropylene and polycarbonate bottles found the same overall pattern, release rising with temperature, while also showing that absolute counts shift meaningfully depending on the detection method used ^[85]. That is what a maturing literature should look like: the same direction, with the magnitudes still being calibrated.

The World Health Organization’s 70 °C guidance exists for a serious reason: to reduce the risk of pathogens such as Cronobacter sakazakii. That guidance should not be discarded. A more practical takeaway is to follow it, then cool prepared formula in glass before transferring it to the bottle when that is workable.

Hot drinks: kettles, paper cups, tea bags, drip coffee

Hot drinks are where the field becomes especially interesting, because the same products often land at opposite ends of the headline scale.

Plastic kettles

On paper, plastic kettles seem like they should be among the worst offenders. Boiling water, plastic walls, repeated heat cycles. Early studies did report substantial particle release when kettles were boiled in deionized water. But a careful series from the Trinity College Dublin group showed something more nuanced: in real tap water, with its dissolved calcium, copper, iron, and bicarbonate ions, those same kettles develop a thin mineral passivation film over the first few weeks of normal use. That film suppresses release by roughly 89 to 99.8 percent compared with the deionized condition ^[32,33]. A 2025 follow up on Australian polypropylene kettles reported first boil nanoplastic release of about 0.011 micrograms per square centimeter, falling to 0.0004 by the 150th boil ^[34].

In other words, release appears to decay with use, and harder tap water seems to accelerate that decline. It is one of the more replicated and less widely known findings in the literature, and a useful reminder that worst case experiments under simulant conditions do not always reflect ordinary use.

Paper cups, which are mostly not paper

What most people call a paper cup is usually paper bonded to a thin polyethylene, polylactic acid, or polypropylene lining. That lining is what holds the liquid. It is also what meets the heat.

A 2020 study filled disposable paper cups with hot water for 15 minutes and reported roughly 25,000 micron sized particles per 100 mL cup, along with submicron particles, dissolved fluoride and chloride ions, and trace lead, chromium, and cadmium ^[35]. Lifetime intake modeling for habitual users centered on about 0.03 mg of microplastics per kilogram of body weight per day ^[36]. Other groups have reported hundreds to a few thousand particles per liter from similar cups ^[37-39], while Indian and European surveys estimate annual exposure from daily paper cup use in the hundreds of millions of particles per year ^[40,41]. A 2023 paper linked paper cup release to compromised catalase activity in vitro ^[42]. Particle characterization also skews small, with most particles below ten micrometers and a longer fiber tail in cups with cellulose containing linings ^[43].

One finding here is especially worth pausing on. A 2023 study comparing compostable PLA lined cups with conventional polyethylene lined cups reported that the PLA cups released 4.2 times more microparticles, along with a cellulose microfiber population unique to the PLA lining ^[44]. In a separate comparison, LDPE lined cups released more nanoplastics than PLA ^[45]. Which material appears worse depends, in part, on whether the focus is micron scale or nano scale release. A 2026 PLA study also reported a hundredfold jump in release between 50 °C and 70 °C, while noting that up to 55% of the detected “nanoparticles” were actually self assembled oligomer aggregates rather than true polymer particles ^[3]. A 2024 pregnancy model study found that paper cup microplastics accumulated in mouse placental and fetal tissues, with an inferred benchmark dose roughly equivalent to two to four cups per day ^[46]. It is one study, in mice, and the dose conversion to humans is not direct. It is also a result striking enough that the underlying signal is hard to ignore.

Tea bags

The best known tea bag paper is still the 2019 study from Laura Hernandez and colleagues at McGill University. Using nylon and PET pyramid tea bags steeped at 95 °C, the authors reported that a single bag released roughly 11.6 billion microplastic particles and 3.1 billion nanoplastic particles into one cup of tea ^[47]. It remains the most cited number in this part of the field.

It is also contested. A methodological commentary argued that interference from the tea matrix itself, including polyphenols, plant colloids, and undigested debris, may inflate apparent counts, and proposed chemical digestion steps to clean up the analysis ^[48]. A 2026 study using optical coherence tomography on similar nylon tea bags reported 16,000 to 24,000 microplastics per milliliter, several orders of magnitude lower than the earlier nano count estimates ^[49]. An independent 2025 study found that release rose over time and then plateaued, again at levels well below the 2019 headline number ^[86]. Other groups have reported particle counts in the billions per liter range, but without measurable cytotoxicity in cultured intestinal cells ^[50-52]. Cross market surveys have also reported hundreds of particles per bag, with phthalate co-release flagged as a concern ^[53,54]. Embryonic zebrafish exposed to tea bag derived particles show oxidative stress endpoints at high doses ^[55]. The European Food Safety Authority’s 2025 review explicitly identifies tea bag studies as a category prone to overcounting.

The fairest summary is that plastic tea bags release particles, the release rises with temperature, and the exact magnitude is still being calibrated across methods. The “billions per cup” figure is one early estimate at the high end of a still settling range, not the only number on the table and not a reason to dismiss the underlying observation. A simple cold water pre rinse before brewing reduces release by 76 to 94% in controlled tests ^[82], which makes it one of the rare low effort habits in this literature with direct experimental support.

Drip coffee bags and herbal pouches

Single serve drip coffee bags, the kind that hang over a mug while hot water is poured through them, release more than 10,000 microplastic particles during a five minute pour at 95 °C, with rayon fibers dominant ^[10]. Heated liquid herbal medicine pouches release roughly 3.5 times the unheated baseline ^[56], and Chinese herbal decoction packages have been reported to release up to 1.21 million microplastic particles and 4.32 billion nanoplastic particles per package ^[57]. A 2025 UK survey of 155 beverages detected microplastics in 100 percent of samples, with hot tea averaging 60 particles per liter and hot coffee 43 particles per liter ^[58].

Taken together, hot drinks do not point to one simple villain. They point to a broader pattern: heat, repeated contact, and thin plastic barriers create opportunities for migration and release, but the scale depends heavily on the material, the drink, and the method used to measure it.

The other heated plastic moments

Takeout containers, reheated

Disposable takeout containers are the most studied product category in this literature. The reported numbers vary widely from study to study, but the directional findings are consistent: higher heat increases release, oil rich foods increase release, foamed polystyrene performs worse than many alternatives, and release tends to rise steeply with temperature ^{[20,22,59-61]}. A 2025 independent study of polypropylene containers in hot and cold water confirmed the same temperature dependence while reporting absolute counts well below the most cited microwave headlines ^[83]. In one set of experiments, microwaving leftovers in oil rich sauces increased release by as much as 125 fold compared with water only conditions ^[62,63]. Rodent studies have also reported intestinal barrier effects from polypropylene cutting board particles ^[64] and from paper cup derived particles at high doses ^[65]. Whether typical human exposures reach biologically meaningful doses remains unknown. The mechanism itself, however, is real and reproducible.

Dishwashers

The dishwasher may be one of the most overlooked heat plus plastic moments in the kitchen. A 2025 study published in ACS ES&T Water measured particle release from common plastic items during mechanical dishwashing cycles and reported roughly 920,000 particles released per cycle, with intensive 70 °C cycles releasing about three to five times more than cold pre-wash conditions ^[66]. It is only one paper, and the only study of its kind in this corpus, so it should be treated as early evidence rather than settled consensus. Still, it points in a familiar direction: heat, water, abrasion, and repeated use tend to work together.

Cookware: nonstick scratches

PTFE, the coating behind most nonstick cookware, is generally stable at normal cooking temperatures. The concern is usually not the intact surface. It is the damaged one. A 2022 Raman imaging study estimated that a single visible scratch on a PTFE pan could release roughly 9,100 particles during one cooking cycle, while broken or flaking coatings could release closer to 2.3 million ^[67]. A 2024 comparison of plastic and non plastic cookware found that non plastic cookware contributed essentially zero microplastics to a food simulant, while plastic cookware contributed several thousand particles per year under daily use assumptions ^[68]. Reusable plastic kitchenware used in the oven adds another layer: in one comparison, melamine released the most particles, along with thermal degradation products of concern ^[69]. A 2024 bakeware study using advanced infrared imaging also found that PET bakeware heated to 220 °C released particles in the 1 to 20 micrometer range, with release increasing across repeated baking cycles ^[70].

Hot cars and PET bottles

The PET water bottle left in the door pocket of a sun warmed car has one of the more established stories in this literature. Antimony, a residual catalyst from PET manufacturing, migrates into water as temperature and storage time increase, occasionally exceeding regulatory thresholds under extreme conditions ^[11,71,72]. A 2025 study exposing PET bottles to natural sunlight over 30 days found microparticle concentrations rising to 14 to 20 micrograms per liter before plateauing ^[73]. A 2026 study that cycled bottles at 60 °C with shaking reported a 9.29 fold increase in nanoparticle concentrations across eight US bottled water brands ^[74]. These are upper bound experiments, not everyday averages. The practical implication is simple enough: water bottles are not ideal for long stretches in hot cars.

Silicone bakeware

Silicone is the quieter part of this story. It appears inert, and at the level of the intact polymer it often largely is. But when heated, silicone can release cyclic siloxanes, including D4, D5, D6, and related oligomers, into both indoor air and baked food. A 2019 study measured indoor air concentrations of D4 to D6 peaking at 301 micrograms per cubic meter immediately after baking, while cakes contained 3.7 to 6.6 mg/kg of cyclic siloxanes ^[12]. A 2025 paper extended that picture, reporting levels as high as 646 micrograms per cubic meter under some conditions ^[75]. At the same time, a 2024 study of 44 EU silicone molds found that all complied with relevant Spanish regulatory limits, with no detectable migration into 3 percent acetic acid or 20 percent ethanol simulants ^[76]. As with other materials, new silicone appears to release more than silicone that has already been used and cured through repeated heating cycles.

What we still do not know

This is the hinge of the article. Almost all of the evidence above tells us what comes out of a cup, a bottle, a container, or a pan under heat. Very little of it tells us what those released particles actually do inside the human body at realistic everyday exposures.

The largest gap is the dose gap between laboratory toxicology and plausible dietary intake. In cell culture studies, researchers often expose cells to released particles at concentrations of 100 to 1,000 micrograms per milliliter for 48 to 72 hours ^[9,16,21,50]. At those levels, oxidative stress, membrane damage, and reduced cell viability are common findings. But those are hazard studies, not exposure matched human models. The concentrations used are often roughly a thousand to a million times higher than even high end estimates of realistic dietary intake from the same broader literature, which generally fall in the nanogram to milligram per day range ^[58,62]. Findings at those doses may help identify what a material can do under stress. They do not, on their own, tell us what it is doing in an ordinary kitchen or inside an ordinary person.

The human evidence is thinner still. One of the most careful human studies in this corpus is, notably, a null result. Curtis Tilves and colleagues followed an infant cohort and found that plastic bottle use at three months was not strongly associated with growth or adiposity, and that early signals in short chain fatty acids and microbiome composition weakened after adjustment for diet ^[77]. A small number of adult cross sectional studies have reported associations between heavy takeout consumption, fecal microplastic counts, and changes in BMI or the microbiome ^[78-80]. But those studies do not establish temporality, and they cannot fully separate plastic exposure from the obvious confounders that travel with takeout heavy diets, including calorie density, sodium, processed fats, and lifestyle. These are correlations. They are not proof of causation.

There is also a more basic measurement problem underneath much of the field: what, exactly, counts as a particle. Many of the submicron or nanoplastic signals reported by light scattering techniques may include not only true polymer fragments, but also dissolved oligomers and additives that cluster together in solution. Fang 2026 reported that up to 55 percent of detected PLA “nanoparticles” were actually self assembled oligomer aggregates ^[3]. Caldwell 2021 raised a similar concern for tea based matrices. EFSA’s 2025 review identifies this as a recurring source of overcounting. So when a study reports nanoplastics, one of the first questions should be whether the method can actually distinguish a solid polymer fragment from an oligomer cloud. In many cases, it cannot.

A third uncertainty is what happens after environmental aging. It is often assumed that sunlight steadily breaks plastics into smaller and smaller particles, eventually driving a cascade toward nanoplastics. But a 2024 study testing polypropylene in surface waters found something more restrained: sunlight did drive oxidative mass loss, but in a slow, roughly linear way, without the runaway fragmentation cascade that many had assumed ^[81]. One result does not settle the question. It does, however, argue for more caution in how confidently the mechanism is described.

So where does that leave the science in 2026? It leaves it in a place that is both clearer and more limited than many headlines suggest. Heat does release particles and chemicals from plastic food contact materials. That much is established. The exact quantities released under realistic kitchen conditions remain uncertain, and are likely lower than the most quoted headline figures. What is released is often a mixture of true polymer fragments, additives, and dissolved oligomers, and current methods still struggle to cleanly separate those categories. At high doses, these materials can produce cellular stress in culture and in animal models. Whether realistic dietary doses produce measurable health effects in humans remains, for now, unknown.

That final point matters most. The science is strong enough to justify caution and to justify a few small changes in habit. It is not yet strong enough to quantify human harm with precision. Both of those things can be true at once. A young field with a consistent signal and unsettled magnitudes is exactly the kind of field where a little proactive care, taken without panic, is the reasonable response.

A pragmatic kitchen checklist

If everything above sounds like an excuse to do nothing, that is not the intended takeaway. A small number of kitchen habits are reasonably well supported, low cost, and worth adopting. They will not eliminate exposure. What they can do is reduce some of the most direct and repetitive contact moments, which is most of what any of us can actually control.

1. Move hot food and drinks into glass or ceramic before they sit in plastic. Well supported. Across polymers and methods, temperature is one of the clearest drivers of release ^[20,60]. Glass and ceramic do not carry the same issue ^[68].
2. Do not microwave food in plastic containers. Use glass or ceramic instead. Supported, with caveat. The most extreme microwave figures come from one contested paper ^[16,17]. But the broader literature still points in the same direction: microwave heating increases release relative to room temperature ^[20].
3. Choose loose leaf tea with a stainless infuser, or pre rinse plastic tea bags with cold water. Supported, though the counts are contested. The headline figures from Hernandez 2019 remain debated ^[48,49]. The release itself is real. In controlled tests, a simple three times cold water pre rinse reduced release by 76 to 94 percent ^[82].
4. Replace nonstick pans when the coating is visibly scratched or flaking. Well supported. Intact PTFE is largely inert at normal cooking temperatures. Damaged coatings are different and can shed measurable particles during cooking ^[67]. If the surface is visibly compromised, it is time to replace it.
5. Do not leave plastic water bottles in hot cars for extended periods. Supported. Antimony migration from PET rises with both temperature and storage time ^[71,72]. Sun exposure can add particle release on top of that ^[73,74].
6. Use a ceramic mug or stainless tumbler for hot drinks instead of disposable cups. Supported. Disposable cups consistently release particles into hot liquids, even if the exact quantities vary by study ^[35,36].
7. Let cooked food cool for a few minutes before sealing it in a plastic container. Mechanism supported, precautionary. Hot food pressed against a softened polymer sits in the same window where many release studies report their highest counts ^[19]. The cooling step costs nothing.
8. For infant formula, follow the WHO 70 °C guidance, then transfer the prepared formula from glass to the bottle once it has cooled enough to feed. Best effort. The WHO temperature recommendation matters for microbiological safety and should not be compromised ^[25,26]. Cooling and transferring is one of the few practical kitchen side adjustments available.

The point is not to do all eight things perfectly. It is to make two or three of them automatic, so they happen without much thought.

A closing note

The science of heat and plastic food contact is newer and less settled than the headlines often make it seem. The same tea bag, measured by two different labs, can produce numbers that differ by orders of magnitude. The most rigorous independent review in this literature, the European Food Safety Authority's 2025 evidence synthesis, concludes that real-world release is likely lower than many of the most quoted figures, while still acknowledging that release does occur and that some combinations of polymer, food, and temperature produce more of it than others.

That is not a reason to dismiss the subject. It is a reason to approach it with care.

Heat is one of the few variables we can meaningfully adjust in everyday life. Not by replacing everything. Not by carrying a sense of contamination into every meal. Just by making a few quieter choices: moving the hot thing into glass before it sits, replacing the scratched pan, drinking tea from a ceramic mug, not leaving the water bottle in the car.

The science will keep sharpening. In the meantime, these habits are modest, inexpensive, and proportionate. They do not ask for perfection. They ask only for a little attention to the moments where material, temperature, and repetition meet. And they leave room for the part of food that has always mattered most: nourishment, ritual, and the people around the table.