Search for garage floor coatings and you will find a hundred pages explaining that epoxy is the classic choice, polyaspartic is the premium upgrade, and polyaspartic is the one that works in the cold.
The first two are fair. The third is the one we want to correct — partly because it decides real money in this region, and partly because we used to repeat it on this very page.
A correction to an earlier version of this guide
Until August 2026 this page said standard epoxy needs a slab at roughly 10°C, and that polyaspartic cures far below freezing. Then we went and read the manufacturer data sheets, which is a slightly humbling exercise. Neither statement holds as a general rule, and the second one is not true of most polyaspartics sold for garage floors.
It is left here rather than quietly deleted, because if you were told the old version by somebody else you deserve to know where it came from.
First, the question this raises: will the finished floor survive −20°C?
Yes. Get this distinction clear before the numbers below, because they are easy to misread.
Application temperature is a process window. It governs whether the chemical reaction completes properly while the floor is being laid. Service temperature is what the cured film tolerates for the rest of its life. They are unrelated specifications.
A floor coated in July at 25°C is fully cured by August, and a cured resin floor handles a Georgian Bay winter without difficulty. Most data sheets do not even quote a lower service limit, because cold is not a failure mode for a cured floor. Where one is published — SaniCrete’s STX system — it is −40°C continuous.
So if the scheduling suits you, booking the work for summer sidesteps the entire temperature question. Every product on the table below is in range in July.
What cold does do to a finished floor
Not cracking from cold. Something slower. Concrete expands and contracts at roughly 10 × 10⁻⁶ per °C. Cured epoxy film runs about 55–60 × 10⁻⁶ per °C — a mismatch of around five to one. Every time the slab passes through zero, the two materials move by different amounts and the bond line takes the difference.
That is the real argument for a more flexible coating in the snowbelt, and it is an argument about the years after the install rather than the day of it. It shows up at control joints and at the door line, which is where you should look on any floor here that is more than five years old.
The number that matters is on the pail, not in the chemistry
Here is what the published data sheets say. Every figure below is a minimum application temperature taken from the manufacturer’s own document.
| Product | Chemistry | Minimum to install |
|---|---|---|
| Rust-Oleum 9100 High Performance | Epoxy | 0°C air, substrate and product |
| Rust-Oleum 9100 WG Winter Grade | Epoxy | 0–15°C |
| Sikafloor-264 | Epoxy | 3°C substrate, 5°C ambient |
| Sikafloor-315 | Epoxy | 10°C substrate |
| ExteriorCoatings Aspartic 85 / 100 low-temp | Polyaspartic | −1°C |
| Deco-Crete Polyaspartic 90 Standard | Polyaspartic | 4.4°C surface |
| Onyx Polyaspartic 87 | Polyaspartic | 10°C |
| Citadel POLY100-SC | Polyaspartic | 13°C optimal |
| Leggari polyaspartic | Polyaspartic | 16°C |
| Elite Crete SPARTIC-ALL MXP | Polyaspartic | 21°C ideal |
Read the two columns against each other and the story falls apart. The coldest-rated product on that list is an epoxy. The warmest is a polyaspartic. Three of the five polyaspartics want a floor warmer than Sikafloor-264 does.
So “polyaspartic works in the cold” is a claim about a category, and a category does not have a working temperature. Only a labelled pail does. When a company tells you they use polyaspartic so they can work through the winter, they have told you something unverifiable — and if the pail in their van is Elite Crete MXP, something untrue.
What the slab lag does on top of that
Whatever the number on the sheet, it applies to the concrete and not to the air. A slab is a large thermal mass sitting on ground that is frozen for months. It lags air temperature by days, and it lags in the direction that hurts: a warm afternoon in April does almost nothing to a slab that has been cold since November.
In an unheated garage between Collingwood and Wasaga Beach the slab sits under 10°C from roughly mid-October to early May. That does not close the season. It closes the part of the shelf that asks for 10°C or more — which, on the table above, is most of the polyaspartics and one of the epoxies.
And the dew point closes days the thermometer would allow
The same sheets carry a second rule that gets ignored more often than the first: the substrate must sit at least 3°C above the dew point, and relative humidity must stay under about 85%. Beside Georgian Bay a mild damp morning off the water can fail that test while a colder dry one passes it. A crew that checks only the temperature is checking half the specification.
What happens if a coating goes down too cold
It is worth being specific, because “it needs to be warm” sounds like a preference rather than a failure mode. This section is about epoxy specifically, though the first and third points apply to any resin below its rated minimum.
Epoxy cure is a chemical reaction between resin and hardener, and reaction rate falls with temperature. Below the product’s working range, three things happen:
- The cure stalls. What should be hard in 24 hours is still soft after three days, and it may never reach full hardness. A soft floor marks under a tire and gouges under a jack stand.
- Amine blush. Cold, damp conditions bring an oily amine residue to the surface as it cures. Anything coated on top of that — the clear top coat, typically — bonds to the blush rather than to the base, and delaminates later.
- Viscosity climbs. Cold resin is thick. It does not wet into the profile of the ground concrete properly, so the mechanical key you paid for by grinding does not get filled.
None of these are visible on handover. All of them turn up in year one or two.
Side by side
| Polyaspartic | Standard epoxy | |
|---|---|---|
| Slab temperature to install | Product-specific: −1°C to 21°C | Product-specific: 0°C to 10°C |
| Realistic season here, unheated bay | Depends on the product, not the chemistry | Depends on the product, not the chemistry |
| Foot traffic | ~24 hours | ~24 hours |
| Vehicles | 3–7 days on the sheets, not the 24–48 hours advertised | 5–7 days |
| UV stability | Stable | Aromatic grades amber, visibly, at the door |
| Freeze-thaw movement | Flexible, tolerates it | Harder, more brittle |
| Film build per coat | Thinner | Thicker; self-levelling grades much thicker |
| Pot life once mixed | Minutes | Hours |
| Cost | 20–35% more | Cheaper |
The reasons this matters more here than elsewhere
Road salt has a season and so does the fix. The damage happens November to March, and what it costs to fix afterwards rises every year you leave it. If the product your installer buys asks for 16°C, your only installation window is May to September — and if the summer books out, a floor that needed protecting last winter goes unprotected for another one. That is an argument for asking about the pail early, not for assuming a chemistry solves it.
Freeze-thaw is not a footnote in the snowbelt. A slab in this region crosses 0°C repeatedly through the shoulder seasons, and it moves as it does. Cured epoxy is harder and more brittle; polyaspartic stays more flexible. Over enough cycles that difference shows up at the control joints and at the door.
Sunlight reaches most garage floors here. A bay that spends summer afternoons with the door open gets direct UV on the concrete inside the opening. Epoxy ambers there and nowhere else, so you end up with a stripe.
When we will tell you to use epoxy
We install both, and there is a real case for epoxy that we would rather state than bury:
- A heated shop. The workshops around Thornbury and Clarksburg, heated through the winter with a stable slab temperature and no direct sun on the floor, remove every reason we normally steer away from epoxy. It costs less and it performs well.
- A genuinely rough floor. Self-levelling epoxy builds far more film in a single pass. If the goal is to flatten a bad surface rather than follow it, that is the product that does it.
- A tight budget on a sound slab. If the slab is in good condition, the bay is heated or the work can wait for July, and the money matters, epoxy is a reasonable, honest choice. We would rather install a good epoxy floor than lose you to a cheap polyaspartic one installed badly.
- A cold slab in late autumn. This is the one that surprises people, and it follows straight from the table above. On a 2°C slab in November, a winter-grade epoxy rated to 0°C is a working product and a polyaspartic rated to 13°C is not. The chemistry that gets sold as the cold-weather answer is sometimes the one that has to wait.
What to ask any installer, including us
Three questions, and they are the whole of it:
- Which product, by name? Not “polyaspartic”. A brand and a product code.
- What minimum substrate temperature does its data sheet give, and what will my slab read on the day? Two numbers. If they cannot produce the first, they have not read the sheet.
- What does the sheet say about vehicle traffic? Compare that to what the salesperson promised you. On the sheets we have read, common polyaspartics give 3 to 7 days. The one-day claim is about the install, not about parking on it.
We put the product name and its minimum on our quotes for exactly this reason. It is a low bar. It is also, oddly, not standard practice in this trade — which is most of why this page exists.
The catch on polyaspartic, stated plainly
Short pot life. Once it is mixed the clock is running hard, and in warm weather an inexperienced crew can lose a batch by being slow with it. There is no opportunity to go back and fix a section the way there is with epoxy.
That makes installer competence matter more than product choice. A well-installed epoxy floor beats a badly installed polyaspartic one every time, and it is not close.
It also will not rescue a bad slab. If moisture is coming up through the concrete, polyaspartic delaminates exactly like epoxy does. Neither product is a substitute for the slab moisture test, and any company selling you polyaspartic as the answer to a wet slab is selling you the wrong thing twice.