Quick answer: A cosmetic oil in PET bottles, sealed with a PE-sealant induction liner, passed 100 percent of release testing at filling. Six weeks into stability storage, seals began to lift. The bond was never weak — the sealant film was being plasticised by the oil, losing cohesive strength from the product side outward. The fix was a sealant polymer change to PET, not a change to the sealing process.
This case is published because it is the most expensive failure mode in sealing liners: the one that ships. At Hangzhou Xinxing Pharmaceutical Packaging we treat product-sealant compatibility as a separate question from seal quality, and this project is the reason that separation is written into our enquiry process.
The situation
A cosmetic brand packing a body oil in PET bottles approached us after a shelf-life failure in a stability programme. The containers had been filled and sealed on a production line with an established induction sealer, tested at release with a vacuum test and a manual peel check, and released. All results were within specification.
At the six-week stability checkpoint, a proportion of containers showed seals that had lifted at part of the circumference. Peel strength on the affected containers was far below the release value. The seal had not leaked yet, but the barrier was compromised and the product would not have met its declared shelf life.
Why release testing did not catch it
Release testing measures the seal as it exists at the moment of filling. A vacuum test asks whether the seal holds pressure today. A manual peel asks how much force is needed to lift the foil today. Neither asks what the product will do to the sealant film over the next two years.
That is not a testing failure; it is a testing scope question. Seal integrity testing answers "is this container sealed". Compatibility testing answers "will it still be sealed in six months". A specification can pass the first and fail the second indefinitely, because the two are measured with different instruments over different timeframes.
What had happened at the seal
The liner used a PE sealant film, which is the correct default for aqueous and mildly aggressive products and is the most widely used sealant in pharmaceutical and cosmetic packaging. The fill, however, was an oil-based formulation with a substantial fraction of low-molecular-weight esters and a fragrance load.
Over weeks in contact, those components migrated into the sealant film and acted as plasticisers. The film softened, its cohesive strength fell, and the bond to the PET rim began to fail first at the edge of the sealing land where the film was thinnest. The failure progressed inward, which is why the first visible symptom was a lifted crescent rather than a complete release.
Crucially, this process does not require a defective liner. The liner met its specification. The specification had simply been written against the container and the line, and not against the product.
The fix
The sealant polymer was changed from PE to PET. PET resists oil and solvent migration considerably better than PE, at a modest unit cost premium and with no change to the sealing process. The line settings that had been qualified were retained, since the foil thickness and the energy requirement were unchanged.
The revised construction went through a compatibility soak before commercial release rather than after: containers sealed and stored at elevated temperature for a period equivalent to the intended shelf life, then tested for peel strength and inspected for channel formation. It passed, and the product went to market on the revised specification.
What the case illustrates
Product aggressiveness is a specification field, not a product category. "Cosmetic" covers creams, lotions, gels, oils and alcohol-based fragrances, and the four behave completely differently against a given sealant. The field that predicts the outcome is the carrier — aqueous, oily, alcoholic, solvent-bearing — plus the fill temperature and the shelf life. It is the single most under-specified field in liner enquiries and the one that determines the largest share of the unit cost.
A PE default is right most of the time and wrong expensively. Because PE is the most common sealant, it is often specified without a decision being made. That is fine for aqueous products. For an oil, an alcohol load, or a surfactant-rich formulation, the default is the failure mode.
Compatibility testing is cheap relative to its consequences. A soak test costs a bench, some containers and a few weeks of calendar time. The alternative is a stability failure that is discovered after the product is on a shelf, which costs a recall, a customer relationship and a re-qualification programme — all of which were on the table in this project before the test was run.
The enquiry question that prevents it
Our enquiry form now asks for the carrier type, the fill temperature, the intended shelf life and any known aggressive ingredient, in that order, before the container dimensions. It is a deliberate ordering. The container determines the die; the product determines whether the die is being cut for something that will still hold together in a year.