Why labels lift at the edges, and how oil migration destroys the adhesive

Oil wicks through paper facestock and plasticises the adhesive, and PP never bonded. Facestock choice, surface preparation and a 72 hour adhesion test at 40 C.

A label that lifted at one corner three weeks after packing is rarely the glue being cheap. Three different failures produce almost the same picture: the adhesive never wet the container at all, the adhesive wet it properly and was then dissolved from behind by the product, or the facestock is stiff enough to peel itself off a curve. They need different fixes, and the container you peeled the label off tells you which one you have.

• Polypropylene: about 29 mN/m
• Facestock: 50 micron BOPP
• Dwell: 24 h before packing
• Test: 72 h at 40 C

Three failures that look identical on the shelf

A pressure sensitive label is three layers: a facestock you print on, an adhesive, and a release liner you throw away. Peel a failed one off deliberately and look at both the container and the back of the label before you change anything.

Interfacial failure leaves the container clean and dry. All the adhesive is still on the label. The adhesive never wetted the surface, and the cause is the surface energy of the plastic, a film of contamination, cold application, or too little pressure at the moment of application. Nothing about the adhesive chemistry was wrong; it never made contact at the molecular scale.

Cohesive failure leaves adhesive on both surfaces, often stringy or visibly wet, and the label may have slid a millimetre or two out of position. The adhesive bonded and was then destroyed. In a balm pack that means oil got into it.

Facestock failure shows in the paper or film itself: darkened and translucent, ink gone dull, edges puckered, or a wrinkle running across a curve. The bond may still be intact underneath. That is a specification problem, and no amount of surface preparation touches it.

Oil migration, and why uncoated paper is the worst possible choice

Uncoated paper facestock, typically 70 to 90 gsm, is a mat of cellulose fibres with nothing in the way of a barrier. Balm oil reaching the printed face, from a smeared lid, a customer's fingers or the film left after filling lip balm tubes, wicks through it by capillary action within days and reaches the adhesive from behind. The visible tell is a translucent halo that spreads outward from the point of contact and dulls the print. By the time you can see it, the adhesive has already been reached.

In the adhesive layer the oil acts as a plasticiser and a solvent at the same time. A pressure sensitive adhesive is a viscoelastic solid tuned to a narrow window: soft enough to flow into surface roughness under thumb pressure, stiff enough not to creep under load afterwards. Absorbed oil swells the polymer network, lowers its glass transition temperature and collapses its cohesive strength. The counterintuitive result is that tack often goes up while holding power goes down, which is why oily labels feel stickier than ever and still slide out of position and ooze at the edges.

Adhesive chemistry decides how long that takes. Tackified rubber hot melts are hydrocarbon systems and a triglyceride oil is a reasonable solvent for them, so they go first. Acrylics are more polar and are usually crosslinked, so they resist longer, and a crosslinked permanent acrylic is the sensible default for anything oily. Films resist entirely, because the oil never gets to the adhesive through the facestock, only around the edges. That is the whole argument for BOPP.

Surface energy: polypropylene was never going to hold

An adhesive can only bond to a surface it wets, and wetting requires the solid's surface energy to exceed the adhesive's own surface tension by a working margin of roughly 8 to 10 mN/m. General purpose permanent acrylics are formulated to wet substrates at about 36 mN/m and above. That one number explains most tube failures.

Two consequences follow that are not obvious. First, corona treatment applied at the extruder is not permanent: it decays over weeks to months, faster in warmth, so a box of tubes that labelled perfectly in January can fail in June from the same delivery. Second, polyolefin containers are usually compounded with slip additives such as erucamide or oleamide, and those additives are designed to migrate to the surface. An older tube has a thin fatty layer on it that no dyne pen reading from the factory ever saw. Both effects punish the maker who buys packaging in bulk and works through it slowly, which is otherwise the sensible thing to do, as sourcing sets out.

A 12 mm tube is a leaf spring

Wrap a flat, stiff label around a narrow barrel and you have stored elastic energy in it. That energy pulls at the leading and trailing edges continuously, and the adhesive only has to creep a fraction of a millimetre for an edge to lift. Nothing degraded; the label simply won.

Bending stiffness scales with roughly the cube of caliper, so thickness matters far more than it looks. An 80 gsm paper facestock at about 90 microns is somewhere near eight times stiffer in bending than a 50 micron BOPP film, for the same width. Curvature matters just as much: a standard lip balm barrel at 12 to 16 mm outside diameter is about the tightest curve in cosmetic packaging, while a 60 mm tin is nearly flat by comparison. That is why the same labelstock behaves impeccably on tins and fails on tubes, and why the format choice covered in tube versus tin quietly sets your labelling budget too.

Three geometry rules follow. Make a wrap label longer than the circumference so it laps onto itself: a 12 mm barrel has a circumference of 37.7 mm, so a 42 mm label gives a 4 mm lap, and film sticking to its own film is a far better bond than film sticking to polypropylene. Radius the corners at 2 to 3 mm, because a square corner concentrates peel stress at a point. And keep the label 3 mm clear of any rolled edge, seam or moulding line, since an adhesive cannot bridge a step without leaving a channel that oil and air then travel along.

The residue left during filling, which causes more failures than everything else

Adhesion happens in the first few nanometres. A monolayer of oil is enough to prevent it, and a balm workshop coats everything in exactly that. Splash from pouring, a run down the outside of a tin, aerosol from a heat gun used to reflow a dipped surface, warm containers handled with bare fingers, and the general film that settles on a bench where wax is melted. None of it is visible on a matt tinplate lid.

Containers also arrive dirty. Pressed metal tins carry residual drawing lubricant from forming, and injection moulded tubes carry mould release plus the slip additive already mentioned. Neither is a defect, and both will stop a label sticking, in the same way that residue left inside a tin sets up rust spots.

You can check for it without buying anything. Run water over a clean metal or glass container: on a genuinely clean high energy surface it sheets out into an unbroken film, and on a contaminated one it beads and breaks. That water break test is borrowed from metal finishing and takes ten seconds. On plastics it tells you less, because the plastic is low energy anyway.

The 72 hour test that settles the argument

Labelstock is sold by construction, not by performance against your formula in your container. The only useful evidence is a test, and this one takes six labels and three days.

1. Prepare six containers in two groups of three. Three exactly as they come off your filling bench, oil film and all. Three wiped with isopropanol on a lint free wipe, one pass, a fresh face of the wipe each time, left to flash off dry. Fill all six with the actual formula, not a stand-in, and cap them.
2. Let them reach room temperature first. A warm container softens the adhesive and a cold one stops it wetting out. Most permanent acrylics have a minimum application temperature around 5 to 10 C, and a chilled container also collects condensation.
3. Apply the labels with real pressure. Roll them down from one edge with a rubber roller or the back of a spoon, not a fingertip press in the middle, which traps air. Mark each container with its group and the date, or give the trial its own code from the batch code generator.
4. Hold 72 hours at 40 C. An oven with a reliable low setting, an incubator, or a warm cupboard with a thermometer in it. This is the same 40 C used for accelerated ageing: it is roughly what a delivery van reaches in summer, it softens the balm so its oil is at its most mobile, and it accelerates adhesive creep.
5. Cool for an hour before testing. A warm adhesive tests soft and will flatter itself.
6. Peel the leading edge. Lift it with a thumbnail and pull steadily back on itself. Record which of the three failure modes you get, and whether the facestock tears or stretches before it releases.
7. Keep three more at room temperature for twelve weeks. Oil migration through paper is slow, and the 40 C test does not fully substitute for it.

The pass criteria are strict on purpose: no edge lifted on its own, no adhesive left on the container, no darkening of the facestock, and enough bond that the label distorts before it releases. Anything less is a problem your customers would have found instead.

Reading the failure

Rescuing the run in front of you

Nothing is wrong with the balm, so do not remelt anything: reprocessing a sound product to solve a packaging fault gains nothing and costs you the crystal structure you spent time on. The decision is about labour and about what the pack must legally carry.

If the labels release cleanly and the containers are not oily, peel, degrease and relabel. Budget 30 to 60 seconds a unit and decide whether that is cheaper than the alternatives at your volume, using the same logic as any other rework in costing and pricing. If adhesive residue has been left behind, relabelling is no longer cheap: the residue has to come off with isopropanol and a plastic scraper, and on a printed or lacquered tin that often takes the print with it.

Above a few hundred units, oversleeving usually wins: a printed paper band, a shrink sleeve or a folding carton carries the mandatory information without touching the failed label. What you must not do is press a lifted label back down. A peeled adhesive has collected dust and lost its wet-out, and re-pressing it buys a fortnight at most.

Specifying so it does not happen again

Then get the application right, because a correct specification applied badly still fails. Work between 18 and 25 C. Roll the label down under even pressure. Allow dwell: a pressure sensitive adhesive develops perhaps a quarter of its ultimate bond at the moment of application, most of the rest over 24 hours as it flows into the surface roughness, and reaches its final value somewhere between 24 and 72 hours. Bagging or boxing labelled stock straight away, while the bond is at a quarter strength, causes edge lift that gets blamed on the adhesive months later. Store label rolls in their bags at 15 to 25 C away from the melting bench, since paper stock is hygroscopic and a cold roll brought into a warm room collects condensation on the outer wraps.

Finally, order a sample sheet before a roll of five thousand. Converters will send them, and one sheet run through the 72 hour test is worth more than any technical data sheet. The same discipline that keeps packaging compatibility testing honest applies here, and the container decisions that feed it are in the packaging guide.

What the test can tell you, and what it cannot

The 72 hour test tells you whether this labelstock survives this formula in this container for three days at 40 C. It does not tell you what happens after nine months on a warm shelf, and it cannot be extrapolated to a different formula. A high castor oil lip balm, a soft sweating shea balm and a hard beeswax salve present completely different amounts of mobile oil to the same pack, and a formula change is a reason to retest.

Supplier data will not close the gap either. Peel figures on a technical data sheet are measured on stainless steel, usually as a 180 degree peel after 20 minutes and after 24 hours, and stainless steel is the friendliest substrate there is. Two stocks described identically as 50 micron white BOPP with a permanent acrylic can behave quite differently against an oily polypropylene tube.

So the decision rule is short. If the container is a polyolefin, assume low surface energy until a dyne pen says otherwise. If the facestock is uncoated paper, assume oil will reach the adhesive eventually. If either is true, run the test before you commit to a roll, and if both are true, change the specification without bothering to test. And accept the trade-off honestly: the aesthetic that sells best in this category, uncoated natural paper with no varnish on a narrow tube, is the construction most likely to fail, and the fix is a different material rather than a better technique. Where that matters most is stock that will sit in heat, which is also the argument made in shipping in hot weather.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.