Bleed water should never be finished back into the slab surface.
Adobe Stock / AH Photos WPG Bleed water should never be finished back into the slab surface.

While some of the things that cause problems with concrete slabs have changed over the past 20 years, many (maybe most) have not. Looking back though the Concrete Construction archives, we find many articles that are as relevant today as they were when written. Here, then, are three common problems with concrete slabs and solutions based on previous CC articles. We’ll cover three more of these in the next issue.

This project had variable slump and setting behavior so the finisher had to squeegee to move the excess bleed water off the slab in some areas. This crew did a great job not finishing the bleed water back into the surface.
Scott Tarr / North S.Tarr Concrete Consulting This project had variable slump and setting behavior so the finisher had to squeegee to move the excess bleed water off the slab in some areas. This crew did a great job not finishing the bleed water back into the surface.

Problem: Excessive Bleeding
Bleeding is a result of the heavy components in concrete (sand and aggregate) settling and pushing the extra water of convenience to the top. Bleeding isn’t always bad. It lowers the water-cement ratio and densifies the concrete. But concrete that bleeds too fast or for too long can cause a number of problems: rock jams in pump lines, sand streaks in walls, weak horizontal construction joints, and voids beneath rebars and
aggregate particles.

Even when bleeding isn’t excessive, finishing concrete flatwork at the wrong time causes different problems. Finishing before bleed water has evaporated can cause dusting, craze cracking, scaling, and low wear resistance. Working bleed water into the surface also increases permeability, allowing water, de-icing salts, and other harmful chemicals to enter the concrete more easily.

Solution: Excessive bleeding can be avoided. Don’t add too much water to the concrete. Most of the water added to make placing easier bleeds out of the concrete. Any time saved during placement will be lost while waiting for the bleed water to evaporate. Place concrete at the lowest possible slump. If you need a higher slump to speed placement, consider using a superplasticizer. Add additional fines in the concrete, such as:
More finely ground cement. Concrete made with high early strength (Type III) cement bleeds less because the cement is ground finer than normal (Type I) cement.
More cement. At the same water content, rich mixes bleed less than lean mixes. But also consider that higher-strength slabs can curl more.
Fly ash or other pozzolans. However, Bruce Suprenant advises finishers to consider the effect of pozzolans on fresh concrete properties. The set time and bleeding of the fresh concrete mix can change so adjust finishing practices accordingly.
If concrete sands don’t have much material passing the No. 50 and 100 sieves, blend in a fine blow sand at the batch plant.
For air-entrained concrete, use the maximum allowable amount of entrained air. Entrained air bubbles act like additional fines. Air entrainment also lowers the amount of water needed to reach a desired slump.

To explore bleeding in more depth, check out these articles:

Curling at the joints raises the edges of the slab as shown by the elevated ends of the level.
Scott Tarr / North S.Tarr Concrete Consulting Curling at the joints raises the edges of the slab as shown by the elevated ends of the level.

Problem: Slab Curling
Curling is the rising up of a slab at corners and edges adjacent to contraction or construction joints or sometimes at a mid-panel crack. It occurs when the top of a slab dries or cools more rapidly than the bottom. As the concrete near the surface contracts, the edges rise off the subbase. Curling caused by surface drying is more common than that caused by surface cooling. “We measure curling using a Dipstick in a line diagonally from corner to corner on panels,” says Scott Tarr, North S.Tarr Concrete Consulting. “I have measured a curl of 1.5 inches in a 12.5-foot-square panel.”

The loss of subbase support that’s a result of curling can put the top of the slab in tension when under load, often causing cracks parallel to the curled joint. Curling also can cause rocking of the slab under load, joint spalling, floor-covering distress, and failure of joint fillers.

Solution: A curled slab can be repaired but it’s costly. Grinding can bring the edges back in line, although there’s always the danger that the curling will relax or continue if done too soon. Jerry Holland, Structural Services Inc., recommends: Grout to restore subgrade support and do nothing else unless you have to. But there are ways to control curling before it happens, involving design decisions, material choices, and construction practices.

Design decisions: When joints in unreinforced slabs are spaced at 15 feet or less, edges don’t curl as high but there are more joints at which curling is possible. At greater joint spacing, however, mid-panel shrinkage cracks may form and the slab can curl there. Joint spacing decisions are a trade-off between the potential for shrinkage cracks and the potential for joint curling. Another approach is to eliminate joints altogether and use mild steel reinforcing bars on supports that keep them 1 inch (for #4 bars) or 1½ inch (for #5 bars) below the floor surface. Other joint-free floor systems, some that include steel fibers, have recently come into greater use.

Material choices: Another way to reduce curling is to specify concrete strengths no higher than necessary for the structural capacity and abrasion resistance. High-strength concrete shrinks more and creeps less, aggravating curling problems. Reduce the paste content by using as much well-graded coarse aggregate as possible at the largest feasible maximum size—include 1½-inch top-size aggregate if possible. Minimize total water content, not water-cement ratio, and avoid soft aggregate that increases shrinkage. Be careful of admixtures that can increase shrinkage but consider shrinkage-reducing admixtures.

Construction practices: Keep the subgrade as dry as possible so it can accept water from the slab—obviously not possible if you are placing directly on a vapor retarder, which you should be for most interior slabs. Make sure the reinforcement stays in the proper location. And never add water on-site to retemper the concrete. Wet curing can allow water to seep through the contraction joints, keeping the bottom of the slab wet.

To explore curling in more depth, check out these articles:

GPR devices can be used to determine slab thickness.
Scott Tarr, North S.Tarr Concrete Consulting GPR devices can be used to determine slab thickness.
Cores taken from various locations in a single slab ranged from 2.5 inches to 8 inches in a slab specified to be 7 inches thick.
Scott Tarr / North S.Tarr Concrete Consulting Cores taken from various locations in a single slab ranged from 2.5 inches to 8 inches in a slab specified to be 7 inches thick.

Problem: Thin Slabs
Owners expect that their slabs are at least as thick as specified. But how often is that truly the case? Project specifications will typically reference ACI 301, Specifications for Structural Concrete, which references ACI 117, Specification for Tolerances for Concrete Construction and Materials. ACI 117-10 (Section 4.5.4) requires slabs-on-ground to have an average thickness no greater than 3/8 inch less than the specified thickness and to have no individual thickness measurement greater than 3/4 inch less than specified. For suspended slabs, the requirement is that the thickness can be no more than 1/4 inch less than specified. There is no longer any tolerance for how much thicker the slab can be than specified.

Slab thickness can be measured by taking cores or using ground-penetrating radar (GPR). Scott Tarr describes how to use GPR: “Unlike impact echo that only spot checks, the GPR allows you to see the continuous (60 to 90 points per foot) variation in slab thickness in real time as you scan the floor or pavement. The newer units are smaller and can be run at walking speed. If the exact thickness is desired, we take cores. The GPR measures the time it takes the signal to reflect back to the receiver (in nanoseconds). This time is correlated to slab thickness by taking and physically measuring the thickness of core samples.”

Solutions: Is the tolerance required by ACI 117-10 realistic? In a June 2000 interview with floor consultants Armand Gustaferro and Eldon Tipping, both stated that the slab tolerances in effect then were unreasonable, ridiculous, and mostly unachievable—current tolerances are only slightly better. “Current technology does not provide the contractor with a cost-effective method of achieving these goals,” Tipping said.

Clear back in 1989, Gustaferro looked at more than 3,000 cores from four different floors in four different states built by four different contractors and found that the average thickness of all four floors was less than the specified thickness. He also calculated the standard deviation of the thickness of each floor, an indication of how much the thickness varies. Assuming a normal distribution of the thickness values (a bell-shaped curve), about 68% of the floor thickness measurements will be within one standard deviation.

So say we have a floor that is specified to be 4 inches thick but the average measured as-built thickness is 3.9 inches and the standard deviation is 0.5 inch. Then 68% of the values will be between 3.4 and 4.4 inches thick. But that also means that 16% will be thicker than 4.4 inches and 16% will be thinner than 3.4 inches, which is getting dangerously close to the ACI 117 minimum thickness for any sample of 3.25 inches.

ASCC Position Statement #9 on this topic comes to the conclusion that based on an analysis of as-built data, for a 4-inch-thick slab the average thickness will be 35/8 inches and the thinnest measurement would be about 21/2 inches. The statement cautions designers that “if this is not acceptable, the specified slab thickness should be increased.”

Another approach for a contractor is to measure the thickness of several of their floors (if the owner agrees) and determine the standard deviation. This indicates how much the thickness normally varies on that contractor’s floors and what average thickness is needed to meet the specified tolerance. Gustaferro’s 1989 article concludes that the slab has to be thicker if the standard deviation is higher. His advice for contractors and specifiers remains true today:

“To comply with the ACI tolerance, contractors would have to closely control fine grading of the subbase and base course. Such extreme control may not be worth the cost in terms of improved slab performance. It might be more cost-effective to make the slab thicker than specified than to control the fine grading so closely for the required tolerance.

“On the other hand, designers should be able to feel confident that a slab specified to be 5 inches thick is reasonably close to that. Is it reasonable that 16% of the floor is less than 4 inches thick and only 30% is thicker than 5 inches? I don’t think so. I think the designer and owner deserve something better.

“A realistic specification could read, ‘The average thickness of a floor slab shall be no less than the thickness shown on the drawings, and not more than 16% of the floor will be thinner than 3/8 inch less than that shown on the drawings.’

“Such a tolerance specification, based on standard deviation, rewards contractors who closely control fine grading and finishing. On a 100,000-square-foot project, a contractor who can achieve a standard deviation of 0.50 inch will use about 154 cubic yards of concrete less than if the standard deviation is 1 inch.”

To explore slab thickness tolerances in more depth, check out these articles: