You are here:
Actual vs. Predicted Lead Scale Observations from the Field
Citation:
Tully, J., M. DeSantis, AND M. Schock. Actual vs. Predicted Lead Scale Observations from the Field. AWWA Water Quality Technology Conference, Toronto, Ontario, CANADA, November 11 - 15, 2018.
Impact/Purpose:
Theoretical chemical models industry uses to predict scale formations within the DWDS are built on the assumption of equilibrium with simple scale-forming lead minerals. DWDSs can contain many variables not accounted for within the models, and a disconnect often arises between the predicted and reality. The actual mechanisms controlling lead release are assessed based on a scale analysis survey from a variety of US Midwestern systems. When possible, this data is coupled with water lead levels to estimate the efficacy of those scales in reducing lead levels. Frequently, the actual mechanisms controlling lead release were unpredictable using equilibrium solubility modeling, and only recognized by analyzing the scale from an in-service pipe.
Description:
The conventional wisdom of lead solubility has been built over the years by geochemical solubility models, experimental studies, and actual field sampling of drinking water lead levels by differing protocols. These models predict that lead solubility and release would be controlled by an expected mineral phase in a system with given specific water quality parameters. Rarely, have the phases formed in real world drinking water distribution system (DWDS) pipes been compared back to theoretical predictions. In this survey, the scale from at least one LSL offered to EPA from 24 DWDSs in the US EPA’s Region 5 were evaluated. These public water systems represent a range of treatments including pH/alkalinity adjustment, blended phosphate, orthophosphate, and lime softening. Based on theoretical modeling 8 are predicted cerussite systems, 5 hydrocerussite, and the 11 systems which use a phosphate inhibitor are predicted to form a lead (II) orthophosphate mineral. The pipe scale analysis used some combination of X-ray diffraction (crystalline compounds), energy- and wavelength-dispersive X-ray spectrometry (elemental composition) to compare between the actual and predicted mineralogy. Corrosion control effectiveness was explored with limited sequential profile water sampling from 7 of the studied systems. This information combined with available plumbing information was used to assess the extent of premise plumbing lead release. For systems without profile data, EPA’s LCR 90th percentile data for the system was used to roughly assess corrosion control effectiveness, even though LCR values are not frequently representative of water that has been in direct contact with the LSL. This review indicated that the scales observed in real world DWDSs often do not follow the model predictions, but rather contain a mixture of expected and unexpected phases. Of the different treatment types, pH/alkalinity adjusted and lime softened systems most frequently matched the predicted phase, whereas phosphate dosed systems were the least predictable. In total, 13 of the 24 systems were found to have <50% of the predicted phase comprising the scale. Unpredictable and unexpected phases in the DWDS can pose problems to water systems as pipe scales can vary by location and small changes in water quality can create detrimental consequences. Further, amorphous materials were discovered within the pipe scales of 8 of the water systems surveyed. This creates an additional complicating factor in predicting lead release as amorphous materials are not amenable to quantitatively predicting their lead release or responses to changes in numerous water chemistry variables. Results of the profile sampling showed that in the cases of the 4 systems that were meeting the LCR action level at the time of sample the average 1st liter lead value was below 8 ppb. However, in all but one of those systems the LSL maximum was measured to be well above the LCR action level, ranging from 36-51 ppb Pb.