In January 2009, Concrete Construction magazine embarked on a study to examine curling in warehouse floors. The plan was to learn more about how concrete mix ingredients impacted the amount of curling. With the help of Greg Scurto, president of Scurto Cement, Elgin, Ill., who wanted to learn more about how to produce flat floors while contributing to the body of knowledge for the industry, CC assembled a committee of floor construction experts to design a study examining nine concrete mixes placed in a 60,000-square-foot floor. The developer of the warehouse promised access to the floor for a two-year period, in order to profile floor surfaces and monitor moisture movement in the floor for each different mix. The study aimed to observe curling in relation to:

  • Image

    Credit: Joe Nasvik

• The inclusion of 1 1/2-inch aggregate in different amounts

• The use of calcium chloride as an accelerating admixture versus nonchloride accelerators

• The influence of polycarboxylate superplasticizers

• Mixes with moderate and heavy dosages of macro fibers

Now two years later, the final profile work is complete, resulting in interesting data and conclusions, as well as more questions regarding floor construction.

Project history

From the beginning, the intent of the study was to examine real concrete applications and not lab-created samples—"real-crete" versus "lab-crete." Curling can be observed better in a real setting where differential drying occurs between the bottoms and tops of large floor panels. Linear shrinkage, on the other hand, can be observed better in the lab. Studying real-crete meant many variables wouldn’t be controlled, but that’s what contractors deal with all the time. The test was an opportunity to view what actually happens to floors in the real-world application.

The floor was placed under winter conditions so there was more than the usual variation in the water content between truck loads due to frozen aggregate piles. There were also significant temperature differences between concrete delivered at the start of the day versus at end of each day. On one occasion, 12 trucks waited in a line to be quality checked and emptied, but there were also times when there were no trucks onsite. But all the concrete met the required compressive strength (see Figure 1), and finishing sequences progressed normally. All placements were hard-troweled with riding finishing machines using pan floats, steel trowel blades, and finally plastic trowel blades. All floor surfaces exhibited a high degree of shine when they were completed.

Due to difficult economic times, the warehouse sat empty during its first year. In the winter, warehouse temperatures averaged about 50° F with ambient relative humidity (RH) readings approximately 40%. During the summer months, room temperatures registered about 70° F with RH readings as high as 100%.

Ten RH probes were installed in January 2010, covering every mix design, and the initial readings were all above 95% (see Figure 2). At the same time, 24 cores were taken in order to begin permeability studies of the surface finishes. Shortly afterward, half of the warehouse was rented to two different businesses, greatly restricting access.

  • Image

    Credit: HR Green

    January 2011. The 3D Scan shows the floor to be essentially flat. The indication of minimal curl at the upper right is probably due to the low angle of the scan at the greatest distance.
The floor stays flat

Two types of instruments were used to profile the floor surface: a 3D laser scanner and a D-Meter. The laser scanner provided a snapshot of the entire floor with an accuracy of approximately 1/8 inch, making it easy to spot trends. The D-Meter, on the other hand, measured a representative sample of panels, providing elevation differences to within 1/1000 of an inch. For the purpose of this study, D-Meter profiles were only conducted on floor panels bounded on all sides by sawed control cuts. Measurements were taken from corner to corner on the diagonal lines—unlike measurements taken for calculating FF.

By the end of the first year, it was thought that significant profile differences would develop between mix designs but that didn’t happen. Instead a regular pattern of low profile curling and then curl relaxation was observed. At present, the floor is without any significant amount of curling anywhere.

High slab relative humidity

Scurto installed a high-quality vapor retarder on top of the finished subgrade, just before concrete placement, which prevented moisture from the earth moving through the concrete and also any water movement from the concrete into the subgrade. So the water-of-convenience could only leave through the top of the slab, except that it was blocked by the dense finish.

RH measurements consistently were high for all nine installed locations throughout the entire two years of the study. Unfortunately, renters covered most of the original probe locations with pallets and equipment, so only four locations could be accessed for the final reading. Those RH readings were 99%, 88%, 96%, and 93%.

What was learned Some shrinkage occurs whether curling does or not. Almost all of the sawed control joints in the floor are activated now. The original width of the joints was approximately 1/8 inch but today a few joints in the floor are as wide as 1/4 inch now, while construction joints (formed joints surrounding each days’ placement) are 3/8 inch wide at all locations. Panel widths for all but one mix were 12 feet 8 inches by 14 feet 2 inches. Panel sizes for the mix with 7 1/2 pounds of fiber/cubic yard of concrete (Mix 6B) were 42 feet 5 inches by 37 feet 11 inches—the only sawed control joints being along column lines. Currently there isn’t a single crack anywhere in the floor.

  • Image

    Credit: Joe Nasvik

RH measurements confirm that water-of-convenience remains trapped in the floor due to the dense troweled surface finish and the vapor retarder underneath. RH readings are considered high at the end of two years. In “Should You Cure Floors?” (see Concrete Construction, August 2010), the statement was made that curing procedures for this floor weren’t necessary. This doesn’t suggest the concrete didn’t require curing, but rather the moisture retained by the dense troweled finish allowed the concrete to cure under ideal conditions for an extended period of time. Given the high RH of the concrete, the floor has remained essentially flat with little to no curling for two years and may remain that way for the foreseeable future.

Concrete is a forgiving material. In spite of the variations in mixes due to weather conditions, batch variability, time delays, and the decisions of placing crews, the entire floor performance to date is positive and uniform.

All the nine mixes placed included either micro- or macrofibers. Microfiber additions were 1 1/2 pounds/cubic yard of concrete. The dosage rates of macrofibers were 3 pounds and 7 1/2 pounds/cubic yard. The ultimate tensile and flexural strengths of concrete increase when macro polymeric fibers are added in the range of 0.5% (7 1/2 pounds/cubic yard) compared to the volume of concrete. At these addition rates, it’s possible to greatly increase the distance between control joints, thereby reducing eventual maintenance costs. Less is known about 3-pound addition rates.

Next steps

Howard Kanare, senior principal scientist at CTLGroup, Skokie, Ill., plans to conduct permeability studies on cores contributed by Sundt Construction, Phoenix, which were taken from a different floor slab placed outside the Phoenix area. When the study is complete, the results should help resolve questions about the effect ambient conditions have on the impermeability of densely troweled finishes.

You can purchase finishing machines as small as a walk-behind trowel weighing about a 100 pounds or as large as a 3000-pound “rider” capable of exerting very large forces per square inch on fresh concrete. Does the weight of the machine relate to the density of the finish layer or are other factors at play?

Retaining water in densely troweled warehouse floors has advantages. Concrete develops better strength over time, and the dense finish is more abrasive resistant. However, floors specified to receive other moisture-sensitive surface finish products must be able to dry out to acceptable levels within reasonable time periods. The thickness of the dense finish layer may relate to how well moisture can move through it.

Kanare completed water vapor transmission (WVT) tests on cores taken from the floor according to ASTM E96-05 “Standard Test Methods for Water Vapor Transmission of Materials.” The resulting WVT rates for the 1/8-inch-thick dense finishes were significantly less than required for curing compounds as defined by ASTM C309 or C1315.

However, the microfinished floor in the sprinkler room of the warehouse was much more permeable and didn’t meet the requirements for a curing agent. Would it be possible to specify floor finishes by the thickness of their dense layer? The committee thinks so and there are plans to begin an effort to define the characteristics of different finish thicknesses.

Learn more about CC's Field Test Study