Peeling Polyurethane on the Clarifiers

This is amine blush (carbamate formation) causing intercoat delamination.

Mechanism: amine-adduct epoxy hardeners at the film surface react with atmospheric carbon dioxide and moisture to form ammonium carbamate - a waxy, water-soluble, alkaline residue. It is not visible in all light, but it destroys the chemical and mechanical bond the next coat needs.

Why here:

• 88% RH and steel only 3°F above dew point - moisture is plentiful at the surface; a 3°F margin is thin and condensation on a cold shell overnight is very likely. The lower third of the shell is the coldest, wettest zone, which matches where the peel is worst.
• 49°F steel is likely below or at the PDS minimum application temperature for the epoxy, so cure is slow and incomplete, extending the blush-forming window.
• No ventilation/dehumidification - still, humid air sits on the film.
• Recoat intervals exceeded: the 24-hour maximum recoat window on the data sheet is referenced to 75°F. Whatever the true window at 52°F, the contractor topcoated on an arbitrary 24-hour schedule rather than a measured cure state.

The polyurethane finish never bonded to the contaminated epoxy; the peel is purely intercoat adhesion loss.

Confirmation checks:

• Water-break / wetting test: water droplets bead and refuse to wet a blush-contaminated surface; on clean epoxy they spread and hold a continuous film.
• pH check: press pH paper against a water-wetted area - carbamate is alkaline, so a reading well above neutral supports blush.
• Solvent rub: blush rubs off as a soapy smear.
• Visual: pale haze or whitish bloom on the epoxy, especially on the cold lower shell.

Additional investigation before repair:

• Determine the actual failure plane with pull-off tests or a knife peel: is it only finish-to-intermediate, or has the intermediate also lost bond to the primer or to the steel? If the epoxy itself is undercured at 49°F, the whole system may be compromised and patching is pointless.
• Gather cure records: surface temperature and RH logs, product batch numbers, mix ratios.
• Check whether the finish failure includes CO2 bubbling/pinholing - moisture-cure urethanes and polyurethanes applied at 88% RH can also show bubbles, which points to an additional application defect.
• Sample low, middle and upper shell - the moisture gradient tells you whether this is a condensation problem (cold wall) or a plant-wide practice problem.

Corrective:

1. Remove the unbonded finish over the whole affected zone - do not spot-patch; delamination at the peel edge means the bond is failing beyond what you can see.
2. Wash the exposed epoxy with warm water and a mild detergent or the manufacturer's recommended cleaner, scrubbing, then rinse with clean water. Blush is water-soluble; washing, not sanding alone, removes it.
3. Abrade the washed surface to a dull, uniform profile, then re-check with a water-break test and allow to dry fully.
4. Re-apply the finish within the recoat window and verify adhesion on a test patch before releasing the rest of the shell.
5. If pull-off shows the intermediate is also undercured or debonded, remove the full system to bare steel.

Prevention / specification changes:

• Mandate a wash-and-abrade step for amine epoxies whenever the recoat window is exceeded, and write it into the specification as a hold point.
• Require heating and dehumidification so steel is at least 5°F above dew point and within the PDS temperature range, with logged readings at the point of application.
• Replace the arbitrary "24 hours" recoat timer with documented PDS recoat windows at the actual temperature, and add a cure check before topcoating.
• Consider a blush-resistant hardener grade for cold, humid work.