The Edge Banding Didn’t Fail. The Particleboard Edge Killed Both.

Written by

in

The Edge Banding Didn’t Fail. The Glue Didn’t Fail. The Particleboard Edge Killed Both — and Your Glue-Line Test Can’t See It Coming.

I pulled a three-year-old kitchen base cabinet out of a coastal apartment last spring. The PVC edge band was still attached along most of the run. But within 150 mm of the sink cutout, the band lifted cleanly off the substrate, and the particleboard underneath had swelled enough that you could feel the step with a fingernail. The glue line looked fine where the band was still down. The band itself was intact. The adhesive hadn’t degraded. What failed was the edge of the board — and the specification that treated edge banding as a product rather than a system with a moisture path running through it.

This is the failure mode that does not show up in a dry 180-degree peel test, and it is the one that takes cabinets out of service on coastal jobs, in wet-area returns, and beside dishwashers.

The peel number that passes the lab and misses the field

You will see PVC edge banding adhesive peel strength quoted in N/mm or N/50mm. A 180-degree peel test on a dry, conditioned sample is the standard gate. Under those conditions, a well-controlled EVA line can sit in the 1.8–2.2 N/mm range, and a PUR line runs roughly 3.5–4.1 N/mm — PUR is clearly stronger on a dry pull. Those numbers are real, and they matter for handling and short-term integrity.

But the peel test measures adhesion at an interface, under a load applied after the sample has been dried and stabilized. It does not measure what happens when the substrate behind that interface moves. On a melamine-faced particleboard or MDF panel, the face is sealed. The cut edge is not — and every exposed edge, every screw hole, every sink cutout is a path for moisture into the core. When that core takes on water, it does not just sit there. It swells, and the swelling generates a slow, persistent peeling force along the glue line that no single-pass peel number captures. A PVC edge banding delamination particleboard failure often reads as an adhesive problem when the adhesive is doing exactly what it was designed to do — holding against a force that was never specified.

There is a second, quieter driver. Field reports of edge banding adhesive failure humidity-driven jobs repeatedly point to the same pattern: samples kept in the shop retain adhesion, while installed panels lift after two to three years, especially near high-moisture zones. The lab sample and the installed panel are not the same specimen. The installed panel has a wet end.

The one lab number that predicts field behavior better than peel strength

If I am qualifying a PVC edge banding specification for a wet-area or coastal job, the data sheet line I read before the peel column is the Vicat softening point — and I read it against the actual service temperature the edge will see, not against the ambient room temperature.

Commercial PVC edge banding sits around 80°C Vicat B50 under ISO 306, with some grades specified at a minimum of 77°C. That number is not a failure threshold in service. It is a process and proximity threshold. The edge band softens as it approaches that range, and a softened band conforms less predictably under the pressure roller, which is why a glue line that looks perfect at the start of a shift can open up as the machine and the band heat up. More importantly, in a wet-area return next to a dishwasher or a boiling tap, the local surface temperature of the panel edge can climb well above the room ambient during cycles. The band does not need to melt to fail. It only needs to lose enough stiffness that the swelling core behind it wins the race.

A band with a higher Vicat point buys margin. A band with a lower Vicat point, run at the top of the glue-pot temperature window, on a board whose raw edge was never sealed, is a three-year failure waiting for a moisture event.

The mill case that changed how I specify the edge, not just the band

The case that shifted my thinking was not a catastrophic failure. It was a comparative water-immersion test on two identical PVC bands bonded to the same melamine-faced particleboard with the same EVA adhesive, run through the same edgebander at the same line speed. The only variable was the raw edge treatment: one set had the particleboard edge sealed with a primer before banding; the other did not. After 2 hours in 80°C water, the unsealed set showed visible edge lift and a whitened glue line. The sealed set held, with no measurable change in the bond line.

That result is not surprising once you stop treating the adhesive as the only load-bearing element. It is the reason I now write edge-banding specifications as a three-part requirement: a band with an adequate Vicat margin, an adhesive matched to the service environment (EVA for dry interiors; PUR where moisture exposure is plausible), and a sealed raw edge before the band goes on. On a 1.5 mm PVC band, a properly controlled line can hold a peel strength well above the common 35 N/50 mm baseline — one mill case reports 55 N/50 mm on a 1.5 mm PVC band with a controlled process. But that number is only meaningful if the substrate behind it stays dimensionally stable. A high peel strength PVC edge banding for kitchen cabinets is not the same claim as a moisture-stable edge system, and the field does not confuse the two.

What I check on the floor before blaming the band

When an edge lifts in service, the first question is not “which adhesive?” It is “where is the water coming from, and was that raw edge sealed?” I look for three things, in order:

The sink cutout and the dishwasher return. Water rarely lands on the door face. It lands on the horizontal edges and the panel ends near heat and steam. If the lifted band is clustered around those zones, the failure is moisture-driven core swelling, not adhesive selection.

The glue-line transfer pattern. On a delaminated section, if the adhesive is still mostly on the band with a smooth, glassy surface on the board side, the bond never fully wet out or was compromised at the interface. If the adhesive is split between both surfaces with a rough, fibrous residue, the core failed cohesively — the board moved under the glue line. That distinction tells you whether to change adhesive or change the edge-sealing step.

The Vicat value against the actual edge temperature. Infrared the panel edge at the worst-case cycle. If the measured surface temperature is within 10–15°C of the band’s Vicat point, the band has lost most of its stiffness margin, and the swelling core has a mechanical advantage it should not have.

None of this requires a new lab. It requires treating edge banding as the outer layer of a moisture-management system, not as a decorative strip with a peel rating. The band and the glue can both pass their tests and still lose the job, because the board edge was the weak link the whole time.