Interior Cement Floors Don't Curl: But Concrete Floors Do Warp, and Joints Suffer!
By Scott M. Tarr, P.E.

When recently confronted with problems on a "cement" floor I thought, "If you actually constructed the floor with cement and not concrete, I would expect problems."

Just because "cement" is often used in lieu of "concrete" as we speak doesn't mean it's the correct material for floors. Yet we seem to get by with the terminology. In much the same way, while used interchangeably for the past century, there is a difference between curling and warping. Although even the American Concrete Institute(1) currently defines the terms similarly, when the two terms were originally described by pavement engineers in the early 1900s there was an important difference in the cause of the distortion(2,3,4).

"Curling" is the deformation of the slab surface due to a difference in temperature between the slab surface and bottom. Like most materials, concrete expands and contracts with change in temperature. If the slab surface is cooler than the slab bottom, the surface will contract, causing the slab edges to curl upward.

In contrast, slab "warping" is the deformation of the slab surface profile due to a difference in moisture between the slab surface and bottom. As with a sponge, if the slab surface is allowed to dry and the bottom is kept saturated, the edges will tend to warp upward. In general, the edges of interior concrete floor slab panels warp upward due to a moisture difference between the slab top and bottom, as shown in Figure 1. In most cases, curling on interior concrete floor slabs is just the sport shown in Figure 2.

While all interior plain concrete slabs can be expected to warp somewhat, there is no standard that specifies an allowable magnitude to this distortion. The acceptable amount of warping is dependent on several factors, including how the floor is used and the stresses resulting from applied loads. Warping causes edges of slab panels to lift off the base. Under heavy repetitive loading, excessive warping leads to fatigue cracking as unsupported edges deflect in the same manner as a diving board, eventually overstressing the slab.

Warping also impacts joint performance by decreasing joint stability. Joint stability is the differential deflection between adjacent slab panel edges(5). Semi-rigid joint fillers should be selected based on the level of joint stability achieved. Important filler properties include hardness and elongation.

If excessive warping is anticipated, mechanical load transfer devices in joints may increase joint stability. However, steel dowels (round, square, plate, or diamond) provide little resistance to warping. The long-term performance of warped joints may be compromised under repetitive traffic if attention is not provided to consolidate the concrete beneath the dowels (especially those with square or rectangular cross section). As a tight concrete/steel interface is required for complete load transfer, with poor consolidation the load transfer at doweled joints can be similar to that at undoweled joints where aggregate interlock provides load transfer. Many believe that aggregate interlock only transfers load when joint width is limited to 0.035 in. However, the research this theory is based on(6) found load transfers of 52 percent at a joint opening of 0.065 in. and 32 percent at an opening of 0.1 in. This may help explain why millions of square feet of non-reinforced concrete slabs on ground have provided many years of serviceability without incorporating the cost and risk of dowels in sawcut joints. Complete load transfer efficiency is not necessary for all wheel loads on all slabs.

Designers should not assume the inclusion of dowels in all joints automatically assures a good slab. While proper dowel alignment is regularly achieved in construction joints where forms help maintain position and allow verification and correction, dowels at control joints are often misaligned. This is routinely confirmed using non-destructive techniques. Dowel baskets at control joint locations can easily be jostled during construction, thus compromising dowel alignment and joint function, which can lead to excessive random cracking. In addition, if slab thickness is reduced by the inclusion of dowels at all joints and cracks occur where dowels are not located, the service life of the slab is substantially reduced. This is a costly risk that must be considered and minimized. The use of mechanical load transfer devices is advisable in some cases. But it's not necessary in all cases, and shouldn't be the only factor considered in slab design and construction.

The key to long-term serviceability of concrete slabs on ground is not the inclusion of steel, which is becoming more costly due to recent shortages, but a focus on reducing the shrinkage potential of the concrete mix. Dowels alone do not assure serviceability and require additional effort to maintain alignment and achieve adequate concrete consolidation. A recent survey of warped doweled joints found 50 percent to be unstable and in need of subsealing.

Whether or not dowels are used, reducing concrete shrinkage potential increases joint stability by optimizing aggregate interlock and minimizing warping. And by doing so, slab serviceability is increased at a reduced construction cost. Along with eliminating expensive steel, the cost is decreased by maximizing the aggregate and minimizing the cement content in the mix(7). With aggregate, cement paste becomes concrete. If slabs on ground were constructed of cement, the high heat of hydration would likely create a variety of distress.

In fact, in retrospect, given the internal heat that would be generated and surface cooling, I suppose a cement floor might actually curl...


  1. American Concrete Institute Committee 116, "ACI 116R-00, Cement and Concrete Terminology," ACI Manual of Concrete Practice, 2004, Part 1.
  2. Tayabji, S.D., and Colley, B.E., "Analysis of Jointed Concrete Pavements," Federal Highway Administration, Contract DTFH61-80-C-00052, October 1981.
  3. Huang, Y.H., Pavement Analysis and Design, Prentice-Hall, Inc., Englewood Cliffs, NJ, 1993.
  4. Tarr, S.M., Okamoto, P.A., and Sheehan, M.J., "Concrete Pavement Warping, Curling, and Subbase Interaction," PCA R&D Serial No. 2210, Portland Cement Association, Skokie, IL, October 1998.
  5. Tarr, S.M., "Industrial Slab on Ground Joint Stability," Concrete Repair Bulletin, International Concrete Repair Institute, Des Plaines, IL, May/June 2004, pp. 2-5.
  6. Colley, B.E., and Humphrey, H.A., "Aggregate Interlock at Joints in Concrete Pavements," Highway Research Record Number 189, 1-18 (1967); PCA Development Department Bulletin D124.
  7. Neuber, J., "If I Can Do It, Anybody Can!," Concrete Construction, Hanley Wood, Addison, IL, March 2004, pp. 47-54.
Scott M. Tarr is a Senior Evaluation Engineer with Construction Technology Laboratories, Inc., Skokie, IL. As a licensed PE, his principal experience is in the area of problem-solving distressed concrete slabs on ground throughout the United States and internationally and he specializes in comparing the as-built to specified load carrying capability of industrial floors and developing repair specifications to restore the serviceability to that designed. An ICRI and ACI member, he serves on several committees including 302, Construction of Concrete Floors, and 360, Design of Slabs on Ground.

He can be reached at

Back to ConcreteNews

© 2004 L&M Construction Chemicals, Inc. | ConcreteNews Fall 2004.

Subscribe to ConcreteNews