I would like to comment on a question in the December 2000 Problem Clinic. The problem was blistering of epoxy-terrazzo applied to a concrete slab with a 4-inch layer of granular fill over the vapor retarder. The question was if a vacuum pump could be used to create a vacuum in the fill and reverse the migration of moisture into the slab.

The blistering of polymer-based coating materials is most often referred to as osmotic blistering. The fact that such blisters have developed is a strong indication that the coating system has not developed proper bond with the substrate. Osmotic blistering does not seem to develop where polymer materials have properly penetrated the substrate and polymerized (cured).

The surface of a slab may appear dry, yet a significant amount of moisture may still be present within the concrete. The concrete surface may also be at the dew point at the time the flooring material is applied. In either case there may be enough moisture present to interfere with the proper penetration and curing of the applied material. The material may appear normal, when in fact it has not achieved tenacious bond with the substrate or proper chemical curing.

As moisture from within the concrete condenses beneath the sealed surface, osmotic cells may develop. A slab need not be on grade for such blisters to develop. We have investigated a number of such problems on elevated slabs as well. In an osmotic cell in concrete, water will be drawn towards the cell in an attempt to equalize the concentration of chemicals in the water on each side of the cell. An osmotic cell can develop several hundred psi of pressure, which accounts for water squirting upward when the blister is broken open. Osmotic cells may be localized zones or simply the weakest points of a more universally affected area.

While it will not correct the basic deficiency, the idea of drying the fill course beneath the slab could conceivably help reduce the potential development of blisters by osmosis. However, I know of no vacuum process to reverse the migration of moisture in the slab other than purposely constructed subslab radon mitigation systems with perforated piping, crushed stone, and piping fans. Coring vent holes, tenting, and using special dehumidification units could help remove moisture from the fill course. The process also would likely expose any other weaknesses in the applied system. Tensile bond tests of the applied system should then be conducted both in nonaffected areas and immediately adjacent to the blistered sites. The results of these tests will help establish whether the system can be repaired locally or if the entire system will need to be removed and replaced.

Peter Craig, Concrete Constructives, Greene, Maine