Adhesion failure gets blamed on the ink almost every time, and the ink is usually innocent. Work this list in order — it is ordered by how often each cause turns out to be the answer.
1. Measure the dyne level on the actual roll
Not the specification. Not a sample from a different batch. The roll that is on the press.
Corona treatment decays. A film treated to 42 dynes/cm at the film plant loses surface energy in storage as the oxidised surface groups reorient inward and as additives migrate outward. Warm warehouses accelerate it.
Use a dyne pen or test inks. Apply a line; if it holds continuous for two seconds, the surface is at or above that value. If it beads, it is below.
You want the substrate roughly 8-10 dynes/cm above the coating's surface tension. Solvent-based inks generally need 38+ on film. Water-based systems, having higher surface tension, generally need 40-44.
If the dyne level is low, stop here. Nothing you do to the formulation will fix an untreated surface. Get in-line treatment, or get fresher film.
2. Check the age and additive package of the film
If the dyne level is adequate but adhesion still fails, ask what the film contains.
Slip additives — erucamide and oleamide are the common ones — are compounded into polyolefin film and migrate to the surface over time by design. That migrated layer is exactly what it sounds like: a low-surface-energy lubricant sitting between your ink and the substrate.
The diagnostic signature is distinctive. Adhesion passes on fresh film and fails on aged stock of the same specification, with nothing changed on press. If you see that pattern, you are looking at additive bloom, and the fix is specifying low-slip film for printed surfaces or printing the film fresher.
3. Look for contamination
Silicone is the classic. It comes from release papers, lubricants, aerosol sprays used near the line, and occasionally from a previous job on the same equipment. Tiny quantities destroy adhesion and produce craters and fisheyes.
Plasticiser migration from certain substrates, residual oils from film handling, and dust all contribute.
A quick test: wipe a small area with isopropanol, let it dry, and coat it. If adhesion improves dramatically on the wiped area, you have surface contamination.
4. Confirm the coating actually dried or cured
An undercured film is soft at the interface even when the surface looks fine.
For solvent systems: check that drying capacity matches line speed. Retained solvent at the substrate interface weakens the bond and shows up later as delamination in the laminate.
For UV systems: check cure at the substrate interface, not the surface. In pigmented films the top of the coating absorbs the UV and the bottom can be starved — particularly in whites and heavy film weights. Undercure at the interface is a common cause of UV adhesion failure. Test by solvent rub on the reverse if the substrate allows.
5. Consider cure shrinkage
This one is specific to UV and catches people out because the wetting was fine.
As acrylate double bonds convert during cure, the film contracts. In a highly functional formulation that shrinkage is significant, and it pulls against the substrate bond. On a marginal substrate, a formulation can wet perfectly, cure hard, and then lift.
If you see adhesion failing on a high-functionality formulation that wets well, try reducing crosslink density — lower functionality oligomer, or more monofunctional diluent — and see whether adhesion recovers.
6. Only now, look at the formulation
If all of the above is clean, the formulation is genuinely in question. At that point the useful questions are whether the binder chemistry suits the substrate polarity, whether an adhesion-promoting component is warranted, and whether the pigment loading is high enough to be starving the binder.
The pattern worth internalising
Adhesion is an interface property. The interface has two sides, and the substrate side changes over time in ways the specification sheet does not record.
When adhesion fails after a period of working fine, and nothing changed in the ink, something changed in the film. That is almost always where the answer is.