EVALUATING RISK TO PYRRHOTITE AFFLICTED CONCRETE FOUNDATIONS IN NORTHERN CONNECTICUT AND SOUTHERN MASSACHUSETTS

To date, the Trinity College pyrrhotite testing program has accumulated over a thousand data points on the concentration of pyrrhotite within concrete foundations of private homes, condominiums, municipal buildings, businesses, and construction aggregate from several active commercial quarries in both Connecticut and Massachusetts. We have established a testing program that combines two independent analyses: magnetic susceptibility loss (MSL) and total sulfur. MSL accurately identifies the presence of, and semi-quantifies the concentration of pyrrhotite. Total sulfur results are used to estimate the concentration of pyrrhotite when they are above typical concrete sulfur levels (ranging between 0.1-0.2% total sulfur, mean 0.14%) and serve as an independent check to the MSL analysis to ensure reliable results. As there are non-pyrrhotite sulfur minerals present in concrete cement and some aggregate, total sulfur or sulfide alone cannot reliably identify the concentration of pyrrhotite.

Assessing risk for concrete deterioration for any individual test result is possible if a relationship exists between pyrrhotite concentration and the degree of deterioration due to the pyrrhotite reaction. We use map (spider) cracking as the diagnostic visual symptom of the pyrrhotite reaction and have recorded the existence of map cracking, and where it occurs, in basement homes at the time of concrete core extraction. We present two methods to assess the likelihood that map cracking may occur: 1) We calculate the percentage of map cracked afflicted homes (regardless of age) for homes within a similar range for both MSL and total sulfur; 2) We calculate a logistic regression for the data set (R software, GLM function) where the binary dependent variable is the presence of (or lack thereof) map cracking at the time of analysis and the independent variables are the concentration of pyrrhotite (using either MSL or total sulfur) and the age of the concrete. Results indicate that map cracking is not observed in concrete with MSL values below 5 × 10^-8m³/kg. Based on the linear relationship between MSL and total sulfur, this translates to approximately 0.15% pyrrhotite by mass in concrete. Both the descriptive statistics and the logistic regression indicate that the probability of map cracking increases incrementally above this value.