Removing Sealers and Curing Compounds
Most of what gets called removal here is not a separate operation at all. The question is whether the treatment sits above the depth your profile pass was cutting to anyway, and for one of the four there is nothing to remove.
Ask where the treatment sits before you ask how to remove it. A thin sealer or curing compound is above the depth your profiling pass was cutting to anyway, so it comes off as part of that pass rather than as a job before it. A penetrating sealer sits below that depth and has to be cut past, which means extra passes you cannot count in advance. A densifier cannot be removed at all, because it reacted into the concrete and there is no layer there. Only a heavy, built up film is genuinely a removal operation, and at that point it is the old coating problem rather than this one.
Removal is the wrong word for three of the four
The removal page on this site makes a point of saying that taking a coating off and profiling the slab underneath are two operations with two rates. That is true of an old epoxy floor and it is the usual reason a removal job loses money. It is not true here, and the difference is worth being precise about rather than carrying the rule across.
An epoxy floor is a build measured in mils, sitting on top of concrete you have not touched yet. A curing compound or a thin topical sealer is a fraction of that, and the profile a coating specifies requires cutting away cement paste to a depth that leaves the film far behind. The film is not an obstacle in front of the work. It is inside the work.
So the useful question is not which chemical or which machine. It is where the treatment sits relative to the cut you were making anyway, and how to tell if concrete is sealed is how you answer that before quoting. This page picks up from that diagnosis.
| What you found | Where it sits | What the job is | What to watch |
|---|---|---|---|
| Thin topical sealer or curing compound | On the slab, above the depth you were cutting to. | Not a separate operation. The profiling pass removes it. | A soft film blinds tooling, so the first pass runs slow and needs clearing. |
| Heavy or repeatedly recoated sealer | On the slab, but thick enough to behave like a coating. | A removal job in its own right, with its own rate and its own line. | Once it shatters, smears or has layers, it is the old coating problem. |
| Penetrating sealer | In the pore, commonly deeper than a profile pass reaches. | Cut past it. The pass count is discovered, not decided. | Penetration is uneven, so one clear test spot does not clear the floor. |
| Silicate densifier | Reacted into the concrete. There is no layer. | Nothing to remove. You are grinding harder concrete. | A tooling and rate problem, not a chemistry one. |
Nothing here is a depth measurement or a product specification. Penetration depth is a property of one product on one slab, which is precisely why the second case is quoted by testing rather than by planning.
The thin film case: one operation, one slower pass
This is the common one, and the good news is that it barely changes the plan. What it changes is how the first pass behaves, and misreading that is the mistake worth avoiding.
A soft film does not cut cleanly. It softens under the heat of the pass, packs into the gaps between the segments and blinds them, and the machine keeps moving while it has effectively stopped cutting. That is the same mechanism that makes a soft coating smear rather than shatter, described on the removal page, just at a much smaller scale. The tell is swarf that is gummy rather than powdery, and segments with material packed across their faces.
The response is to plan a slower first pass and to stop and clear the tooling rather than pushing through it. What you should not do is read the slow start as a hard slab and change your tooling for it. Loading and hardness feel similar from the handles and want opposite responses, and bond vs concrete hardness is the page that separates them.
The penetrating case: grind, test, repeat
A silane or siloxane is in the pore rather than on the surface, and how far in depends on how porous the concrete was and how much product went on. Both vary across a floor, and neither is written down anywhere you can read. So the honest position is that you cannot decide the pass count in advance, and pretending otherwise is how this job goes over.
What you can do is measure progress directly, using the same water drop test that found the problem. Grind a pass, vacuum properly, then drop water on the freshly ground surface. If it still beads, you are still in treated concrete. If it darkens and soaks in within seconds, you are through to clean substrate and the remaining passes are ordinary profiling.
Two disciplines make that loop trustworthy. Vacuum before you test, because grinding dust holds water and will fake a pass. And test several spots per pass rather than one, because the whole reason this case is hard to quote is that the depth is not uniform. One clear reading in the middle of a bay does not clear the bay, for the same reason the prep pillar says one good spot does not clear the room.
Price the passes you can predict, then test for the ones you cannot
A thin film adds no pass, it just slows the first one. A penetrating sealer adds passes you only discover by grinding and retesting. Model the profile you can plan for, then treat anything past it as a contingency you have told the customer about in advance.
The seeded rate assumes the tooling is cutting. A film that blinds the segments is not cutting, whatever the machine sounds like.
Open the live calculator →The densifier case: there is nothing to remove
This one is worth stating plainly, because people go looking for a product that will take a densifier off and there is not one. A silicate densifier reacted with free lime in the concrete and became hardened material inside the pore structure. It is not a layer resting somewhere. It is part of the slab.
So the work is grinding harder concrete, and every decision in it is a tooling decision. Expect a slower cut than the same floor untreated, expect tooling to last less time, and choose the bond for a hard, dense surface. Nothing about this stops you coating the floor and nothing about it is a defect. It simply costs more time than the untreated version of the same slab, and the only real failure mode is having quoted the untreated rate.
What does not work, and why
- Acid etching. Acid works by dissolving cement paste, and a film holds it away from the paste. On a sealed slab it is not a weaker option, it is not an option, and grinding vs acid etching covers what it does and does not reach
- Waiting for a dissipating compound. They break down under UV and traffic, and neither is spread evenly across a floor. Waiting gives you a slab that is clean where you will test it and sealed under the racking, which is worse than one that is sealed everywhere
- Strippers on open floor. The residue to clean up and the water you introduce are much the same whether the film was twenty mils or half a mil, so on a thin film the trade is poor. On a penetrant or a densifier there is no film to act on at all
- Testing only where it is easy. Every failure mode on this page is uneven by nature. The perimeter, the corners and anywhere that stayed covered are where the treatment survives and where the coating will let go first
Where a stripper genuinely is the answer, usually detail work or somewhere a machine cannot reach, the cautions are the ones the removal page sets out: residue left behind is a bond breaker under the new floor, and a stripped and rinsed slab has taken on water it now has to lose.
The moisture gate reopens, and that is the real risk
Everything above is a time and tooling problem, which means it costs money and can be planned for. This one can stop the job, and it is the reason a curing compound is not merely an inconvenience.
A curing compound is a film applied specifically to slow water leaving a slab, and it has been doing that for as long as it has been in place. An old topical sealer does much the same thing incidentally. Take either off and the concrete is free to release moisture it had been holding, so the slab you are handing to a coating is not the slab you tested.
The fix is sequence rather than product. Test moisture on the prepared surface, after the film is gone, and treat a reading taken through a sealer as no reading at all. An in situ relative humidity probe also needs time in the slab before it reads true, so this belongs in the schedule rather than in the last hour before coating. If the prepared slab comes back high, the answer is either time or a barrier, and moisture vapor barrier sizing covers what that costs. Gate four of the prep sequence is the only one of the four that more grinding cannot fix.
Where to go next
If you have not settled which treatment you are dealing with, how to tell if concrete is sealed sorts the four and names the test for each. If what you found turns out to be thick, layered, or failing in patches, it is a coating rather than a sealer and how to remove old epoxy and paint is the page for it.
For the profile the prepared slab has to reach once the film is off, CSP explained covers how open the surface needs to be and why over profiling is a failure too. For the whole sequence this sits inside, how to prep concrete for epoxy runs the four gates in order.
Frequently Asked Questions
Do I have to remove a sealer, or will the grinder take it off anyway?
For a thin topical sealer or a curing compound, the profiling pass takes it off, and that is the normal answer. A film like that is a fraction of the depth you have to cut to reach the profile a coating specifies, so it is gone well before you get there. Unlike an old epoxy floor, this is one operation rather than two.
What it does change is how that first pass behaves. A soft film does not cut cleanly. It softens, packs into the gaps between the segments and blinds them, so the machine looks like it is working while it has stopped cutting. Plan for a slower first pass and for stopping to clear tooling, and do not read the slow start as a hardness problem when it is a loading problem.
How do I know when I have ground past a penetrating sealer?
Use the water drop test as a progress check rather than as a diagnosis. Grind a pass, vacuum the area properly, and drop water on the freshly ground surface. If it still beads, you are still in treated concrete and you have not reached clean substrate. If it darkens and soaks in within seconds, you are through.
Run it in more than one spot, because penetration depth is not uniform: it follows how porous the concrete was and how much product went on, both of which vary across a floor. This loop is the reason a penetrating sealer cannot be quoted from a pass count decided in advance. You grind, you test, and the floor tells you when to stop.
Can I remove a densifier?
Not in any useful sense, and it is worth saying plainly because people go looking for a product that will do it. A silicate densifier is not a layer sitting somewhere. It reacted with free lime in the concrete and turned into hardened material inside the pore structure, so it is part of the concrete now. There is no film to lift and nothing for a stripper to attack.
So the job is not removal, it is grinding harder concrete, and it is a tooling question rather than a chemistry one. Expect a slower cut and choose the bond for a hard, dense slab. Nothing about this stops you coating the floor. It just means the profile costs more time than the same floor untreated would.
Will a chemical stripper work on a sealer or a curing compound?
It can soften an acrylic film, and on small detail work or somewhere a machine genuinely cannot reach that may be the only route. On open floor it is usually a poor trade, and the reason is scale. The residue you have to remove afterwards and the water you introduce are much the same whether the film was twenty mils of epoxy or half a mil of curing compound, so on a thin film you are adding nearly as much of a problem as you are taking away.
The two failures are the same ones any stripping job has. Softened film and stripper residue left in the surface are a bond breaker under the new coating, and a stripped and rinsed slab has taken on water it now has to lose. On a penetrating sealer or a densifier a stripper does close to nothing at all, because there is no film there for it to act on.
The curing compound is meant to break down. Can I just wait for it?
No, and this is the one that costs people a callback. Dissipating curing compounds are designed to break down under UV and traffic, which sounds like a schedule you could work with until you notice that neither UV nor traffic is evenly distributed across a floor. The middle of a lit, driven bay loses it. The perimeter, the corners, and everything under racking or stored material keep it, potentially for years.
So waiting does not give you a clean slab, it gives you a floor that is clean where you will test it and sealed where you will not. Prep as though the compound is fully present everywhere, and confirm on the prepared surface rather than on the schedule.
Does removing a curing compound change the moisture situation?
It can, and this is the part that turns a prep job into a delay. A curing compound is a film applied specifically to slow water leaving the slab, and it has been doing that job for as long as it has been there. Take it off and the concrete is free to release moisture it was previously holding.
The practical consequence is about sequence. Test moisture on the prepared surface, after the film is gone, rather than accepting a reading taken through it. An in situ relative humidity probe also needs time in the slab before it reads true, so this is a scheduling item rather than a same day check. A slab that read comfortably through a sealer can fail once that sealer is off, and finding that out before the coating goes down is much cheaper than finding it out afterwards.