Sand Under Slab-On-Grade

[CanuckPE]

I've specified a thin layer (2") of sand directly under concrete for several slab-on-grade projects per a professional development course I participated in a few years ago. As I understand it, wetting the sand prior to concrete placement helps mitigate warping of thinner slabs during curing by equalizing moisture migration from the top and bottom of the slab. I've done about five projects with the sand layer, and the field reviews went reasonably well without any negative feedback from the client or contractors. For all five projects, the slabs were relatively thin (as thick as 8" with a single layer of reinforcing steel).

I just field reviewed a mat foundation with a relatively heavy reinforcing steel cage on it, and I utilized the thin layer of sand approach. There is a 20M bar @ 12" o/c e/w top and bottom. The mat thickness is between 22" and 12" depending on the location. The bar tying folks built the cage in-situ and set the cage on brick chairs about every four feet on centre. I'm told the compaction tests were completed for several inches of structural fill below the sand.

Upon field review of the cage, I found the brick chairs were embedded into the sand and the bottom cover is now way under that specified (3"). I haven't done any checks on bearing pressure (cage weight over bearing area of the bricks), but I suspect the weight of the cage divided by the bearing area of the bricks results in a high pressure. The sand is not not very compact, as it's been disturbed by all of the construction activity, but it's under a double layer of reinforcing steel, so I think further mechanical compaction is out of the question unless the cages are removed. Perhaps we could hand-tool it. If I had been the contractor, I would have had the cages preassembled, chairs prepped/measured then set the cages onto the chairs using a telehandler, but I did not specify this technique.

A couple of questions:
1) What are your thoughts on the relatively loose layer of sand between better-compacted fill layer below and the reinforced concrete above? Would it contribute significantly to settlement/differential settlement if left loose as it is? Should we remove the layer of sand? Should we try to hand-tool for compaction?
2) Does anyone have any similar experiences with reinforcing steel resting on a sand layer for S-O-G? What are your experiences?

 

[NOLAscience]

I wouldn't worry about compacting the sand. It is probably still in decent shape to support the concrete to be poured over it. Maybe they could smooth out the high points and kick some sand into low spots with a trowel or something. Just do the best they can in an hour's work.

You could try to lift the cages (if they contractor has the means to do so) and slip some thin flat plate under each chair, smoothing the sand to near level before placing the plate. I'd say that it's imperative that you have 2" of cover over the steel rebar. That would be the minimum if you were pouring against a concrete slab or a form.

 

[CanuckPE]

Thanks for your reply NOLAscience!

The contractor has offered to lift and rechair the bars with a telehandler and another palette of bricks. I'm planning to review again shortly following their rework.

The mat is for rail traffic. Localized bearing pressure considering the thickness of the slab under wheel loads with a bearing stiffness of 600 lb/in3 is in the range of 1300 psf (which is relatively small), but the system will be sensitive to settlement and differential settlement.

 

[kingnero]

Do you pour the concrete directy on the sand? I've always seen a plastic barrier inbetween sand and concrete, to avoid loss of moisture at the bottom.
This is not a water/vapor barrier. Only purpose is to retard the drying out of the concrete at the bottom, which is what you are trying to do with wetting the sand before pouring, if I understand it correctly.

 

[XR250]

I believe the intent of the sand is to allow some loss of moisture on the bottom of the slab to minimize curling

 

[cvg]

the sand thickness will vary due to foot traffic during the steel placement and there is no way to screed it after steel is placed. this results in a slab of varying thickness which may not be desired. typically we use an aggregate base course material, compacted to 95% which is very stable. subgrade is always wet down prior to pouring concrete to slow down moisture loss into the subgrade.

 

[CanuckPE]

Thanks for your replies!

kingnero and XR250, the concrete goes directly on the sand immediately following a wetting of the sand. The goal for curing is equalizing moisture migration from the top and bottom of the slab. The wetted sand helps equalize the moisture loss from the top and bottom. When moisture migrates more heavily to the top rather than the bottom, the slab tends to warp by curling up on the sides and corners in particular. Thicker slabs tend to not warp as much. A vapour barrier sends moisture only up and tends to amplify warping.

cvg, I agree. This project is likely the last time I will place sand under reinforcing steel for a SOG. I suspect the results would be similar for warping by wetting clean granular fill.

 

[geotechguy1]

I've had several jobs where these sand layers specified by structural engineers under slab on grades got washed out somehow (flood or leaking waterline) and then the slab ended up moving significantly. Not the question you're asking but worth making sure there isn't a pathway anywhere for this to happen.

 

[Enable]

Specifying sand to mitigate curling is not the right approach (though least you didnt sandwich it between a retarder and the slab). See for example Concrete Floor Problems by J.L.

You want a low w/c combined with wet curing. Removing a vapor retarder from the assembly due to curling concerns is absolutely silly. You are going to cause quite the vapor drive leading to all kinds of issues with your floor finish if not properly taken into account post-slab (e.g. ditra or epoxy). You're literally adding dollars / ft2 to the floor assembly instead of laying down a plastic sheet.

 

[Lomarandil]

To be clear, I ask for sand above a vapor/moisture barrier.

----
just call me Lo.

 

[Enable]

Lo stop that. Stop that right now.

God help you if they actually pre-wet it. The sand is going to retain the moisture and create a massive vapor drive posing all sorts of unnecessary issues for your SOG floor finishes. See link I posted above.

 

[BigH]

Why not place a thin blinding layer of concrete instead?

 

[SJBombero]

Enable - I read the article and I am convinced. Any trick on how to convince the AHJ not to insist on the sand/gravel layer over the vapor barrier? They are pretty insistent here even though technically it is not a structural issue and has no place on our drawings.

 

[Rabbit12]

Enable is right on for floors with finishes. You don't want anything between your slab and vapor barrier. You can deal with curling by having a well graded coarse aggregate in your concrete mix. It's been several years, but I developed specifications for my former employer on the concrete mix design. I believe ACI has a guide for slabs on grade that gives guidance to the gradation.

If you have no floor finishes you have a lot more options for what your base looks like.

pvchabot, I disagree. Although the vapor barrier isn't structural it's installed by your concrete slab contractor. If it's not on the structural drawings you run the risk of it getting missed. As I noted above there are also issues with curling that need to be considered in the mix design so it affects structural in that way too.

 

[Enable]

Any trick on how to convince the AHJ not to insist on the sand/gravel layer over the vapor barrier?

Fortunately I have not had to deal with such things since in my Canadian jurisdiction it's mostly done the proper way. I have other battles up here (like always putting a vapor retarder inboard of cavities in residential basement walls...FFS). But I do have some ideas (not tested at all so YMMV):

A) Add a note on the drawing or a specific leader pointing to the SOG / base interface that says "see architectural plans for control layers". The control layers are supposed to be an architectural / building science thing anyways so they should be on those drawings. Make sure to talk to the architect about how to do it right / where it should be placed.

B) If you don't like your architect then try talking to your geotechnical engineer about recommending the correct ordering in their report. In my experience, AHJs will be much more nit picky over structural design elements than anything coming from a geotech. And do the same thing with the leader / note "subsurface control layers to be per geotechnical report" and include the geotech report or snippet where it says vapor retarder overtop the capillary break in with the project documents.

BTW if this is not typical for where you are do the contractor a favor and make sure you point it out to them. These things are very easy to miss if they've always done something one way and they'll probably assume you just made a mistake on the DWGs (I know because I am one...help a brother out!)

 

[CanuckPE]

Thank you all for your replies.

Thanks for the heads up geotechguy1. Considering the arrangement, I don't think piping will be an issue. The sand is deep and well confined.

Enable, we have a different climate and standard practices that we work with in Canada. There are no T.O. concrete finishes for this mat that were referred to in your linked article. It's an industrial floor (which I failed to mention). We're semi-arid, and I only specify a vapour barrier if we're seeking to retard contaminant migration (up or down). A vapour barrier is typically required for residential basement slabs with NORM such as Radon in the soil.

BigH, that's a good idea. Would a thin and wetted leancrete layer have the same, or similar, effect on curing as the wetted sand layer?

 

[Enable]

I live and work in Ontario. I suspect I am acutely aware of your local climatic conditions

Absolutely you don't need a vapor retarder for an industrial facility where floor finish is likely to be minimal over the service life and where the diffusion wouldn't pose a problem for indoor air quality or mold concerns (such as non-conditioned space). Incorrect that it is only for residential with radon concerns.

Sand is still not the way to go though, you'd want more granular material that acts as a better capillary break plus is easier for compaction / setting.

As to adding an additional layer beneath...for the love of god just get a C2 or C1 mix and wetcure it (without air if not required to be air entrained). It'll be cheaper and your curling will be minimum to nill. Though if it is an industrial facility the mix will be "value engineered away" by the contractor pretty quick so maybe doing something completely different even if more expensive is the way to achieve your result. Meh.

 

[CanuckPE]

Enable, I didn't realize you were in Canada, eh. The linked article referred to Florida and California. I just noticed the reference to the Canadian jurisdiction.

With the exception of floor finishes and keeping radon out of conditioned spaces (residences or otherwise), what benefits do vapour barriers under SOG's provide? And, for what applications do vapour barriers add value?

S2 is specified.

 

[Settingsun]

I've never come across a loose sand layer. Occasionlly used to fill in particularly rough granular sub-base to protect the plastic sheet from punctures, but not a layer per se.

I don't deal with slabs that have floor finishes so the plastic sheet just seems to be a matter of preference. For slabs with bottom reo, we (Australia) can use 10mm less cover if the plastic sheet is used, but it's an extra step so not a clear cost win to include it. I haven't seen a sand layer on the sheet before.

I specify watering down bare granular sub-base before the pour, mainly because because a very good engineer told me to years ago.

 

[Enable]

Hope you enjoyed your Canada Day!

If you neglect to place a vapor retarder under your on-grade elements but have control layers (vapor,air,thermal) on the roof / walls, you have created a very effective way to mine moisture. You have a constant source of moisture (grade) diffusing into the conditioned space (above grade), and have trapped it there because we've placed vapor retarders on every surface it could otherwise go through. Whether or not it becomes a problem will depend on the particulars. Such as, can the HVAC system keep up? What is the temperature across the wall assembly and where is the dew point? If you are mining moisture but have a dew point inboard of the exterior sheathing you are going to have problems (cant dry out due to retarder on other side and wont dry in because, well, that's the high concentration source we've mined).

If an on-grade element is below conditioned space, I cannot fathom why one would want to remove the retarder from the assembly. But this is for building science considerations, not structural. Sometimes finishers will complain about drying times but they will complain about everything, so I don't really care for it lol

BTW S2 will work gangbusters as the w/c is 0.45! You said it's an industrial facility so I imagine you are dealing with a FF 80-100 floor combined with an application of a densifier? Wet curing in that case may be an issue depending on schedule / how quick it is recommended to densify. But if possible wet cure with NCF blankets. Guarantee you'll have no curling issues at joints or sides again.

EDIT - And if you do use a densifier make sure you consult the manufacturer as removing the vapor retarder from the assembly may create problems. The densified surface will be far less permeable than the concrete below, and the drive may cause issues at the interface.