Removing Carpet and Tile Adhesives
Unlike an unknown old coating, you can usually find out what this is, and here it changes the answer. Soft adhesive and black cutback smear and blind tooling. Thinset is not an adhesive at all, and grinds like the concrete it is made of.
Find out what it is, and here you usually can. The floor covering that came up names the material underneath it, and that changes the tooling rather than merely the schedule. Soft carpet adhesive and black cutback are polymer and bitumen: they soften, smear, and blind the segments, so they want tooling that shears and a pass that stays cool. Thinset under ceramic tile is not an adhesive at all, it is a cement mortar, and it grinds like the slab. On any pre 1980s tile floor, settle the asbestos question before a machine touches it, because the worst smearer and the likeliest asbestos are the same material.
This is the one removal job where the material is knowable
The advice on how to remove old epoxy and paint is that you usually cannot discover what an old coating is and that it rarely matters, so you classify it by how it behaves when you attack it. That is right for a coating. It is the wrong instinct here, and the difference is the reason this page exists separately.
Adhesive comes with a clue nobody has to guess at: whatever was on top of it. Carpet, old vinyl tile, ceramic tile and modern LVT each sit on a different material, and you generally know which one came up because somebody just took it off. That is worth using, because two of the four want opposite tooling from the third, and bringing the wrong set costs a morning rather than a percentage.
| What came up | What is stuck to the slab | How it behaves | What works |
|---|---|---|---|
| Carpet or sheet vinyl | Soft adhesive, often water based | Softens and smears. Packs into the tooling and blinds it. | Shearing tooling, light passes, stop and clear often. |
| Old vinyl or asphalt tile | Black cutback mastic, asphalt based | Thermoplastic. Gets softer as the pass heats it, so pushing harder makes it worse. | PCD or a scarifier. Keep the heat down. Settle the asbestos question first. |
| Ceramic or porcelain tile | Thinset mortar, a cement | Hard and brittle, and left in notched ridges rather than an even layer. | Aggressive metal bond tooling. Treat it as concrete, because it is. |
| Modern LVT or vinyl plank | Acrylic pressure sensitive adhesive | Thin, and often designed to release. Usually the easiest of the four. | Frequently comes up with a scraper, then a normal profiling pass. |
A floor can carry more than one of these, because buildings get refitted. A cut through to the slab at a doorway or a former partition line often shows two eras of flooring stacked up, and the machine will meet them one after another.
Why black mastic gets worse the harder you push
Cutback is asphalt based, which makes it thermoplastic: it softens with heat. Grinding is a heat generating process. So the intuitive response to a pass that is not removing much, which is to push harder, puts more energy into the film, softens it further, and packs it deeper into the gaps between the segments. The machine sounds like it is working and it has stopped cutting entirely.
This is the smearing case described on the removal page in its most extreme form, and the response is the same in kind: tooling that shears rather than abrades, usually polycrystalline diamond, or a scarifier where the build is heavy enough to plane. What is specific to cutback is the heat. Lighter passes and stopping to clear the segments beat a heavy pass, which is the opposite of what works on a hard slab and is why bond vs concrete hardness is worth reading alongside this: the tooling that shears a soft adhesive is not the tooling that cuts a cement mortar.
Thinset is a cement, and treating it as glue is the expensive mistake
Thinset is a cement mortar with polymer additives. Nothing about its behavior under a machine belongs with adhesive, and the word is what leads people to bring soft bond, shearing tooling to it and then wonder why the floor is barely marked after an hour.
Treat it as concrete: hard, brittle, and removed by aggressive metal bond tooling that cuts. What makes it genuinely awkward is not the material but the shape. Thinset is troweled with a notched trowel, so it is left in ridges rather than as an even film. You are flattening a corrugated surface, which means the high points come off long before the low points, and a pass that looks finished from standing height still has adhesive sitting in the valleys. Check it at a low angle with a light across the floor rather than from above.
The asbestos question, and the convergence worth knowing
On a pre 1980s building this is settled before anything disturbs the floor, not after. Resilient tile of that era and the black cutback mastic under it were both common asbestos containing materials, and they are usually found together, because one was bedded in the other.
That produces a convergence that is worth stating plainly. The material that behaves worst under a grinder, and therefore most tempts somebody to attack it harder and drier, is the same material most likely to contain asbestos. The instinct the previous section warns against on production grounds is the same instinct that turns a survey question into a contamination event.
If the age and history of the building make it plausible, have it tested by a qualified party before you quote, let alone before you grind. A positive result does not add a step to the job. It changes what the job legally is, who is permitted to carry it out, and how the waste leaves site. That is a scoping question rather than a site question, and finding it out at quote time is the difference between a job you decline cleanly and a job you have to stop halfway through.
Stained afterwards is not the same as contaminated
Adhesive that has soaked into the pore commonly leaves a shadow, and no amount of further grinding removes it without taking away more concrete than the job needs. Whether that matters depends on a distinction worth being clear about with the customer before you start.
- Color that is dry, hard and part of the surface is cosmetic. Under an opaque coating it is invisible and it does not affect the bond
- Residue that is still soft is a bond breaker, whatever it looks like. The new floor would be bonded to adhesive rather than to concrete
- The test is the usual one. Drop water on the prepared area: if it darkens and soaks in within seconds you have concrete, whatever color it is. If it beads, something is still on there
- Under a clear or decorative finish the appearance question returns, because the shadow will show through. That belongs in the conversation at quote stage rather than on the day
How to tell if concrete is sealed covers that water test properly, including the two ways it can mislead you and the one treatment it cannot see at all.
Removal is still not a profile
As with any removal job, getting the adhesive off does not leave a surface a coating can bond to. Removal tooling is built to take material away, not to open the slab evenly, and what it leaves is uneven, often gouged where a scarifier or PCD tracked, and nowhere near the profile the next product specifies.
So the same rule holds here as on the coating removal page: take it off, then prepare the slab underneath exactly as you would prepare any bare slab, and price the two separately. How to prep concrete for epoxy covers what that second stage still has to satisfy, including the moisture gate, which a floor covering may have been masking for as long as it was down.
Where to go next
For the general version of the behavior classification used here, how to remove old epoxy and paint sets out the three ways a material responds to being attacked and why a well installed floor is the expensive one. For the tooling decision underneath both, bond vs concrete hardness explains why one set will not do both jobs.
Once the slab is bare, how to prep concrete for epoxy runs the four gates in order, and CSP explained covers how open the surface has to be and why over profiling is a failure too.
Frequently Asked Questions
Do I need to know what the adhesive is, or can I just grind it?
Here you should find out, and that is the opposite of the advice for an unknown old coating. With a coating the chemistry is usually unknowable and rarely matters, so you classify by how it behaves under attack. With adhesive you have a strong clue nobody has to guess at: what was on the floor. Carpet, old vinyl tile, ceramic tile and modern LVT each sit on a different material.
It matters because the two commonest cases want opposite tooling. Soft carpet adhesive and black cutback are polymer and bitumen: they soften, smear, and pack into the gaps between the segments until the machine is not cutting at all. Thinset under ceramic tile is a cement mortar, so it is hard and brittle and grinds much like the slab. Bring soft-adhesive tooling to thinset and you will barely mark it. Bring thinset tooling to cutback and you will glaze it in a few minutes.
Why does black mastic get worse the harder I grind?
Because it is thermoplastic. Cutback is asphalt based, so it softens with heat, and grinding is a heat generating process. Pushing harder puts more energy into the film, which softens it further, which packs it deeper into the tooling. The machine sounds like it is working and it has stopped removing anything.
That is why aggressive grinding is the wrong instinct on cutback specifically. What works is tooling that shears rather than abrades, usually polycrystalline diamond, and keeping the heat down: lighter passes, and stopping to clear the segments rather than pushing through. It is the smearing case described on the coating removal page, in its most extreme form.
Could the adhesive under an old tile floor contain asbestos?
Yes, and on a pre 1980s building this should be settled before anything touches the floor. Both the resilient tile itself and the black cutback mastic under it were common asbestos containing materials, and the two are usually found together.
This produces a convergence worth knowing about: the adhesive that behaves worst under a grinder, and therefore most tempts somebody to attack it harder, is the same one most likely to contain asbestos. If the age and history of the building make it plausible, have it tested by a qualified party. A positive result does not add a step to your job, it changes what the job legally is, who is permitted to do it, and how the waste leaves site. Grinding first and asking afterwards turns a survey question into a contamination event.
Is thinset really not an adhesive?
Not in any way that matters to your tooling. Thinset is a cement mortar with polymer additives, so its behavior belongs with concrete rather than with glue, and calling it adhesive is what leads people to the wrong tooling and a lost morning.
Treat it as concrete: it is hard, it is brittle, and it comes off with aggressive metal bond tooling that cuts rather than shears. What makes it awkward is not the material, it is the shape. Thinset is left in ridges under the trowel notches, so you are not removing an even layer, you are flattening a corrugated one, and the high points come off long before the low points do.
The adhesive is off but the slab is still stained. Is that a problem?
It depends entirely on whether what is left is soft or merely colored, and those need different responses. Adhesive that has soaked into the pore can leave a shadow that no amount of grinding removes without taking off more concrete than the job needs. Under an opaque coating, color that is dry, hard and integral with the surface is cosmetic and does not affect the bond.
Residue that is still soft is a different matter and is a bond breaker. The test is the one used everywhere else in prep: drop water on the prepared area. If it darkens and soaks in within seconds you have concrete, whatever color it is. If it beads, something is still on there and the coating will bond to that instead. Under a clear or decorative finish the appearance question returns, and that is worth agreeing with the customer before you start rather than after.
Are chemical adhesive removers worth using?
On small areas, in detail work, or where a machine genuinely cannot go, sometimes. Across an open floor they usually trade one problem for a worse one. A softened adhesive is not a removed adhesive, it is a wet mess that still has to be scraped up and disposed of, and what is left behind in the pore is exactly the bond breaker you were trying to eliminate.
The two structural failures are the same ones any stripping job has. Residue has to come out completely or the new floor is bonded to it, and a stripped and rinsed slab has taken on water it now has to lose before it can be coated. On an asbestos suspect floor a remover is also not a way around the survey, because the risk is in the material rather than in the method.