Article pubs.acs.org/est
Potential Reversal and the Effects of Flow Pattern on Galvanic Corrosion of Lead Roger B. Arnold, Jr. and Marc Edwards* Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University, 407 Durham Hall, Blacksburg, Virginia 24061, United States S Supporting Information *
ABSTRACT: Simplistic conventional models predict that a greater mass of lead will be released from lead pipes exposed to higher velocity and flow durations. However, if galvanic Pb−Cu connections are present, or if a highly protective Pb(IV) scale can be formed, reduced flow can markedly increase the mass of lead release to water and resultant consumer exposure. Three chemical mechanisms were identified that can reduce lead release at higher flow including (1) formation of Pb(IV), (2) potential reversal of Pb:Cu couples, after which galvanic corrosion sacrifices copper and lead is protected, and (3) reduced formation of corrosive microenvironments at lead surfaces in galvanic couples. Potential reversal occurred only in the presence of free chlorine with continuous flow, and it did not occur with chloramine, with intermittent flow, or if orthophosphate was present. For both disinfectants, electrochemical measurements supported a mass balance of lead release demonstrating that a greater total mass of lead release occurred with intermittent flow than with continuous flow.
■
INTRODUCTION Consumer water usage habits affect water flow velocity, flow duration, water stagnation patterns, taste and odors, and corrosion of premise plumbing systems.1−3 Recently, reduced water demand and long stagnation times associated with new “green” construction, have been associated with elevated lead in water from leaded brass plumbing devices.4,5 These findings prompted concern regarding impacts of water conservation on lead release from millions of legacy lead service line pipes in the United States and around the world.6 Existing conceptual and numerical mass-transfer models of soluble lead release from lead service lines have proven useful in understanding certain trends in potable water lead concentrations as a function of hydraulic conditions,7,8 but they do not consider possible impacts of stagnation on galvanic corrosion or the role of water chemistry in evolution of protective scales (e.g., Pb(IV)). As a result they invariably predict that a greater mass of lead will be released by lead pipe when exposed to higher velocities and flow duration. Likewise, simplistic conceptualizations of particulate lead release from service lines by scouring and erosion corrosion,9 would also predict increased mass of lead release to water with a higher velocity or flow duration. Recent observations of anomalously high lead concentrations in situations with low water usage4,6,10−12 gave rise to concerns that the conventional wisdom regarding the relationship between flow and mass of lead release to water is not always valid. While field observations are hopelessly complicated by variations in plumbing from site to site and cannot provide proof of cause and effect, especially when considering dilution effects (i.e., the same mass of lead released to much lower © 2012 American Chemical Society
volume of water results in higher concentration), certain data have suggested that a greater mass of lead might actually be released to water in situations with lower water use and longer stagnation times. To the extent this is true, implementation of water conservation might unintentionally increase the mass of lead released to water. In many instances where anomalous behavior was suggested, the leaded plumbing material was galvanically connected to copper pipe. In parallel with the field observations, recent research on galvanic corrosion6,13,14 identified several mechanisms by which a reduced mass of lead release to water could be expected with increased flow duration. Specifically, during stagnation, a highly corrosive microenvironment characterized by a high chloride concentration and low pH (300 ppb frequently measured sporadically in consumer’s homes from 2001 to 2004.35 With the exception of the condition with free chlorine and continuous flow, the trends in galvanic current observed during flowing conditions were sustained after the switch to stagnation, most likely because the flow pattern and extent of exposure to oxidants also influenced important aspects of the scale present on each sample. With both disinfectants, the galvanic current after the switch to stagnation was higher in the pipes previously exposed to frequent flow events and lowest in the pipes previously exposed to continuous flow. Mass Balance. A mass balance of lead release across the experiment was performed, and the total mass of measured lead was compared to the predicted mass of lead corroded solely by the galvanic current, which was obtained by integration of current data and Faraday’s law. To the extent that lead corrosion occurred by uniform (nongalvanic) corrosion, higher levels of lead would be released to water when compared to this prediction based on galvanic current alone. The total mass of lead measured was the sum of lead measured in the bulk water, in the bottom of the basins, and in the scale from all surfaces including the lead pipe, copper pipe, and the tygon connector. Scale mass measurements confirmed visual observations that a greater mass of lead scale buildup occurred on pipes subjected to intermittent flow, and a higher percentage of the total measured lead was found in the scale on pipes with intermittent flow. The presence of such an accumulation of lead deposits has been linked to sustained problems with particulate release to
Figure 4. Total mass of lead release during final two weeks of normal exposure. During this period, nitric acid and hydroxylamine were used to improve total mass recovery of lead from the basins. The blue fraction represents lead captured in bulk water samples prior to acidification (not including particulate lead), and the red fraction represents lead captured in samples after acidification (including particulate lead).
Consistent with previous work regarding health concerns due to the sporadic release of lead particulates,13 a massive and seemingly random lead release occurred in one condition during the final two weeks of the experiment that was not necessarily representative of average lead release throughout the experiment. Effect of Long Stagnation. Following the termination of recirculating flow conditions, each of the eight lead−copper couples was subjected to stagnant conditions with initial water chemistry and disinfectant dose matching the recirculating flow phase. Upon the onset of stagnation when filling the Pb−Cu couple previously exposed to free chlorine and continuous flow, the lead pipe was still cathodic to copper with a galvanic current of −26 μA. However, after 4 days, the current reversed, causing the lead to once again become anodic with a current of 12 μA. During recirculating flow, the chlorine concentration was maintained between 3.5 and 4.0 mg/L; however, during stagnation, free chlorine was consumed in the corrosion process, and when the chlorine residual disappeared, the oxidation reduction potential of the water dropped below the threshold required for stabilization of Pb(IV), and breakdown of the Pb(IV) scale could begin. After the first week of stagnation, refilling the couple with fresh chlorinated water did not cause electrochemical reversal to occur again, as the current sacrificing the lead was maintained at 14 μA. The destabilization of the presumed Pb(IV) scale resulted in a massive release of accumulated scale that had formed during continuous flow with free chlorine. After imposing stagnant conditions, the average lead concentration in the water contacting the pipes was about 130 000 μg/L, which was more than 4 times higher than the average lead concentration in other conditions during long stagnation (Figure 5). 10945
dx.doi.org/10.1021/es3017396 | Environ. Sci. Technol. 2012, 46, 10941−10947
Environmental Science & Technology
Article
ultimately, to reduce lead corrosion rates. Contrary to simplistic expectations based on conventional models of lead release, in systems with galvanic corrosion, a greater mass of lead can be released to water in situations with lower flow duration, presumably due to enhanced formation of microcorrosive environments during stagnation events and improved passivation of Pb−Cu couples with continuous flow. Even when electrochemical reversal does not occur, continuous flow pipe-loop experiments can produce drastically different results than experiments conducted under stagnant conditions. Thus, in addition to water chemistry, flow patterns must be considered when comparing experimental results. In recirculating tests, sorption and settling of particulates, which can account for the majority of the total mass of lead release, can create difficulties in conducting rigorous mass balances. Other issues including highly recalcitrant solids formed in the presence of phosphate interfered with definitive closure of mass balances. In systems in which electrochemical reversal has occurred and lead is protected from corrosion by galvanic connection to copper, prolonged stagnation events (>4 days) or a switch to chloramine could reverse the protective galvanic effect and create conditions that could induce release of extremely high lead levels to potable water. To the extent such problems occur in practice, appropriate flushing of lead service lines after long stagnation events might mitigate public health concerns and (if chlorine was present) could potentially assist in re-establishing the potential reversal and its benefits for controlling lead in water. Homes with low flow patterns or prolonged stagnation may represent an additional risk factor for lead release. It has long been understood that schools and homes with low water use may have higher first and second draw lead concentrations, but the possibility that a greater mass of lead may also be released in such circumstances deserves increased scrutiny. As consumers respond to efforts for water conservation and given increased popularity of “green buildings”, it is important to understand how reduced water use could impact consumer exposure to elevated lead in drinking water.
water and increased risk of consumers to lead that poses an acute health risk.8,29 For all conditions without orthophosphate, approximately 100% or more of the lead predicted to be corroded by the galvanic current was measured in the experiment. The slightly higher values beyond the 100% value are certainly expected due to nongalvanic lead corrosion processes, but for these conditions the vast majority of the lead release was likely due to the galvanic connection to the copper. The overall weight loss of the lead coupons from the beginning of the experiment to after the removal of scale was also calculated. In general, weight loss corresponded well to the total mass of lead measured in the experiment. The mass balance also indicates that far more lead than predicted based on galvanic current alone was measured in the condition with continuous flow and free chlorine in which Pb(IV) formed. This result is not surprising because the net galvanic corrosion was very low due to the potential reversal, and Pb corrosion was occurring due to reactions on the lead surface, as per lead pipe alone. While a minimal amount of dissolved lead release and some sporadic particulate release continued to occur while the galvanic current was negative, the majority of the lead measured in this condition was likely derived from the beginning of the experiment, when lead corrosion proceeded rapidly with a galvanic current exceeding 100 μA. The total mass of lead recovered was much less than the overall weight loss only in this condition, confirming that a significant amount of particulate lead was probably lost in the beginning of the experiment when the basin was not acidified with nitric acid. When orthophosphate was present, significantly less than 100% of the predicted lead was measured during the experiment. This is a perplexing result that defied explanation based on attempts to improve recovery and account for the lead that was “missing”. However, it was discovered that the lead scale formed in the presence of the phosphate was extremely recalcitrant. Specifically, when samples of scale collected from each lead pipe with phosphate were acidified in a solution of 1% nitric acid, only 10−20% of the lead present was recovered after 2 h in a 1% nitric acid solution, whereas lead scale from pipes without phosphate dissolved completely in the same time. Recovery of the scale from lead phosphate system increased to 70−80% after 16 h in a 1% nitric acid solution. Thus, even the 2-h exposure to 1% nitric during the experiment probably underestimated the total lead released to the basin in conditions with orthophosphate. Future work would be needed to determine if this fully explains the discrepancy and to identify the exact lead species present in these conditions. Consistent with these results, other researchers have reported formation of lead phosphate solid phases (e.g., hydroxypyromorphite Pb5(PO4)3OH or lead orthophosphate Pb3(PO4)2) that are extremely insoluble in acid.1,21,36 These results have obvious practical implications for lead monitoring at the tap for drinking water systems with orthophosphatea very aggressive preservation and digestion protocol might be required to fully detect the actual concentration of lead in such samples if recalcitrant particulates are present.33 It is also possible that much of this lead would not be bioavailable. Practical Implications. In drinking water systems with free chlorine disinfection, potential reversal can occur for lead pipe and copper couples. This phenomena is accelerated by increased flow duration. The net result of potential reversal is to essentially eliminate galvanic corrosion of lead and,
■
ASSOCIATED CONTENT
S Supporting Information *
Details of the lead release mass balance, photos showing lead scale formation on the galvanically coupled coupons, additional electrochemical data, and water quality information. This information is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Tel:(540) 231-7236; e-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge the financial support of the National Science Foundation under grant CBET-0933246. Opinions and findings expressed herein are those of the authors and do not necessarily reflect the views of the National Science Foundation. 10946
dx.doi.org/10.1021/es3017396 | Environ. Sci. Technol. 2012, 46, 10941−10947
Environmental Science & Technology
■
Article
(25) DeSantis, M. K.; Welch, M.; Schock, M. Mineralogical Evidence of Galvanic Corrosion in Domestic Drinking Water Pipes. In AWWA Water Quality Technology Conference Proceedings, 2009. (26) Schikorr, G. The Cathodic Behavoir of Zinc versus iron in hot tap water. Trans. Electrochem. Soc. 1939, 76, 1. (27) Francis, R. Galvanic Corrosion: A Practical Guide for Engineers; The Corrosion Society, NACE International, 2001. (28) Xie, Y. J.; Giammar, D. E. Effects of flow and water chemistry on lead release rates from pipe scales. Water Res. 2011, 45, 6525−6534. (29) Cartier, C.; Arnold, R. B.; Triantafyllidou, S.; Prevost, M.; Edwards, M. Effect of Flow Rate and Lead/Copper Pipe Sequence on Lead Release from Service Lines. Water Res. 2012, 46, 4142−4152. (30) Edwards, M.; Dudi, A. Role of chlorine and chloramine in corrosion of lead-bearing plumbing materials. J. AWWA 2004, 96 (10), 69−81. (31) Giammar, D. E.; Wang, Y.; He, J.; Cantor, A.; Welter, G. J. Experimental investigation of lead release during connection of lead and copper pipes In AWWA Water Quality Technology Conference Proceedings, 2011. (32) Deshommes, E.; Laroche, L.; Nour, S.; Cartier, C.; Prevost, M. Source and occurrence of particulate lead in tap water. Water Res. 2010, 44, 3734−3744. (33) Triantafyllidou, S.; Parks, J.; Edwards, M. Lead Particles in Potable Water. J. AWWA 2007, 99 (6), 107−117. (34) Hulsmann, A. D. Particulate Lead in Water Supplies. Water Environ. J. 1990, 4, 19−25. (35) Elevated Lead in D.C. Drinking Water − A Study of Potential Causative Events, Final Summary Report; EPA 815-R-07-021; U.S. Environmental Protection Agency: Washington, DC, 2008. (36) Xie, Y.; Giammar, D. E. Equilibrium Solubility and Dissolution Rate of Lead Phosphate Chloropyromorphite. Environ. Sci. Technol. 2007, 41, 8050−8055.
REFERENCES
(1) Schock, M. R.; Wagner, I. Corrosion and Solubility of Lead in Drinking Water. Chapter 4 in Internal Corrosion of Water Distribution Systems; AWWA Research Foundation, 1985. (2) National Research Council. Alternatives for Premise Plumbing. Chapter 8 in Drinking Water Distribution Systems: Assessing and Reducing Risks; National Academies Press: Washington, DC, 2006. (3) Zhang, Y.; Love, N.; Edwards, M. Nitrification in Drinking Water Systems. Crit. Rev. Environ. Sci. Technol. 2009, 39, 153−208. (4) Elfland, C.; Scardina, P.; Edwards, M. Lead Contaminated Potable Water from Brass Devices in New Buildings. J. AWWA 2010, 102 (11), 66−76. (5) Nguyen, C. K.; Elfland, C.; Edwards, M. Impact of advanced water conservation features and new copper pipe on rapid chloramine decay and microbial regrowth. Water Res. 2012, 46 (3), 611−621. (6) Triantafyllidou, S.; Edwards, M. Galvanic corrosion after simulated small-scale partial lead service line replacement. J. AWWA 2011, 103 (9), 85−99. (7) Kuch, A.; Wagner, I. A mass transfer model to describe lead concentrations in drinking water. Water Res. 1983, 17, 1303−1307. (8) Cardew, P. T. Development of a convective diffusion model for lead pipe rigs operating in laminar flow. Water Res. 2006, 40, 2190− 2200. (9) Britton, A.; Richards, W. N. Factors influencing plumbosolvency in Scotland. J. Inst. Water Eng. Sci. 1981, 35, 349−364. (10) Lambrinidou, Y.; Triantafyllidou, S.; Edwards, M. Failing our children: Lead in US school drinking water. New Solutions 2010, 20, 25−47. (11) Nguyen, C. K.; Stone, K. R.; Dudi, A.; Edwards, M. A. Corrosive microenvironments at lead solder surfaces arising from galvanic corrosion with copper pipe. Environ. Sci. Technol. 2010, 44, 7076− 7081. (12) Breese, S. Factors influencing the Rate of Lead Release in Domestic Plumbing Systems. AWWA Annual Conference and Exhibition, 2008, Atlanta, GA. (13) Edwards, M.; Triantafyllidou, S. Chloride-to-sulfate mass ratio and lead leaching to water. J. AWWA 2007, 99 (7), 96−109. (14) Nguyen, C. K.; Stone, K. R.; Edwards, M. A. Chloride-to-sulfate mass ratio: Practical studies in galvanic corrosion of lead solder. J. AWWA 2011, 103 (1), 83−92. (15) Hu, J.; Gan, F.; Triantafyllidou, S.; Nguyen, C.K..; Edwards, M. A. Copper-Induced Metal Release from Lead Pipe into Drinking Water. Corrosion 2012, In press. (16) Wang, Y.; Xie, Y.; Li, W.; Wang, Z.; Giammar, D. E. Formation of Lead(IV) Oxides from Lead(II) Compounds. Environ. Sci. Technol. 2010, 44, 8950−8956. (17) Lytle, D. A.; Schock, M. R. Formation of Pb(IV) oxides in chlorinated water. J. AWWA 2005, 97 (11), 102−114. (18) Reiber, S.; Dufresne, L. Effects of External Currents and Dissimilar Metal Contact on Corrosion of Lead from Service Lines; Final Report to USEPA Region III: Philadelphia, PA, 2006. (19) Reiber, S. Galvanic Stimulation of corrosion on lead-tin soldersweated joints. J. AWWA 1991, 83 (7), 83−91. (20) Lytle, D. A.; Schock, M. R.; Scheckel, K. The Inhibition of Pb(IV) Oxide Formation in Chlorinated Water by Orthophosphate. Environ. Sci. Technol. 2009, 43, 6624−6631. (21) Lin, Y. P.; Valentine, R. L. Reduction of Lead Oxide and Release of Pb(II) in Mixtures of Natural Organic Matter, Free Chlorine and Monochloramine. Environ. Sci. Technol. 2009, 43, 3872−3877. (22) Dryer, D. J.; Korshin, G. V. Investigation of the Reduction of Lead Dioxide by Natural Organic Matter. Environ. Sci. Technol. 2007, 41, 5510−5514. (23) Switzer, J. A.; Rajasekharan, V. V.; Boonsalee, S.; Kulp, E. A.; Bohannan, E. W. Evidence that Monochloramine Disinfectant Could Lead to Elevated Pb Levels in Drinking Water. Environ. Sci. Technol. 2006, 40, 3384−3387. (24) Xie, Y.; Wang, Y.; Giammar, D. E. Impact of Chlorine Disinfectants on Dissolution of the Lead Corrosion Product PbO2. Environ. Sci. Technol. 2010, 44, 7082−7088. 10947
dx.doi.org/10.1021/es3017396 | Environ. Sci. Technol. 2012, 46, 10941−10947
