The surplus paste water is the most mobile and least dense of the four principal concrete components, and the only one capable of changing phase from liquid to gas. Immediately following placement, this surplus water starts to percolate upward, and the slab begins a process of sedimentation wherein the solid particles of cement, sand, and stone sink to the bottom while the lighter liquid water works its way, or bleeds, to the top.

Starting with the largest capillary pores, bleeding and self-desiccation combine to dewater the paste. While the larger pores are being emptied and the menisci forming that remain are more than 400 water molecules wide, very little horizontal shrinkage occurs. Though strongly adhered to the surrounding solids, these larger menisci are just too large to translate their surface tensions into enough force to pull the sides of the pore inward. Once the dewatering process reaches the pores that are less than 400 water molecules wide, however, the surface tensions of the smaller menisci being formed do become sufficient to draw in the pore walls. The aggregation of these myriad local surface tension-induced microcontractions is the phenomenon that makes the concrete shrink en masse.

Rule No. 12a: Drying shrinkage is caused by what happens inside the pores that are less than 2 millionths of an inch across.

Once the bleed water has made its way to the top and removed, the paste at the surface is exposed directly to the air. If the air’s temperature is higher than the dew point, then the air’s relative humidity will be less than 100% and the paste water will start to evaporate. Then the vaporization rate will be determined by four main factors: the air’s relative humidity, the air’s movement, the paste’s surface roughness, and the paste’s temperature. Indeed, every experienced finisher knows that his Window of Finishability—the imaginative term coined by Bill Phelan 25 years ago—always will close as the environment gets more arid, the concrete gets hotter, the wind blows harder, and/or the surface is left more open.

Rule No. 12b: In the first two years, a slab’s top layer will shrink about 1/32-inch for every 5 feet of its uninterrupted length.

Suppose a new plain, portland cement concrete slab is cut into 15x15-foot squares. If the concrete exhibits normal shrinkage and no visible cracks develop, the average sawcut width increases by about 3/32 inch. If a high-shrinkage concrete is used instead, the average joint grows about 5/32 inch. Conversely, if the concrete is of the low-shrinkage variety, then the sawcut widths only grow about 1/32 inch. Relative to the joint growths that are normally expected, the practical performance differences between high and low shrinkage concretes are barely significant—especially because any amount of joint growth, no matter how small, immediately raises the specter of stability problems. In short, the hand wringing that often attends consideration of a mix’s shrinkage potential is largely pointless.

Rule No. 12c: Regardless of slab thickness, at age D days the drying shrinkage experienced t inches from the top surface will approximately equal {-E+[t(E-t)]½}/10,000 inches/inch, where E=5{1-exp[-D/90]}.

The figure shows this formula plotted at D=360 days. The fact that this curve is unaffected by the slab’s depth has profound implications regarding the proper use of embedded reinforcement.