Ordering Information:
ORDER NUMBER:  91210
DATE AVAILABLE: Winter 2007/2008

Printed Report		PDF
Subscribers	Order Report
Non-Subscriber	Coming Soon - AWWA Bookstore	Coming Soon
	Coming Soon - IWA Publishing	Coming Soon

PRINCIPAL INVESTIGATORS:
Jeffrey Parks, Marc Edwards, Peter Vikesland, Matthew Fiss, and Abhijeet Dudi

OBJECTIVES:
The objective of this project was to examine the effects of bulk water chemistry on concrete corrosion and autogenous repair of concrete.

BACKGROUND:
Little research expenditure has gone towards the investigation of processes responsible for the corrosion of concrete drinking water infrastructure. One important mode of attack on concrete involves access of corrosive water to underlying reinforcing steel through cracks. Under some circumstances it is known that cracks in the concrete surface can repair themselves through reactions with constituents in the water. This phenomenon has been termed “autogenous healing.”

HIGHLIGHTS:
This project identified that autogenous healing can be controlled by water chemistry, and it can be an effective method of sealing small cracks that initiate in concrete before they develop into larger cracks that can cause failure. When high levels of magnesium and silicon are present this seal can develop high strength. Carbonation of internal surfaces was not detected when cracks were sealed via autogenous healing. In some cases, water permeability and chloride diffusion are impeded by autogenous healing regardless of strength attained.

APPROACH:
A new technique was developed to simulate autogenous repair of cracks in concrete structures. Concrete was prepared from ½” maximum coarse aggregate and Portland cement. Small coupons were cut into two halves and the crack width was controlled through the use of 0.25 mm thick polyethylene spacers. These coupons were suspended in two liter containers with water containing various combinations of calcium, magnesium, phosphorus, silicon, zinc, natural organic matter, and carbonate at pH 7, 8, 8.5, or 9.5 and gently agitated. Experiments were conducted in triplicate by suspending three specimens in each container. Changes in water chemistry were monitored and specimens were subjected to strength testing, chloride diffusion testing, and carbonation testing. Surface precipitates were evaluated by scanning electron microscopy with energy dispersive spectroscopy (SEM-EDS) and Raman spectroscopy.

RESULTS/FINDINGS:
Results indicated that autogenous healing is most certainly controlled by the water chemistry. The formation of a precipitate that restored strength to the concrete, or “glued” the separated halves together, only occurred consistently in waters with high levels of magnesium and silicon. In some cases the strength of the glue was quite strong. For example, an average force of 215 N was required to break apart the autogenously healed specimens in waters that had pH 9.5, 40 mg/L magnesium, and 35 mg/L silicon. SEM-EDS analyses indicated the presence of magnesium and silicon on the crack surface after these higher strength specimens were sheared apart.

The extent of healing that occurs can be quantified in a number of ways. First, the most beneficial type of autogenous healing involves filling of the crack with restoration of strength with a magnesium silicate solid. In other instances, the crack was filled by precipitates that did not restore strength to the separated sections of concrete, but nonetheless, could serve as a barrier to diffusion of chloride and other corrosive constituents (e.g., CO₂ ) to the underlying rebar and concrete. Formation of a physical barrier without restoring strength was shown to significantly impede transport of chloride.

IMPACT:
The use of the Langelier Index (LI) to determine whether concrete corrosion will occur in distribution systems is widely accepted although it is not always reliable. LI cannot indicate when appreciable autogenous healing of cracks in concrete will occur with restoration of strength. Obviously, other factors such as magnesium and silicon concentrations play an important role and must be explicitly considered if an index were to be developed that describes autogenous repair.