What actually gets oil out of concrete?

Something that brings it back up. That is the whole test, and it is a more useful way to shop than reading strengths off a label. Oil in a slab is not a film sitting on top. It is inside the pore structure, which is why every page here says degrease before you grind. So the only products that remove it are the ones with a mechanism for moving it upward: a poultice, which holds an absorbent carrier in contact with the surface and wicks the oil back out as it dries, or a microbial or enzyme degreaser, which breaks the hydrocarbon down in place while the slab is kept damp. A surfactant cleaner, meaning the ordinary scrub and rinse degreaser, emulsifies what it can physically touch and rinses that away. On a fresh spill that is the right tool. On a slab that has been absorbing oil for years it cleans the top of the pore network and leaves everything below it, which is exactly the part that ruins a coating. The three differ by reach rather than by strength, so buying a stronger version of the one with the least reach does not help.

Will a solvent degreaser work?

It is the one to be most careful with, because it can leave the slab worse than it found it. A solvent dissolves oil, which sounds like removal but is not. Dissolving it also thins it, and thinned oil moves through a pore structure more freely than the original oil could. Unless something is there to absorb the mixture and carry it back to the surface, the solvent evaporates and the oil it mobilized settles deeper than where it started, below the depth any grinding pass is going to reach. That is the same failure as grinding a contaminated slab, arrived at from the opposite direction. Both the mechanical route and the obvious chemical route push the contamination further down. It is also a different problem from the solvent rag warning on the fisheyes page: that one is about spreading a contaminant sideways across a floor, this one is about driving it vertically into a slab. A solvent used inside a poultice, where an absorbent holds it in place and takes the mixture back up, is the version that works.

How long should a degreaser sit on the slab?

Long enough that the answer is measured in hours or days rather than in minutes of scrubbing, and this is the expectation worth setting before the job is quoted. Extraction is a dwell process. A poultice has to stay wet against the surface and then dry slowly, because the drying is what pulls the oil up, and covering it to slow that down usually makes it work better rather than worse. A microbial degreaser needs the slab kept damp for as long as it takes the organisms to consume the hydrocarbon, which is a multi-day timescale rather than an afternoon. Follow the dwell the product itself specifies rather than a general rule, because the carriers and the organisms differ. The practical consequence is scheduling. Effort does not substitute for time here, so a contaminated bay is not something to discover on the morning you planned to grind. Heavily soaked areas commonly need more than one cycle, and you find that out by testing after the first one rather than by deciding in advance.

How do I tell how deep the oil has gone?

Grind a small test patch and look at the freshly cut face, because the depth is the thing that decides whether this slab is a cleaning job or a specification decision, and nothing on the surface tells you. A cut through contaminated concrete shows a darker band, and where that band stops relative to the depth you have taken off is the measurement you actually need. Vacuum before you read it, because grinding dust obscures the color. Two cheaper tells are worth knowing. Oil rises in heat, so a bay that looks acceptable on a cold morning can show the staining again once the sun reaches it or once a grinding pass warms the slab, and that re-emergence is a genuine result rather than a nuisance. Old contamination is also rarely uniform: it follows where the vehicle or the machine actually sat, and it runs deeper along cracks and joints, which are the paths that were open to it.

How do I confirm the degreasing worked?

Not with the water drop test, and this is the trap, because that test is the standard check everywhere else in the prep sequence. A water drop reads the top of the surface and nothing below it. A surfactant clean produces precisely the state that fools it: a genuinely clean top layer over oil that is still in the slab. The floor darkens, the water soaks in, the test passes, and the contamination that will cause fisheyes is untouched. A clear water test after degreasing is not evidence, and it is the one place in the prep sequence where that test gives a confident wrong answer in the direction that costs money. Use the test patch instead, the same way the sealer page uses a grind and retest loop. Cut a patch, vacuum it, and read the cut face for a darker band. Do it in several places, because contamination is local by nature, and do it where the oil was worst rather than where the floor is easy to reach.

Degreasing Oil-Soaked Concrete

Every prep page on this site tells you to degrease before you grind. None of them says what that word covers, and the three products a contractor reaches for fail in three different ways. One of them makes the problem worse.

Quick answer

Choose a degreaser by how deep it reaches, not by how strong it claims to be. Oil is in the pore structure rather than on the surface, so the only products that remove it are the ones that carry it back up: a poultice, or a microbial degreaser given days rather than an afternoon. An ordinary scrub and rinse cleaner reaches the top of the pore network and no further. A solvent left to evaporate makes it worse, because thinning oil helps it travel deeper. And the water drop test cannot confirm any of this, because it only ever reads the top layer.

The site keeps saying degrease. This is what the word is hiding

Every prep page here gives the same instruction and gives it in one word. The prep pillar puts degreasing before grinding because oil sits below the surface and a machine drives it deeper. Grinding vs acid etching notes that acid does nothing through an oil film. The fisheyes page names oil as one of the two contaminants that cause the defect. All of that is a verdict, and none of it is a method.

The gap matters because the products sold for this job are not variations on one thing. They work by different mechanisms, they reach different depths, and two of the four take days rather than minutes. A contractor who buys the strongest label on the shelf usually buys the one with the least reach.

How the four degreasing methods differ.
Method	How it acts	How deep it reaches	What it costs you	Where it fails
Surfactant cleaner (scrub and rinse)	Emulsifies oil it can physically touch, then rinses it away.	The surface and the top of the pore network.	Minutes. The only one that is a scrubbing job.	Leaves a clean top over contaminated concrete, which is the state that passes a water test.
Solvent degreaser, applied and left	Dissolves the oil, which also thins it.	As deep as the thinned oil then travels, which is deeper than it was.	Fast to apply, and the damage is done as it evaporates.	Moves contamination down instead of out. The one method that can leave the slab worse.
Poultice (absorbent plus solvent)	Holds solvent in contact, then wicks the mixture back up as it dries.	Into the pore, limited by dwell and by how many cycles you run.	Hours to days per cycle. Drying slowly is what makes it work.	Rushed or uncovered, it dries before it has drawn anything back.
Microbial or enzyme degreaser	Breaks the hydrocarbon down in place while the slab is kept damp.	Wherever it stays wet long enough to keep working.	Days. Not something to start on grinding morning.	Allowed to dry out, or given an afternoon when it needed a week.

Reading across a row: the method, the mechanism it uses, how far into the slab that mechanism can act, the time it needs, and its characteristic failure. No dwell figures are given because they are a property of a particular product rather than of the category, so use the dwell the manufacturer specifies.

The solvent trap: thinning oil is not removing it

This is the finding worth carrying off the page, because the instinct it corrects is a strong one. Solvent dissolves oil, and a floor treated with it looks better immediately. Dissolving is not extraction.

Thinning oil lowers its viscosity, and thinner oil travels through a pore structure more freely than the original oil could. If nothing is present to absorb the mixture and carry it back to the surface, the solvent flashes off and leaves the oil it mobilized sitting deeper than where it started. Frequently that is below the depth any profiling pass will ever reach, which converts a cleaning problem into a specification problem.

That is the same failure as grinding an oily slab, reached from the opposite direction. The prep pillar already warns that a machine drives contamination deeper and smears it along its path. The obvious chemical shortcut does the same thing vertically. Between them they describe the two ways this job is usually made unrecoverable, and both feel like progress at the time.

It is also a distinct problem from the solvent warning on the fisheyes page, which is about a rag spreading silicone sideways across a floor. Here the direction is downward and the floor area is unchanged. Solvent inside a poultice, where an absorbent holds it against the surface and takes the mixture back up, is the version of this that works, and it is the difference between the third row of the table and the second.

Extraction is a dwell, not a scrub

The two methods that genuinely pull oil out of concrete both work by waiting. A poultice has to stay wet against the surface and then dry slowly, because the drying is what draws the oil upward into the absorbent, and covering it to slow that drying usually improves the result. A microbial degreaser needs the slab kept damp for as long as the organisms need to consume the hydrocarbon, which is measured in days.

Effort does not substitute for time in either case. Scrubbing harder, or buying a concentrate and applying it thicker, does not compress a dwell. Heavily contaminated areas commonly need more than one cycle, and the honest way to find that out is to test after the first one rather than to decide the number in advance.

That makes contamination a scheduling item rather than a morning task, which is the same shape as the finding on grinding vs acid etching: the chemical route is not the fast route. There it is the neutralizing, rinsing and drying that puts the job behind. Here it is the dwell. Either way, discovering an oil soaked bay on the day you planned to grind is a lost day at best.

The water drop test cannot confirm this one

The whole prep sequence on this site runs on the water drop test, so it is worth being blunt about the one place it fails. How to tell if concrete is sealed already covers oil as a false alarm: a contaminated slab beads water with no sealer on it. The inverse is the problem here, and it is the more expensive of the two.

A water drop reads the top of the surface and nothing below it. A surfactant clean produces exactly the state that satisfies it: a genuinely clean top layer sitting over oil that is still in the pore structure. The floor darkens, the water soaks in within seconds, the test passes cleanly, and nothing that will cause fisheyes has been removed. It is the only step in the prep sequence where the standard check returns a confident wrong answer in the direction that costs money.

So a clear water test after degreasing is not evidence of anything. What it tells you is that the surface is clean, which was never the question.

Read the depth off a test patch

The check that works is the one removing sealers and curing compounds uses for a penetrating sealer, pointed at a different question. Grind a small patch, vacuum it properly, and read the freshly cut face rather than the surface.

Contaminated concrete shows a darker band through the cut, and where that band stops relative to the depth you have removed is the measurement that actually decides the job. Vacuum first, because grinding dust obscures the color and will hide the band you are looking for. Run the patch where the oil was worst rather than where the floor is easy to get to, and run more than one, because contamination follows where a vehicle or a machine actually sat and is never uniform.

Heat brings it back. Oil rises as the slab warms, so a bay that looks acceptable on a cold morning can show its staining again once the sun reaches it or once a grinding pass heats the concrete. Treat that re-emergence as a result rather than as bad luck
Cracks and joints run deeper. They were the open paths, so contamination follows them further down than it goes through sound slab. Check them separately rather than assuming the field reading covers them
One clear patch does not clear the floor. The same discipline the prep pillar applies to the water test applies here, and for a stronger reason: oil is local by definition
Recoat a test patch if you are still unsure. Contamination repeats in the same spots, which is what separates it from a gas defect. The defect router sets out that test

When the slab cannot be degreased

Some floors will not come clean, and saying so early is worth more than another cycle of product. A bay that spent twenty years under a press or a leaking vehicle can hold oil deeper than any extraction will reach and deeper than grinding can economically remove. Buying more degreaser is the expensive way to find that out.

At that point this stops being a prep question and becomes a specification one. The honest options are a coating chemistry tolerant of a compromised substrate, cutting out and replacing the contaminated area, or declining the floor as it is written. All three are conversations with the customer rather than costs to absorb quietly, and all three are cheaper before the coating goes down than after it lets go.

There is also a moisture consequence that is easy to miss, because both methods that genuinely extract oil put water into the slab and keep it there for days. Test moisture on the prepared surface after the degreasing is finished and the slab has dried, not on the reading you took before it started. That is the same sequencing rule the sealer removal page applies to a curing compound, and if the prepared slab comes back high, moisture vapor barrier sizing covers what the answer costs.

Where to go next

For the sequence this step belongs to and the four gates a prepared slab has to pass, how to prep concrete for epoxy runs them in order. If the slab is beading water and you are not sure whether that is oil or a sealer, how to tell if concrete is sealed separates the two.

If the coating is already down and pulling away in round marks, the diagnosis is on fisheyes and craters, which is where uncleaned oil usually announces itself. For the other contamination that has to come off before profiling, removing carpet and tile adhesives covers the case where the material is knowable rather than hidden in the slab.