Here is a pop quiz for you. There is a desert and a swamp. We pour a concrete slab-on-grade in both a desert and a swamp. Which place will have more slab moisture problems? Yes, you guessed it, the desert. Why? Because they place sand between the concrete slab and the moisture vapor barrier.

It is hard to understand the practice of placing a layer of sand over the top of a moisture vapor barrier under a concrete slab. That this is an often recommended practice by geotechnical engineers that results in sputtering frustration among those who know better but can do nothing about it.

Almost everywhere that slabs are placed on the ground, a moisture vapor barrier is located directly under the concrete slab and on top of the ground (Photograph 1).

This plastic sheet is in direct contact with the concrete, as it should be. Note the continuous polyethylene sheet extending to the perimeter formwork completely isolating the concrete slab from ground contact.
But when we end up with sand or granular material between the plastic sheet and the concrete (Photograph 2), the sand layer becomes saturated with water from wet curing (water added to the top of the slab leaks down into the sand layer through service penetrations) or from irrigation (water added to the ground at the building perimeter enters the sand layer from the side or from underneath) or from rain (see wet curing and irrigation).

The sand layer cannot dry downwards by vapor diffusion because of the plastic sheet—it is a vapor barrier after all. In fact that’s why we have it under the slab. We want a vapor barrier under the slab. We want to prevent water vapor from coming up from the ground into the concrete slab. But it works both ways–it prevents water vapor from going down as well. It gets worse. The water also can’t drain out of the sand because it is held there by capillary forces. So even if we perforate the plastic sheet, drainage won’t happen. You can’t drain water out of a sponge. What if we used “pea gravel”? No, large quantities of water are still held in the pea gravel. Think surface area of the gravel and the fines in the pea gravel. Pea gravel does drain, but it retains huge amounts of water even though it drains.
Now for the ugly part. The sand layer is wetted by “liquid phase wetting” in a time frame measured in minutes. The sand layer can only dry upwards by “vapor phase drying” in a time frame measured in years. You can do a simple experiment at home boys and girls. Take a paper napkin and wet it by putting it in a cup of water. Now pull it out. It got wet in seconds didn’t it? Now hold it horizontally in the air and wait for it to dry. Takes a long time doesn’t it? Now do it again by placing the wetted paper napkin over a plastic sheet and covering the top of it with a layer of concrete. In the real world a 15-minute liquid phase wetting event can take years to dry.
Guess what? No sand layer, no reservoir to get wet and hold and store water to cause grief. No problem. So where did the idea of putting sand between plastic and concrete come from? Like most bad ideas, it arose from good intentions. It was an attempt to reduce the curl of concrete slabs due to uneven drying. If a slab dries only from the top, the top becomes drier than the bottom. The top shrinks relative to the bottom and the slab edges curl upwards. This is not a big movement and is not a big deal except where you need flat floor systems. Slab curl is very easy to control, use a low water-to-cement ratio (less than 0.5) and wet cure the top (wetted burlap works—Photograph 3). Or just use a low water-to-cement ratio and forget about doing anything else.

Photograph 3 Burlap Cure

A fluted steel deck (Photograph 4) that is going to get a concrete slab placed directly on it. Steel decking is a pretty good vapor barrier. Anybody every see a sand layer between the steel decking and the concrete slabs that are cast over them? So why do we need a sand layer over the plastic for the slab on the ground? Apparently the laws of physics change between the first floor and the second floor…

Photograph 4 Steel Decking

Okay, so we don’t need the sand layer to handle the “curl” thing. But what about using the sand layer to protect the moisture vapor barrier? The polyethylene does need some protecting. You can have holes in it, and pretty much have your way with it as long as it is in direct contact with the concrete. But, vapor barriers have to be continuous and free from any holes. Actually, no. Air barriers need to be continuous and free from holes, but moisture vapor barriers do not need. Most of the vapor moves by air movement, not much of the vapor moves by vapor diffusion. The concrete slab is the air barrier, and the punctured moisture vapor barrier is the vapor barrier. Diffusion is a direct function of surface area. If I get 95 percent of the surface covered I am pretty much 95 percent effective—and the parts that are left I have filled with concrete which is also pretty good as a vapor barrier. But put that sand layer in there and you are doomed.
If I don’t put the sand layer in there it will take too long to finish the floor? Yes, that’s true if you use concrete with too much water in it. The easy answer is don’t concrete with too much water in it. See water-to-cement ratio less than 0.5 above.

Thanks again to Tim Mcadoo