Concrete Control Joints & Cracks

The Behavior of Concrete Slabs With Control Joints and Cracks
by A. Kester

A lot of architects, contractors, and even engineers, do not have a good understanding on the behavior of concrete slabs-on-ground (SOG). I am just one structural engineer and these are my thoughts based on my research and experience. I have worked with concrete slab construction my entire career, and have done a lot of investigations of cracks and settlement damage to SOG, and have a pretty good general understanding of them. This is a review of the subject and I hope it helps people in the flooring industry in dealing with different types of slabs with control joints and cracks.


All concrete shrinks when it cures, due to loss of excess water leaving the concrete. However, only part (40-70 percent) of the shrinkage is recoverable with future wetting cycles, so that means a good percentage of the shrinkage is permanent no matter what happens to the slab in regards to moisture content in the future. Adding water at the chute during placement is a BIG no-no, even though it increases workability at the site, all that additional water must come back out of the concrete and this leads to more shrinkage. The next time your wife is baking a cake or something just before she is done throw an extra half cup of water in the bowl and see what happens….

ACI 302.1R-60, 9.10: Most shrinkage takes place within the first 4 years, with the most significant in the first 60-90 days. It is dependent on many factors like mix design, weather, curing, and the spacing and location of the joints.

Slabs may be restrained by walls/foundations at the perimeter edges and by the ground/vapor barrier along the bottom of the slab. These restraints from shrinkage movement cause tensile stresses in concrete. Concrete is great in compression, but weak in tension (about 8-10% of the compressive strength). So these tensile stresses cause cracks. Control joints purposefully weaken the slab at designed locations to CONTROL the cracking, but of course, not completely. In commercial construction, control joints are often formed by saw-cutting and this should take place within the first 4-6 hours after placement, or as soon as you can walk on the slab and saw-cut without dislodging any aggregate from the concrete.

If a slab cracks at a control joint, then it has done its job. There is no such thing as a crack-free slab, but there are slab systems that use rebar and post-tensioning that can greatly reduce the WIDTHS of the cracks. Welded wire reinforcing (looks like chicken wire) is OK in theory, if it is placed in sheets in the top 1/3 of the slab, which it never is, so in practice it is not of much value. If I specify it, I go heavy and space the bars at least 14 inches so a workman can step between the bars, and specify flat sheets. Polypropylene fibers (fiber mesh) and the like (seen a fuzzy slab?) are not proven to provide adequate resistance to shrinkage cracks, but can reduce plastic shrinkage cracks.

Back to control joints, or sometimes called contraction joints – not to be confused with expansion joints, which are placed in exterior walls in buildings, retaining walls, bridges, heavy roadways, etc., but not normally in slabs-on-grade because expansion is rarely an issue, especially in a climate controlled environment. Thermal and moisture changes in buildings rarely cause enough expansion in concrete slabs to cause problems, especially if there were control joints that cracked due to shrinkage, these may compensate for any expansion, though that is not what they are specifically designed for.

Control joints are rarely used in residential construction outside of garages and driveways, but they are the same as in any other construction. Often in driveways you see pressure treated wood used to separate sections of a slab, this is essentially a contraction joint that allows the individual slab sections to shrink. Also, you can form a construction joint this way, which is an intentional joint in a concrete slab that is used when concrete placement cannot be completed in one continuous operation (between trucks, days between placement, ran out of concrete, etc.).


So what do you fill Control Joints with? It really depends on the floor slab, the finish, and the intended use. Exposed slabs at home improvement stores that support forklift traffic need a semi-rigid material that both bonds and protects the edges of the cracks, can support compression (woman’s high heel probably the worst case), but still allow for some movements (elasticity). One of the links I posted below goes into a lot of detail on this subject and that type of application.

Since this is a flooring website, you are dealing mostly with enclosed, climate controlled spaces without forklift or similar traffic. So you should be concerned with the age of the slab (is it mostly done shrinking?), flooring type, and flooring material. The million dollar question is, will this joint move in the future? Well, nobody can say for sure, probably any joint will move slightly due to slight thermal and moisture expansion and contraction cycles as discussed above. But if the majority of the concrete shrinkage is done (60-90 days on average), and the building is being climate controlled, then we can consider the joint stable and then choose the proper sealant.

I think the worst solution is a cementious-based patching compound that you spread on the surface. This has a high chance of failure because, as was pointed out above, concrete and cement-based products are bad in tension. If that crack widens even slightly, it will likely crack that type of compound. It could also compress and force the rigid and brittle material up and cause a failure of the flooring. ACI and most product manufacturers recommend elastomeric sealants, which are flexible and allow for some movement of the joint; or semi-rigid epoxy or polyurea fillers, which are less flexible but more rigid and better in compression (good for exposed slabs subject to forklift and pallet jack wheels). Vacuum the joints rather than use compressed air per ACI, though not sure why other than maybe blowing compressed air just forces some of the debris deeper into the crack.

If you are dealing with a joint or crack that you believe is unstable either vertically or horizontally, and this is incompatible with your flooring (such as tile), then an epoxy product that “welds” the two sides of the slab together may be the best solution. If done correctly, this will prevent movement along the crack/joint, and provide a stable surface to place flooring across. An extra step of using a fiberglass or similar mesh in epoxy over the crack/joint may be necessary; but on an older slab that appears otherwise stable, probably unnecessary. If possible, placing expansion/contraction joints in the tile that mirror the control joints would be ideal, which the Tile Council of America recommends for ceramic tile expansion (though I don’t really agree that they help, but that is a different subject). But I can see how this could be a nightmare for your tile layout and in reality probably isn’t that feasible.


If you have a slab that is at least 60-90 days older in a controlled environment, and you have no reason to think the joints/cracks are unstable, use an elastomeric sealant, or semi rigid epoxy or similar. If any movement across the crack or joint is undesirable or may cause damage to the flooring, then a full depth epoxy sealant may be required.

Additional Reading

Contraction/Control Joints | Portland Cement Association (PCA)

American Concrete Institute (ACI), the national technical authority on most things concrete.

I found this article that does a pretty good overall job of explaining control joints, possible movements, and different types of sealants, though it appears to be more aimed at exposed slabs such as those in warehouses, home improvement stores, etc.: Rethinking Control Joints – Concrete Construction

This is from a company, and is a longer read, but has a lot of good information and good illustrations: Crack and Joint Repairs on Industrial Concrete Floors and Other Areas PDF

Slab curling is caused by drying and shrinkage, which is a separate issue, and this article explains it well: American Concrete Institute | Curling in Concrete

About the author:

A. Kester basks in the Florida sun. He currently works in the forensic engineering field, performing structural investigations, mostly of buildings and houses with foundation problems, including flooring issues. He is also an avid DIYer that has done a fair amount of flooring, including laminate and quite a bit of tile. His passions are surfing, the outdoors in general, travel, Florida State and the Green Bay Packers.

If you would like to discuss this article, or any of the other articles you have found here at, please join us in the Articles Discussion Forum. If you are not yet a member of The Floor Pro Community, we invite you to register now, it’s free and gives you the opportunity to enjoy all the features of the site.

Leave a Reply

Your email address will not be published. Required fields are marked *