Pull on an adhesive bond and the contact area shrinks until detachment. This is the standard picture — every textbook model, every failure criterion. Tension reduces contact. But this is only true on short timescales.
On long timescales, the contact area under sustained tensile load expands. The mechanism is a competition between two viscoelastic processes. Interfacial viscoelasticity tries to shrink the contact (the expected behavior). Bulk creep tries to spread it (the unexpected one). At long timescales, bulk creep dominates. The adhesive material flows outward under load, increasing the bonded area even as the load tries to peel it away.
This expansion phase is not strengthening. It is the precursor to catastrophic failure. The adhesive spreads thin, the interfacial stresses redistribute, and when the creep-driven expansion can no longer compensate for interfacial degradation, the failure is sudden and complete.
The engineering consequence: an adhesive joint that appears to be getting stronger — contact area growing, load borne stably — may be accumulating the conditions for abrupt release. The appearance of persistence masks the approach of failure. This explains why structural adhesives can hold for years and then fail without warning: the expansion phase looked like stability, and the inspection criteria confirmed it, right up until the moment of collapse.
Persistence is not the same as stability.