Last Updated on September 26, 2026 by Umar Farooq
Removing antifreeze stains is the easiest part of dealing with an antifreeze spill, and treating it as the whole job is how the dangerous part ends up spread across thirty square feet of driveway instead of sitting in one place where you can pick it up. Coolant is water-miscible. It comes off concrete far more readily than oil does, and that is the one piece of straightforwardly good news on this page. The catch is that the thing taking it off is water, and water does not destroy ethylene glycol. It relocates it.
So three rules, and between them they are the article. Get something absorbent onto it before anything wet touches it. Keep the hose off it for as long as you can stand to. And expect the mark to vanish well before the hazard does — which is backwards from every other stain on this site, where the mark outlives the problem that made it.
Four things this page deliberately does not do twice. The four stain families, the spot tests and the pre-wet, dwell, agitate, rinse sequence that every concrete stain shares are set out in cleaning stains from concrete, and nothing about a water-soluble spill changes them. Whether the runoff may legally go on the grass, down the gully or nowhere at all is decided by your own municipality, and that chain is worked out in pressure washing runoff rules. What a wash solution does to a planting bed, and the difference between diluting something and diverting it, belongs to protecting plants when pressure washing. And if you are not yet certain which of the five fluids a garage slab collects you are looking at, the identification table is in cleaning garage floor concrete.
A Cat’s Lethal Dose Is Measured in Teaspoons
Start with the number, because it is the reason this page is ordered the way it is.
The Merck Veterinary Manual’s toxicology section gives it plainly: “The minimum lethal dose of undiluted ethylene glycol is 1.4 mL/kg in cats, 4.4–6.6 mL/kg in dogs, 7–8 mL/kg in poultry, 2–10 mL/kg in cattle, 1.6 mL/kg in macaques, and 6.61 mL/kg in guinea pigs” (Merck Veterinary Manual, Ethylene Glycol Toxicosis in Animals). It adds that most cases involve antifreeze, “in which concentrations of ethylene glycol up to 95% are common,” and that younger animals may be more susceptible.
Run that through a household. A four-and-a-half kilogram cat is at 6.3 mL of undiluted ethylene glycol, which is about a teaspoon and a third — that conversion is my arithmetic from Merck’s figure, not a published dose. Out of a jug of prediluted 50/50 coolant it is roughly two and a half teaspoons of liquid. That is a small enough volume that “it was only a splash” and “it was a lethal dose” are not mutually exclusive statements.
Now open the bottle’s own paperwork, because section 3 of a safety data sheet is where a composition is stated rather than implied. Prestone’s sheet for its Antifreeze/Coolant with Leak Detect lists ethylene glycol at 40 to under 55 per cent by weight, diethylene glycol at 1 to under 5 per cent, and 2-ethylhexanoic acid at 1 to under 5 per cent, with the note that “Exact percentage of ingredients is withheld as a trade secret.” The three ingredients it names for labelling purposes are ethylene glycol, 2-ethylhexanoic acid and diethylene glycol, and the label carries “Harmful if swallowed”, “May damage the unborn child” and “May cause damage to organs through prolonged or repeated exposure.”
Two glycols, then, not one. Everybody writing about this treats “antifreeze” as a synonym for ethylene glycol, and the sheet in front of me says there is a second one in there at up to five per cent. I went looking for a minimum lethal dose for diethylene glycol in dogs or cats to set beside Merck’s figures, and the only readable copies were third-party reproductions of a veterinary toxicology textbook rather than a source I would cite. So I am not giving you a second number. What the sheet establishes on its own is enough: “ethylene glycol” is shorthand for the contents of the jug, not a description of them.
How these cases end I can count rather than assert. A retrospective series from a Canadian veterinary teaching hospital covering seventeen years reports twenty-one hospitalised cases — fifteen dogs and six cats — with survival to discharge of “33% of dogs (n = 5) and 0% of cats (n = 0)” (Sauder et al., *Canadian Veterinary Journal*, 2026). Twenty-one animals at one referral hospital is thin evidence and I will not pretend otherwise; a teaching hospital sees the cases that got bad enough to be referred, which biases the outcome. But it is a real count, and no cat in it went home.
The Sweet-Taste Story Does Not Survive the Cat
Here is where the received explanation falls over, and the correction changes what you actually do about a spill.
The sweet taste is real and it is documented in the right places. PubChem, the NIH compound database, files ethylene glycol under CID 174 with the experimental taste entries “Sweet taste” and “Bittersweet taste”; propylene glycol, CID 1030, is recorded as “Practically tasteless.” Merck names palatability as a contributing factor directly: “The widespread availability of antifreeze, its sweet taste, and small minimum lethal dose, along with improper storage, use, and disposal contribute to the frequency of ethylene glycol toxicosis.”
So far so good. The problem is which species the story is usually told about.
Cats cannot taste sweetness. Not “prefer it less” — cannot detect it. The mammalian sweet receptor is built from two proteins that have to pair up, and in cats one half of the pair has stopped being a working gene. Li and colleagues reported in *PLoS Genetics* that the cat *Tas1r2* gene carries “a 247-base pair microdeletion in exon 3” producing a frame shift and a premature stop codon early in exon 4, with further stop codons downstream, so “a functional sweet-taste receptor heteromer cannot form” and “the cat lacks the receptor likely necessary for detection of sweet stimuli” (Li et al., *PLoS Genetics* 1(1): e3, 2005). Seven other obligate carnivores have independently lost the same gene.
Line the two facts up. The species with the lowest published lethal dose, and the species that did not survive in that case series, is the species that provably cannot taste the property everyone gives as the reason it drinks the stuff. Whatever is happening, sweetness is not the mechanism in cats.
What does fit is duller and more useful. Merck’s list of exposure routes includes cutaneous absorption of products containing ethylene glycol, and notes that this is particularly documented in cats. A cat does not have to like a puddle to end up with a dose from it; it has to walk through it and then groom. The Canadian series is honest about the limits of what anyone knows here — exposure was “unwitnessed in a large proportion of the cases described and the sources of EG accessed by the animals could not be reported,” with only seven of twenty-one ingestions actually witnessed.
Which settles the practical question without needing the mechanism resolved. A coolant puddle on a slab is a hazard whether or not anything is tempted to drink it, and a shallow film spread over ten square feet by a rinse is more contact area than a deep puddle in one spot, not less. That is the whole argument for the order of operations, and it does not depend on a taste receptor at all.
The Case for the Propylene Glycol Bottle
Propylene glycol coolant is sold as the pet-safe option and the label is doing some work, so it is worth establishing what is actually true.
Merck’s companion entry on propylene glycol toxicosis is measured about it: “Although less toxic than ethylene glycol, propylene glycol, when ingested, may be associated with a syndrome similar to the acute phase of ethylene glycol toxicosis.” It gives an oral LD50 in dogs of about 9 mL/kg, and — the detail that tells you most about how the two compare in a clinic — says that for propylene glycol “the use of alcohol dehydrogenase inhibitors is not indicated,” because treatment is supportive rather than a race to block a metabolic pathway.
What I am not going to do is put 9 mL/kg next to 6.6 mL/kg and tell you propylene glycol is about a third safer. An LD50 and a minimum lethal dose are two different quantities measured two different ways — one is the dose that kills half a test group, the other is the lowest dose observed to kill at all. Dividing them produces a ratio of nothing. The direction is what the sources support, and the direction is genuinely lower risk.
The cat, predictably, has its own footnote, and it runs the other way. Merck records that in cats “ingestion of a diet containing 6%–12% propylene glycol can result in Heinz body formation and decreased RBC survival” — damage to red blood cells from repeated exposure rather than acute kidney failure from one drink. So the substitution does not scale the same risk down uniformly. It swaps a fast renal injury with a narrow treatment window for a slower haematological one, and in the species that already has the worst outcome.
Verdict, since the question deserves an answer rather than a hedge: yes, propylene glycol coolant is the genuinely lower-toxicity bottle, on the mechanism and on the treatment, and that is a good reason to buy it. It is not a reason to leave a puddle of it on the drive, and it changes nothing about the method below.
Where Removing Antifreeze Stains Sends the Glycol
This is the section the internet is missing, and the mechanism that makes it necessary is one line in a chemical database.
PubChem records ethylene glycol as “Miscible with water” and, in the same solubility entry, “practically insoluble in benzene, its homologs, chlorinated hydrocarbons, petroleum ether, oils.” Read that twice, because it means a coolant spill and an oil spill on the same slab are chemically opposite problems. Oil resists water, which is why an oil stain needs a degreaser to emulsify it and why pressure washing an untreated oil spill drives it deeper. Coolant dissolves in water without being asked, so water alone genuinely lifts it out of the pores.
All three pages I opened for this query have noticed the second half. The method they publish is to wet the area because coolant is water-soluble, add detergent, scrub and rinse — and it works, on the concrete. The split in what happens next is worth noticing. The one written for commercial operators tells you to block the drains and recover the rinse with a berm and a shop vacuum. The two written for homeowners do not mention where the water goes at all, and one of them carries a section on toxicity to people and animals four paragraphs earlier.
The manufacturer has a different method, and it names no water at all. The same Prestone sheet, section 6, opens its containment advice with “Covering of drains”, tells you to “Wipe up with absorbent material (e.g. cloth, fleece)” and to “Collect spillage: sawdust, kieselgur (diatomite), sand, universal binder”, and then, under environmental precautions: “Keep away from drains, surface and ground water. Retain contaminated washing water and dispose of it.” Section 13 adds “Do not empty into drains.” The company that makes the fluid prescribes an absorbent, a covered drain and retained wash water. The consumer SERP prescribes a hose.
One thing I am not going to inflate, because the same document declines to support it. Section 12 of that sheet states the product “Shall not be classified as hazardous to the aquatic environment.” So this is not an ecotoxicity argument and I am not going to dress it up as one. It is a mammalian exposure argument, and the exposure you actually control is the one on your own property — the twenty feet of driveway a dog crosses, the border a cat sits in, the strip of lawn a toddler kneels on.
So the opinion, and it is the strongest one on the page: do not bring the pressure washer to a fresh coolant spill, and the reason has nothing to do with the concrete. A jet does one useful thing to a stain, which is break the bond between the deposit and the surface — and there is no bond here, because the glycol is in solution rather than stuck to anything. All the machine contributes is water and velocity. Water is the thing you were trying not to add, and velocity turns one contained puddle into a fine film over a much larger area, plus aerosol. The trigger belongs at exactly one point in this job: the final rinse of a slab that has already had the fluid lifted off it.
When you do reach that rinse, the water leaving it is carrying dissolved glycol and dissolved dye, which puts this job in the second of the site’s three rinse-direction cases rather than the usual first one. On a sloped drive or an apron, work from the low end upward so the loaded water never crosses concrete you have not reached yet. That inversion is backwards from most of this site and the principle behind it is argued properly in removing graffiti — all this stain adds is that the thing dissolved in your rinse water has a published lethal dose.
Nothing in the Bottle Evaporates Except the Water
Here is why there is a mark at all, given that the fluid dissolves in water and mostly is water.
Ethylene glycol boils at 197.3 °C and has a vapour pressure of 0.06 mmHg at 68 °F, both from PubChem’s experimental properties. Water boils at 100 °C. At driveway temperatures the water fraction of a 50/50 puddle leaves and the glycol fraction does not, which means a coolant spill does not dry so much as thicken. It gets *more* concentrated while it sits there, and the ethylene glycol is hygroscopic besides, so what you find two days later is a syrup rather than a residue.
That is most of what a coolant “stain” is on a slab that has had a drip on it for a fortnight. Not pigment driven into the pores — a sticky, non-volatile film holding two weeks of grit, tyre dust and pollen against the surface. Which is genuinely encouraging, because a film holding dirt is the easiest thing on this site to remove. Detergent and warm water and a nylon brush, and it goes.
The dye is the part people expect to be difficult, and it is where this stain separates cleanly from its nearest neighbour. Removing transmission fluid stains owns the marker-dye argument in full — what the red is, why it was specified, and why the colour outlasts everything else. The reason that page reaches a pessimistic conclusion and this one does not comes down to one property: the dye in automatic transmission fluid is fat-soluble by design, sitting in an oil that water will not touch. A coolant colourant is dissolved in a mixture that is half water, so it is water-soluble by construction — it could not be in the bottle otherwise. Carrier and colourant are the same phase as your rinse water, and there is nothing for the colour to hide in. I could not find a coolant maker publishing which specific colourant is in a given product, and I am not guessing at one from a dye supplier’s catalogue. The solubility argument does not need the name.
(Prestone sells one called Antifreeze/Coolant with Leak Detect — a coolant carrying an additional tracer whose entire job is to be conspicuous. There is a product on the shelf designed to stain your driveway as a feature.)
That gives you a free diagnostic no other page on this site can offer. Wet the mark with plain water and watch the colour for thirty seconds. Coloured water lifting off the slab means coolant, and the job is easier than you feared. Colour that sits there unmoved while the concrete darkens around it means the mark is carried in something oily, and you are on the wrong page.
What the Coolant Picked Up Inside the Engine
The last question is what survives a detergent wash, and the answer is the part of the fluid that was never glycol.
Coolant spends years circulating past cast iron, aluminium, brazed copper, brass and solder, and it picks some of them up. EPA’s *Antifreeze Recycling* fact sheet, written for repair and fleet shops, puts it in the context of disposal: waste antifreeze “may contain heavy metals such as lead, cadmium, and chromium in high enough levels to make it a regulated hazardous waste,” and recycling residues “may be contaminated with metals such as lead, chromium, cadmium, copper, or zinc.”
Scale that honestly, because it is written about a drum of drained coolant at a garage rather than a pint under your wheel arch, and EPA says *may* twice for good reason — the levels depend entirely on the vehicle and vary enough that shops are told to test rather than assume. What it establishes is the direction, not a quantity: the fluid on your slab is not a clean laboratory glycol solution. It is glycol that has been inside an engine, plus an inhibitor package — that sheet names 2-ethylhexanoic acid, a carboxylate corrosion inhibitor, at up to five per cent — plus whatever metal came along.
The glycol, the water and the dye all leave with the rinse. Those other things are inorganic salts and metal compounds, and they do not. So the residual mark after a good wash, if there is one, may not belong to the same stain family as the spill that caused it. A faint rust-coloured shadow left where an old green puddle sat is an iron problem, and iron wants an acid or a chelator rather than the detergent that worked on everything else — the four families and which chemistry each one answers to are laid out in cleaning stains from concrete. Reaching for more detergent, or more pressure, on a metallic residue is the classic wasted afternoon.
Which brings the whole thing to one plain statement, since the alternative is the machine. On a coolant spill, an absorbent followed by detergent beats any amount of pressure, and it is not close. Pressure breaks bonds and carries away what chemistry has already loosened; there is no bond here and nothing loosened. The absorbent is the only step in the entire process that removes glycol from the site rather than moving it around the site, and everything after it is housekeeping. A bag of clay cat litter costs less than a nozzle and is, for this one job, the better tool.
And the honest limit, which is a short one. If the mark still will not go after the colour has lifted and a second detergent cycle has changed nothing, you are almost certainly not looking at coolant any more — you are looking at whatever else was under the car. A dark brown mark that does not release colour into water was oil before it was anything else.
FAQ
Does antifreeze stain concrete?
Less than almost anything else that leaks out of a car. Ethylene glycol is miscible with water, so it does not resist a rinse the way oil does, and the dye in it is water-soluble too. What marks the slab is usually the residue rather than the fluid: glycol has a vapour pressure of 0.06 mmHg at room temperature, so the water evaporates and the glycol does not, leaving a sticky non-volatile film that collects grit for as long as you leave it there.
How do you get antifreeze off a concrete driveway?
Absorbent first, always. Clay cat litter, oil-dry, sawdust or sand, trodden into the texture, left for an hour or more, swept up and bagged — that is the step the fluid’s own manufacturer prescribes, and it is the only step that takes the glycol off the property. Then warm water, a detergent, a nylon brush and a rinse. Judge it once the slab is dry, because wet concrete hides everything.
Is antifreeze still dangerous once it has dried?
It does not really dry. Ethylene glycol boils at 197 °C and is hygroscopic, so the water leaves and the glycol stays behind as a concentrated syrup that pulls moisture back out of the air. A spill that looks like it has gone off is often more concentrated than it was when it landed, not less. Treat a tacky coloured film on a slab as live, not historical.
Will a pressure washer remove antifreeze from concrete?
It will, and that is the problem. A jet’s useful job is breaking the bond between a deposit and a surface, and there is no bond here — the glycol is already in solution. So the machine contributes water and velocity, and between them they turn one contained puddle into a thin film over a much larger area plus aerosol. Use it for the final rinse after the fluid has been absorbed and washed, not before.
Does bleach remove antifreeze stains?
No, and it is a waste of a dwell. Bleach is an oxidiser aimed at organic colour — algae, mould, tannin — and a glycol film is neither an organic pigment nor something that needs breaking down chemically. Warm water and a detergent do the job on their own. Never put bleach on a slab in the same session as an acid, whatever you are cleaning.
Is pink or orange coolant safer than green?
Colour tells you the additive technology, not the toxicity. Both conventional and long-life coolants are overwhelmingly ethylene glycol-based, and the dye is there so that a leak can be identified and so that the wrong two products do not get mixed. The distinction that matters for safety is ethylene glycol versus propylene glycol, and that is printed in the ingredients rather than visible in the jug.
How long should I keep pets off the area?
Until the absorbent has been lifted, the slab has been washed with detergent, rinsed and allowed to dry completely — and then check it by touch with a gloved hand, because a glycol film stays tacky when a water film would have gone. There is no published clearance time for a driveway that I could find, so the honest answer is a state rather than a clock: dry, non-tacky and colourless.
The July Puddle Is the One That Matters
Ten of the twenty-one animals in that Canadian series were presented in June, July or August. Two came in across the whole of winter. It is one hospital over seventeen years and I would not build a rule on it, but it fits the mechanism better than the story everyone tells, because a cooling system fails under load — a hose lets go on a hot afternoon, a radiator seam weeps in traffic, and the puddle appears on a warm slab with the engine still ticking. December has the reputation. July has the leak.
So when coloured water turns up under the car in the middle of summer, the first thing to reach for is the bag of cat litter and not the trigger. The mark it leaves behind is the most forgiving stain on this whole site — water-soluble carrier, water-soluble dye, no film, no bond, no poultice. It will wait for you until Sunday without getting any worse, which is more than can be said for anything else on the slab.
