Concrete Failure
Discovering What Went Wrong

APS is frequently contacted to perform petrographic analysis to determine the cause of failure in the installation or repair of concrete. The following example illustrates a typical failure and the forensic process of determining the cause.

American Petrographic Services, Inc. (APS) teamed with a national parking ramp consultant to analyze a concrete repair failure from a parking garage. The repair material chosen by the designer was a silica fume-modified, air-entrained, cast-in-place concrete . The patch, a maximum 45 mm (1 3/4?) thick, had delaminated and a dinner plate-sized piece of the failed repair was submitted to our laboratory for petrographic analysis. Hand samples and thin sections were cut and polished for analysis.

The success of a concrete repair depends on several factors:

  • Removal of unsound concrete before the repair is attempted 
  • Preparation of the surface to receive the repair
  • The bonding method
  • Selection of repair materials
  • Skill of the craftsman
  • Curing

As part of the repair procedure, the contractor applied a cement slurry bonding agent between the patch and the substrate concrete (the existing concrete garage slab). Review of the failure specimen indicated the majority of the patch failed at the contact point between the bonding agent and the substrate concrete. Two other failure zones were noted: the patch material located above the bonding agent was poorly consolidated, and separation had occurred within the substrate concrete.

Magnified inspection of the failure plane revealed an excess of entrained-size air bubbles in the slurry along the contact point with the substrate concrete. The formation of these bubbles was possibly the result of using a type 1-A cement for the slurry and/or over-brushing/agitation of the slurry into the substrate. Our observations also disclosed the presence of sand and cementitious paste debris which adhered to the bottom surface of the cement slurry bonding agent.

Numerous, large consolidation voids present within the repair concrete suggested poor workmanship or poor workability of the repair material. Small pieces of the substrate concrete adhering to the bottom surface of the slurry suggested either localized areas of good bond, or damaged or ?bruised? substrate concrete. The latter could have been caused by over-impacting the substrate concrete with a jackhammer used to remove unsound material from the patch area. Excessive energy during the preparatory process probably produced planes of weakness such as micro-cracking in underlying sound concrete. Using the proper equipment initially, or if necessary, incorporating a secondary, less-aggressive method, may be needed to remove compromised material and other loose debris before performing the repair.

Numerous plastic shrinkage micro-cracks were present at the top surface of the repair, and carbonation (exposure of plastic concrete to carbon dioxide, resulting in a softer paste at the surface) was significant for a silica fume-modified concrete. These characteristics are indicative of ineffective curing. Poor curing may also produce curling stresses from shrinkage of the material, pulling the edge of the repair away from the substrate.

Although the design may have addressed all of the critical factors necessary for a successful repair scenario, work and quality control practices associated with the various steps will control the outcome of the repair. The distressed repair contains all the forensic information needed to diagnose what went wrong. The science of concrete petrography seeks to answer questions about why the repair failed, or in general, to solve construction material-related problems.

In addition to identifying causes of concrete failure,  APS can assist in mix design options and repair alternatives to maximize the longevity of concrete structures.