Epoxy Self-Levelling on a Damp Slab over Old Alkyd: Blisters at 36 Hours
A 96 m² workshop floor (12 m × 8 m) in a coastal town was previously sealed with a solvent-borne alkyd gloss. The floor was swept, degreased with a water-based alkaline detergent, hosed down and left to dry overnight. The weather was humid: 22–26 °C with 70–80% relative humidity and no forced ventilation. Next morning the surface looked dry to the touch, so the applicator poured a solvent-free two-pack epoxy self-levelling coat at about 1.5 mm wet thickness directly onto the old alkyd, with no primer. Each 5 kg kit was mixed for 2 minutes and poured immediately.
Thirty-six hours later the floor shows blisters from 5 to 20 mm across, mostly in the centre of the room and over old oil stains. Where a blister is cut open, the epoxy peels off in flexible sheets, the underside is wet and whitish, and the old alkyd underneath is still glossy. In other areas the film is hard but carries a faint greasy bloom that does not wash off with water.
1. Identify the failure mechanisms visible in the evidence and rank them by how much of the floor they explain.
Read the worked answer
1. Water vapour pressure from the slab (dominant). Blisters concentrated in the room centre, liquid water and a whitish deposit on the underside of the lifted film, and a floor that was hosed and only 'dry to the touch' after one humid night all point to moisture trapped under a virtually impermeable solvent-free epoxy. The film has no way to release vapour, so osmotic and vapour pressure lift it. The room centre is the last place to dry, which is exactly where the blisters are.
2. No key and no primer on a glossy alkyd. A solvent-free epoxy bonds mainly by wetting and mechanical key. Applying over a hard, glossy, release-agent-free-but-still-slick alkyd film gives adhesion only at the interface; once water pressure develops, the weak link is the epoxy/alkyd boundary, so the epoxy peels as a sheet and the alkyd stays glossy - which is what was found. A primed, abraded system would have failed cohesively instead.
3. Contamination. Blistering over the old oil stains is contamination wicking through. Alkaline detergent residues also attack adhesion.
4. Amine blush (the greasy bloom). Amine-cured epoxy at 70–80% RH and cool temperature forms surface amine carbonate; it is not removed by plain water and will destroy adhesion to any next coat.
5. Application errors. Two minutes' mixing with no induction or degassing time traps air and leaves unmixed resin; a single 1.5 mm pour on a non-porous substrate adds shrinkage stress. These worsen the picture but do not by themselves explain the water under the film.
2. What tests and records should be required before any rework, and what acceptance limits would you set?
Read the worked answer
Moisture in the slab. Use the plastic-sheet test (a 1 m² sheet taped down for 16–24 hours; condensation or a dark damp patch means the slab is still emitting moisture), and confirm with a calibrated moisture meter or an in-situ relative humidity probe. Set the limit from the coating datasheet; for a solvent-free epoxy on a slab a common working rule is no more than about 4% moisture content by meter, or below 75–80% equilibrium RH, and no visible damp under the sheet. Hosing must stop at least several days before application and the slab must be dried, not just surface-dried.
Dew point. Record air temperature, relative humidity and slab temperature with a sling psychrometer or hygrometer and a contact thermometer at the start of each shift. Substrate must be at least 3 °C above the dew point - on a humid coastal morning this is frequently the binding constraint, not the calendar.
Adhesion of the existing coating. Cross-hatch or pull-off test on the old alkyd: if it lifts, it must be removed, not overcoated. A solvent rub test tells you whether it is a sound, fully cured film or a soft, wax-contaminated one.
Contamination. Water-break test and a solvent wipe to detect oil; the alkaline detergent must be rinsed with clean water and the pH checked.
Trial patch. Lay a 1 m² patch to the intended specification in the worst area (centre and over the stains) and assess after 7 days: no blisters, hard film, good adhesion. Keep dated records and photographs - this file is the evidence if the client disputes the cause.
3. Write the corrective specification and the application controls that would have prevented this failure.
Read the worked answer
Strip and prepare. Remove all failed and suspect epoxy by grinding or light shot blasting. Remove the old alkyd completely - overcoating an unknown gloss alkyd in a wet area is not worth the risk. Degrease, rinse, and dry to the moisture limits above. Fill cracks and blowholes, and grind the surface to a mechanical profile.
Moisture control. If the slab continues to emit moisture, apply a suitable moisture-suppressing or surface-tolerant epoxy primer at the manufacturer's rate, or install a designed damp-proof membrane - a primer is not a substitute for a slab that is actively wet. Alternatively dry the room with heaters and forced ventilation and verify with the plastic-sheet test.
System. Surface-tolerant epoxy primer, then the self-levelling epoxy in two coats of about 0.75 mm each rather than one 1.5 mm pour, with the second applied inside the recoat window. Where a primer is not used, at minimum abrade and solvent-wipe the substrate and check the datasheet for direct-to-substrate approval.
Application controls. Mix with a timer to the exact ratio, scrape the sides of the pail, then allow the stated induction or sweat-in time (commonly a few minutes) before pouring; keep pot life in mind, since a 30 °C floor shortens it sharply. Use a spike roller to release air. Maintain air and substrate temperature above the minimum and substrate at least 3 °C above dew point, with relative humidity inside the datasheet range - dehumidify or ventilate if needed. Remove amine blush by washing and abrading before any subsequent coat, and protect the fresh floor from water for the full cure time stated on the datasheet.