Identify Your Coating Defect
You do not need to know what it is called to find it here. Start with what the floor actually looks like, narrow it with one physical test, and land on the page that explains the cause and the repair.
Work from the floor, not from a word. Find the group below that matches what you can see, read the one line that separates its members from each other, then run the test in that line if looking is not enough. If you already know what the defect is called, the troubleshooting index lists every page by name and will be quicker. If none of the groups quite fit, or two of them do, the three question version also asks when the fault appeared and what the conditions were, and ranks what is left.
Two rules that settle most of it
Before the groups, the two observations that do the most diagnostic work. Both come from watching what a defect does rather than what it looks like, and between them they separate causes that look identical on a cured floor.
Gas moves. Contamination repeats.
- Recoat a small test patch and watch where the marks land
- New marks somewhere else: a gas defect. That air or vapor has already escaped, so the second coat gasses from wherever the slab is releasing next
- Same marks in the same spots: contamination. It is still sitting on the slab, and no number of extra coats will bury it
- This one test separates a moisture problem from a cleaning problem, and they have nothing in common except the shape they leave
For round marks, look at the middle
- There are only three possibilities, and each points somewhere different
- Closed dome: gas still trapped under an unbroken film
- Open puncture: the gas got out and the film was too stiff to close behind it
- Coating pulled back: nothing was punctured. The resin met a spot it could not wet
Match the floor to a group
Each group opens with the question it answers. Work down until one of them describes your floor, then use the tell to pick between that group's members.
Round marks in the film
Are there round marks in the coating, raised, open or bare?
These start as gas, as liquid, or as a spot of dirt. What separates them is whether the mark is a closed dome, an open hole, or a place the coating pulled away from, and for a dome, what comes out when you cut it: air is gas from the slab or the mix, liquid is water drawn up through the film by osmosis. One case sits outside all three: on a metallic floor, round open areas showing the coat beneath may be cells the installer induced on purpose, and those belong with the metallic effect rather than here.
Domes in the film that appeared weeks or months after the floor was finished, often in the same low or damp areas each time, and often growing slowly rather than arriving all at once.
- The tell: The dome is closed, like a bubble, but it turned up long after the coating cured rather than during it. Bubbles are born in the wet film; these are not.
- Confirm it: Cut one open and catch what comes out. Liquid, rather than air, is the whole diagnosis, and it is often slightly soapy or alkaline to the touch.
Raised round domes in the film, from pinprick size up to a coin, sometimes in drifts across one bay and sometimes over the whole floor.
- The tell: The dome is closed. Press it and it gives, because the gas is still trapped under a film that never opened.
- Confirm it: Puncture one with a blade. If air escapes from a film that was intact until you cut it, it is a bubble and not a pinhole.
Small open holes through an otherwise flat film, closer to the surface of a cracker than to a blister. Easiest to see with a light held low and raked across the floor.
- The tell: The mark is an open puncture with coating right up to its edge. The gas got out, and the film was too stiff or too thin to close behind it.
- Confirm it: Draw a fingernail across one. A pinhole catches as a pit in a level film, where a bubble stands proud of it.
Circular bare or thinned spots with a raised rim, as though the coating was pushed away from a point. Usually visible within minutes of laying the coat, long before it gels.
- The tell: Nothing punctured the film. The coating pulled back from a spot it could not wet, so the center is thin or bare rather than holed.
- Confirm it: Recoat a test patch. Contamination repeats in the same spots because it is still on the slab, where a gas defect moves, because that air has already gone.
The color or clarity is wrong
Is the film sound but the color or pattern wrong, whether uneven, too flat, cloudy, hazy or ambered?
The film is intact and stuck down. Start with what the floor was specified to look like, because that decides which half of this family you are in: on a floor meant to be one solid shade any variation is the fault, while on a metallic floor the variation is the product and the fault is that there is too little of it, or that it opened up. Among the solid colors, what separates them is when the color went wrong and where it sits: a shade that shifted after cure maps to where the light falls, a haze lying on the surface washes off while one cured into the film does not, and unevenness that went down with the material is cured into the film and runs all the way through it.
Bands, swirls or blotches of lighter and darker shade across a floor that was meant to be one solid color, often strongest where one kit ran out and the next began.
- The tell: The color is uneven rather than wrong, and it was uneven from the moment the material went down. Nothing about it changed after the floor cured.
- Confirm it: Scratch through a light area and a dark area to the substrate. If both are the same film and the shade difference runs all the way through, the pigment is unevenly distributed rather than something sitting on top.
A metallic floor that came out flat and lifeless instead of deep, or blotchy and muddy instead of flowing, or peppered with open craters that go through to the coat underneath.
- The tell: The floor was never meant to be one shade, so the fault is not that it varies. It is that the variation is too weak, at the wrong scale, or torn open. On a solid color floor the same appearance would be a different page entirely.
- Confirm it: Hold the floor against the sample board the client approved. A metallic floor has no uniform standard to measure against, so the board is the only reference that settles whether what you are looking at is a defect or the product doing what it does.
A polyaspartic topcoat that dried milky, pearlescent or white rather than clear, often worst in the coolest part of the floor or wherever the air was dampest while it cured.
- The tell: The haze is inside the film rather than on it. It does not wipe off, it does not wash off, and it was there as soon as the coat cured rather than appearing later.
- Confirm it: Wash a test patch with warm water and a scrub pad. Amine blush lifts and takes the haze with it. If the patch comes up clean and still cloudy, the whiteness is cured into the coating and this is the page.
A cloudy, milky or greasy film lying on the surface, often worse near doorways, cold walls and anywhere the air was damp while the floor cured.
- The tell: The haze sits on the film rather than in it. Scratch a corner and the coating underneath is sound and the color it should be.
- Confirm it: Wash a test patch with warm water and a scrub pad. Amine blush is water soluble and lifts, where a defect cured into the film does not.
An amber or yellow cast across a floor that went down clear or white, usually strongest wherever daylight lands longest.
- The tell: The color shifted after the floor cured, and it maps to light. Shadowed edges and covered areas stay closer to the original shade.
- Confirm it: Move a mat or a rack that has sat in place for weeks. A sharp line at its edge is UV; an even shift across the whole floor is not.
The coating is coming off
Is the coating lifting, flaking or peeling away from what it was laid on?
What separates these is what the underside brings with it and where it happened: bare slab means the bond to the concrete let go, coating on the underside means one coat let go of another, and damage only where a vehicle parks is a third thing again.
Bare concrete showing through in patches about the size and shape of where the tires sit, sometimes with the tread pattern still readable along the edge of what is missing.
- The tell: It is only where a vehicle parks. Walkways, the middle of the bay and the floor under shelving are untouched, and the coating that is missing is usually still stuck to the tire.
- Confirm it: Look at the tire as well as the floor. Coating transferred onto the tread settles it, and the slab exposed underneath will be clean and sound rather than powdery, chalky or damp.
The top coat lifting, sheeting or flaking away while the coat underneath stays put, often peeling in wide thin sheets that keep their shape as you pull them.
- The tell: Lift a flake and read its underside. Coating rather than concrete means one coat let go of another, and the layer still on the floor is usually sound and well stuck.
- Confirm it: Score a small square through the top coat only and lift a corner with a blade. If it releases cleanly and the coat beneath is intact and glossy, the bond that failed was between coats rather than to the slab.
Coating lifting, flaking or coming away in sheets, sometimes releasing across a whole bay once a scraper gets under an edge.
- The tell: Lift a flake and read its underside. Bare concrete or loose dust means the bond to the slab never formed, which puts the cause in the prep.
- Confirm it: Score a small square through to the concrete and lift a corner. Clean release off the slab is a prep failure, not a coating failure.
It set too slowly or too fast
Did the coating set on the wrong schedule, either still soft long after it should have hardened, or going off before you could spread it?
What separates these is which way the clock ran. Too slow is a fault in the chemistry, and there uniformly soft points at the ratio while soft in patches points at the mixing. Too fast is usually not a fault in the material at all but the temperature of what it was poured onto, which compresses every stage of the cure at once rather than spoiling any one of them.
A floor still soft, gummy, rubbery or taking a fingerprint long after the data sheet said it would be walkable, sometimes with a greasy or wet feel that never dries.
- The tell: Uniformly soft across the whole pour points at the ratio. Soft in streaks, swirls or patches with hard material between them points at the mixing.
- Confirm it: Press a thumbnail into an out of the way spot. Material that dents and stays dented has not cured. Material that is only tacky on top but hard underneath is a surface problem, so read the blushing page instead.
Polyaspartic that thickens in the bucket before the batch is out, pulls into ropes and drags under the squeegee, or sets so fast that the edge has gone off before you reach back to it.
- The tell: The material stopped being workable, rather than the finish coming out wrong. If the floor laid down normally and only shows banding at roller width, that is the applicator rather than the clock, and it is a different page.
- Confirm it: Take the temperature of the slab, not the air, and compare the pot life your product is rated for against what you actually got. A slab in sun runs well above the air around it, and every stage of the cure scales with the slab.
The surface texture or pattern is wrong
Is the film sound and the right color, but the surface itself wrong, whether rough, ridged, patchy, unevenly covered or slick underfoot?
Every member here has a film that is whole: if it is split or crazed you want the cracks family instead. What separates these is where the texture came from: banding at the spacing of the passes follows the applicator, roughness you can feel particles in came from the material or from what landed in it, and a bare patch in a decorative layer is an absence rather than a texture at all. One member sits outside all of that: a floor that is merely too smooth has no defect in it anywhere, and what is wrong is that nothing was ever put in the coating to give it grip, which makes it a specification that was missed rather than a fault in the work.
Stripes, ridges or bands of differing sheen running the width of the floor, usually at the spacing of the roller and usually most obvious in raking light or from a low angle.
- The tell: The pattern follows how the floor was worked rather than anything in the slab. Marks run parallel, at roller width, and line up with where each pass started and stopped.
- Confirm it: Look along the floor from a low angle rather than down at it, and check whether the bands line up with the edges of the passes. A mark that maps to the applicator is a technique mark, not a material fault.
Thin or bare patches in a flake floor where the base coat shows through, or a floor that reads as blotchy because the flake is dense in some areas and sparse in others.
- The tell: Nothing is wrong with the film. What is wrong is what is missing from it, and the pattern usually maps to where somebody stood, how far they could throw, or where the base coat had already started to set.
- Confirm it: Look at the edges of a thin area rather than its middle. A broadcast problem fades gradually from dense to sparse, where a spot that lost its flake later has a defined edge and loose flake around it.
A finished floor that feels like fine sandpaper, or has scattered hard specks and bumps in an otherwise glossy clear coat. Usually felt before it is seen.
- The tell: The film is continuous and the right color. What is wrong is that something is in it or on it, or that the coat was too thin to bury what it went over.
- Confirm it: Run a gloved hand across it, then a blade. Grit that scrapes off was lying on the surface, grit that stays is cured into the film, and roughness that maps to the flake or the profile underneath was never buried in the first place.
A finished floor that looks exactly as it should, and is dangerously slick the moment anything wet reaches it. Often reported after the first mop, the first rain or the first spill rather than at handover.
- The tell: There is no defect to find. The film is sound, bonded, the right color and the right thickness, and the complaint is entirely about what happens underfoot.
- Confirm it: Wet a small area and walk it in the shoes people actually wear there. If the floor is fine dry and treacherous wet, nothing has failed: the floor has no slip resistance in it, because none was ever specified.
Slippery Epoxy Floor: How to Add Grip to a Floor That Already Cured
The film is cracked or split
Is the coating itself broken open, whether one line running across it or a fine crazed pattern all over?
What separates these is whether the crack started above or below. A single line that runs the length of the floor and reappears in the same place after every repair is a crack in the concrete underneath, coming up through a coating that is otherwise sound and stuck down. A crazed or wrinkled pattern in no direction anybody applied never involves the slab at all: that is the film moving against itself while it cured, and the tell is that it stops at the edges of whatever was coated rather than continuing under them.
Ridges, creases or a fine crazed pattern like hide across the surface, running in no direction anybody applied. Sometimes a whole bay, sometimes only where the sun fell on it.
- The tell: The ridges follow nothing you did. They are not at roller spacing, there is nothing sitting in the film to feel, and the coating is still stuck down.
- Confirm it: Press a ridge and cut through one with a blade. A skin over softer material underneath means the surface set while the body was still moving, which is the whole of this defect.
A crack line showing through a finished floor, usually in exactly the place a crack was filled before coating. Sometimes visible as a hairline within days, sometimes opening into a split you can feel months later.
- The tell: The coating is sound and firmly stuck down everywhere including right up to the line. Nothing has lifted, blistered or peeled. There is simply a crack running through it, following something underneath.
- Confirm it: Look at where it runs. A telegraphed crack follows the slab, so it carries on in a straight line across bays, runs to a wall or a column, and reappears in the same place after every repair. A defect in the film itself stops at the edge of whatever was coated.
Before you repair anything, write down the conditions
Most coating failures are conditions failures, and the conditions are the one piece of evidence that disappears completely if nobody notes them. A data sheet states limits for temperature, humidity and the recoat window, and a floor cannot be checked against those limits from memory. Record the date and time of the coat, the air and slab temperature, the relative humidity, the product and batch, the mix ratio and pot life, and how long the coat sat before the next one went on. Photograph the defect with something in frame for scale.
Where more than one defect is present, and that is common rather than unusual, work in order of what has to be fixed first. Anything to do with the slab, meaning moisture, contamination and preparation, comes before anything to do with the film. A perfect topcoat over an unresolved slab problem fails again on the same schedule the first one did.
Frequently Asked Questions
How do I tell a gas defect from a contamination defect?
Recoat a small test patch and watch where the marks land. This one test separates most of the round defects, and it works because the two causes behave differently once you disturb them.
A gas defect moves. The air or vapor that caused it has already escaped, so the second coat gasses from wherever the slab happens to be releasing next, and the new marks appear somewhere else.
Contamination repeats. Silicone, oil or a release agent is still sitting on the slab, and the fresh coat pulls away from exactly the same spots it pulled away from the first time. Marks in the same places mean you have a cleaning and preparation problem rather than a moisture problem, and no amount of extra coats will bury it.
The floor has round marks. How do I tell which one I have?
Look at what the middle of the mark is doing. There are only three possibilities and each points at a different page.
A closed dome that gives when you press it is a bubble: the gas is still trapped under an unbroken film. An open puncture through an otherwise flat film is a pinhole: the gas got out, and the coating was too stiff or too thin to close behind it. A circular bare or thinned spot with a raised rim, where nothing was ever punctured, is a fisheye: the wet resin met something it could not wet and retreated from it.
Bubbles and pinholes share their causes and need different fixes. Fisheyes share neither, which is why the three way test is worth doing before you decide anything else.
Why is this page separate from the troubleshooting index?
Because they answer different questions. The index lists every defect by name, which is the fastest route if you already know the word for what you are looking at.
This page assumes you do not. It is organized by what you can see from standing height, so it works when all you have is a floor with something wrong on it and no vocabulary for it yet. If you know the name, the index is quicker.
What should I record before I start repairing anything?
The conditions and the material, because both get lost within a day and both decide whether a repair holds.
Write down when the coat went down, the air and slab temperature and the relative humidity at the time, which product and batch it was, what the mix ratio and pot life were, and how long it sat before the next coat. Photograph the defect with something in frame for scale.
Most coating failures are conditions failures, and conditions are the one piece of evidence that disappears completely if nobody notes it. A data sheet states limits for temperature, humidity and recoat window, and you cannot check a floor against those limits from memory.
Can more than one defect be present at once?
Often, and it is worth expecting rather than being surprised by. A slab pushing moisture can bubble the base coat and peel it later, and the same damp conditions that drive bubbling also encourage blush on the topcoat.
Work through them in order of what has to be fixed first. Anything to do with the slab, meaning moisture, contamination and preparation, comes before anything to do with the film, because a perfect topcoat over an unresolved slab problem fails again on the same schedule as the first one did.