Non-Uniform Copper Corrosion in Potable Water: Theory and Practice
While it has long been known that water flow can influence non-uniform corrosion of copper pipe (Rushing 2002, Marshall 2004), it has never previously been considered a primary contributor to the problem. This work is the first to describe a fundamentally important phenomenon in aqueous non-uniform corrosion: flow electrification. A conceptual framework for flow electrification was developed from prior work on non-aqueous fluid flow in pipes, where the primary concern was prevention of electrical explosions. Thereafter, a series of experiments was aimed at monitoring flow electrification, quantifying its practical effects, and examining aspects of non-uniform copper corrosion in real situations.
Under conditions with little or no flow, in a high pH and high chlorine water known to cause pinholes in copper pipes, while water chemistry influenced corrosion, non-uniform corrosion was not sustained. But when flowrates were higher, flow electrification contributed to severe and sustained non-uniform corrosion, with the most serious attack manifested in the first section of pipe that encountered the flowing water. The magnitude of flow electrification increased with chlorine concentration, pH and flowrate. Containers dosed with inhibitors such as zinc or phosphate experienced lower electrification, current, voltage, scale resistance, and corrosion potential measurements when compared to a control without an inhibitor. Additionally, systems dosed with inhibitors had reduced rate of chlorine decay, weight loss, pit density and/or maximum depth. Zinc orthophosphate had the largest positive impact on electrochemical measures of pitting. However, experimental studies suggested that if zinc orthophosphate was dosed to a system for a period of time, and dosing was then stopped, very serious pitting could occur. A practical case study seemed to strongly confirm this hypothesis in one system.
The presence of sulfides caused the separation of anode and cathode along a pipe section, from electrification, to reverse relative to what was observed in the system with high chlorine and high pH. Below a certain level of sulfides, electrification ceased. It seems likely, based on measurement of electrochemical potential (Ecorr) in waters of this type as a function of sulfide concentration, that the onset of pitting would be associated with decreasing Ecorr with time. If so, the fundamental basis of tracking Ecorr rise with time to predict pitting propensity would be invalidated.
Electrochemical noise programs were applied to try to differentiate between systems of low and high pitting propensity. Amplitudes of potential noise and current noise measurements drastically increased with the presence of sulfides or chlorine, confirming that tracking electrochemical noise may indicate the presence of a pitting agent. However, the electrochemical noise measurements are at best, an indirect indicator of copper pitting, and their interpretation is complicated by the co-occurrence of flow electrification.
Attempts were made to apply these insights to a case study examining pitting in two real waters in Maryland and to examine the effects of orthophosphate, chlorination, chloramination and enhanced coagulation on copper pitting propensity. Tracking of Ecorr suggested that the control water (without phosphate) had the greatest pitting propensity. The free chlorine system with orthophosphate maintained Ecorr values below the control signifying that the presence of phosphate lowered the corrosion propensity. Waters with chloramine and phosphate had the lowest Ecorr values and also had the least variability in Ecorr due to the stability of the disinfectant. There is considerable ambiguity in the results, since the copper pipes exposed to the waters in question did not develop serious pitting over the several months of the study.