Scanning electron microscope image of hard-troweled concrete cross section from upper surface to approximately a ¼-in. depth.

Credit: CTLGroup

In February of 2009, CONCRETE CONSTRUCTION and Scurto Cement, Gilberts, Ill., decided to do a two-year warehouse floor study to monitor curling. Scurto placed and finished the concrete, while CC periodically gathered profile data on the 60,000-square-foot floor surface to record any movement in floor panels. Scurto hard-troweled and burnish finished the floor in a manner consistent with what most big box owners currently require. Moisture in the top surface region of the slab was pressed downward during the finishing process, resulting in compacted densified concrete. Given the very dense finish, we began to wonder how possible it was for moisture or water vapor to pass through the surface layer. Howard Kanare, senior principal scientist at CTLGroup, Skokie, Ill., and an advisory committee member for this project, removed core samples of each mix design in the warehouse for petrographic study and for water permeability testing. Among other things, we wanted to learn whether this dense compacted surface also qualified as an effective curing membrane.

Mixes and ambient conditions

The primary purpose of the study was to profile floor panels to follow curling and moisture migration over a two-year period, noting any differences between the 10 concrete mixes placed. Concrete was placed in early February 2009 inside a well insulated building, which remained empty during the year that followed. Ambient temperatures were approximately 50° F during the winter months and 70° F in the summer with relative humidity's (RH) between 70% and 85%.


The size of the machine matters. This ride-on trowel produced a 1/8-in.-thick densified layer.

Credit: Joe Nasvik

RH measurements in the concrete for each of the mixes remained high throughout the first year and a half of the study—never less than 94%. At first we thought this was because the building wasn't in use, with ambient conditions for concrete being nearly ideal. But half the building was rented at the one-year anniversary and for the six months that followed the rented portion of the building was heated. But six months later the floor RH were the same for all areas while ambient RH levels remained lower than the slab's RH.

The concrete mixes were designed to evaluate the amount of curling related to ingredients such as calcium chloride, different sizes of large aggregate, the inclusion of standard dosages of microfibers versus 3- and 7½-pound dosages of macrofibers per cubic yard of concrete, and the inclusion of polycarboxylate superplasticizers for reducing the w/c.


This small ride-on trowel produced a micro-thin densified layer.

Credit: Joe Nasvik

Hard troweling

Scurto placed 20,000 square feet of floor on each of the three days of construction. The placing and finishing process was completed in a manner consistent with current big box floor construction: a 3-D laser screed struck off the concrete, workers used pan floats mounted on riding trowel machines weighing between 1500 to 3000 pounds for the first couple passes, followed by several passes with steel blades (increasing the blade angle with each pass), and finally using plastic trowel blades to provide a burnish finished. Jay Allen, president of Allen Engineering, Paragould, Ark., says the static weight of their 1660-pound riding trowel (add 200 pounds for the operator) is 323 pounds per square inch with steel blades pitched to their sharpest angle. But the pressure applied to the concrete surface is probably higher when the true contact pitches and forces of the blades agitating the concrete are considered.

Additionally, finishers working on kneeboards hand-finished concrete located around steel columns and alongside exterior walls of the building.

The top 1/8 inch

Hard-troweling the floor affected the top 1/8 inch of the floor slab the most. The troweling machines “compression dewatered” the floor surface, forcing a little water to the surface and the rest into the fresh concrete below. This dewatering process left the surface layer with an extremely low w/c ratio. The dark appearance of this surface layer is due to the low w/c ratio restricting cement hydration and concentrating cement particles—the cement acting as a dark gray pigment. The lighter gray color of non-hard-troweled floor surfaces is caused by the higher w/c ratio at the surface and in some cases by carbonation—calcium hydroxide (CH) in solution moves through capillaries to the surface where it combines with atmospheric carbon dioxide (CO2). However, hard-troweled surfaces greatly reduce the interconnectivity and size of capillaries and pores, thereby restricting the movement of CH to the surface. Dense, hard-troweled surfaces carbonate very slowly.

The hard-troweling process compresses the surface and removes pores that are usually visible in the microscope. The dark 1/8-inch-thick layer can be seen at the top of the core in the top photo on page 32. At a magnifacation of 45X, the scanning electron microscope photo on page 31 shows the top ¼-inch profile of the floor. The white colored grains of material congregated at the top are residual unhydrated cement particles. (The particles appear white with the microscope but are dark to the naked eye.) The high concentration of hard particles provides further evidence of the low w/c ratio.

In a normal concrete mixing process, there are always residual particles of cement scattered throughout fresh concrete. Steve Lewis, a chemist for Lehigh Cement, Doraville, Ga., points out the finishing process causes a buildup of cement paste and fine aggregate at the surface, increasing the concentration of residual cement particles.

With most of the water pressed out of the surface during finishing, there is no opportunity for most of these particles to produce hydrate material.